4 How to read the guide book? 11 Smart energy balancing of neighbourhoods 27 Energy efficiency and resilience for smart- grids 39

Transcription

1

2 4 How to read the guide book? 5 What is e-balance? 8 Who is the guide book for? 9 Apply our results 10 Disclaimer 11 Smart energy balancing of neighbourhoods 12 Challenges and goals 15 The e-balance energy management platform 16 The Energy Balancing Algorithms 18 Integrating Systems and Solutions 23 How to deploy the solutions 24 Bronsbergen experience 25 Vision for future developments of energy management systems: transforming academic simulations into virtual advisors 27 Energy efficiency and resilience for smartgrids 28 Challenges and goals 30 Smart-grid Low Voltage features 32 How to deploy the solutions 32 Batalha experience 34 Smart-grid Medium Voltage features 37 How to deploy the solutions 37 Batalha experience 39 Business models in a nutshell 41 Empowering new market players 45 Sustainability of the e-balance platform 48 Social studies: end-user perspective 49 Challenges and goals 54 Activities and results 56 Lessons Learnt 61 Acknowledgements and project consortium

3 IHP GmbH Innovations for High Performance Microelectronics Peter Langendörfer Krzysztof Piotrowski INESC INOVAÇÃO INSTITUTO DE NOVAS TECNOLOGIAS Mário Nunes Augusto Casaca António Grilo Universidad de Málaga Manuel Díaz Daniel Garrido Jaime Chen CEMOSA Noemi Jiménez Redondo Juan Jacobo Peralta Escalante Ana Navas Torrejón University of Twente Marco E. T. Gerards Marijn R. Jongerden Alliander N.V Marcel Geers University of Lodz / Management Faculty / Computer Science Department Bozena Ewa Matusiak Witold Bartkiewicz Ośrodek Przetwarzania Informacji (National Information Processing Institute) Jaroslaw Kowalski EFACEC Energia Alberto Jorge Maia Bernardo ISBN Graphical design by CEMOSA. e-balance consortium, Page 3

4 Dear reader, This guide book describes the experience of the e-balance consortium during the last four years developing different ICT tools and a platform to manage the electric energy in smart neighbourhoods. e-balance is not only about developing an innovative ICT tool or platform to manage several low voltage and medium voltage grid services. In addition, it has carried out an intensive social study in three European countries (Poland, the Netherlands and Portugal). Furthermore it has also designed new business models to create a sustainable framework to deploy the results and get an effective energy balancing, where DSO, energy aggregators and prosumers play in a fair win-win environment. This document starts with the description of the energy management platform and the different functionalities and use cases that covers software (middleware and balancing algorithms), privacy protection, and smart-appliance integration by using fractal-like system architecture. The third chapter describes a new management system for LV and MV grids to increase the resilience with several functionalities such as fraud detection and location, calculation of losses, detection and location of fused luminaires, among others. The fourth chapter proposes an innovative business model framework to enable the e- balance system and other demand response solutions, where the energy aggregator and the prosumers play the central role of satisfying DSO criteria whilst they increase their revenues and reduce the energy costs respectively. The fifth chapter describes the actions aimed to increase the energy awareness of endusers (prosumers). Surveys, user-centred analysis and other engagement methods have been used for this purpose. Finally, in the last chapter a list of lessons learnt are presented to advise future users and implementers which are the most important actions that should be done or should be avoided to succeed. We hope you will enjoy discovering e-balance! Page 4

5 The e-balance blueprint The e-balance project aims at developing hierarchical management system architecture for future smart-grids inspired by the fractal nature of electric and energy grids. As fractal systems, the electric grid is structured using the same organisation (connections) at different levels, what means that the same or at least very similar management approaches can be applied at different levels (e.g. household, neighbourhood, district, etc.). Example of fractal system (source: Wikipedia) The core of the e-balance architecture are units denoted as management units, which control the different fractal levels in the electric grid and scale from the customer management units (dealing with a single household) via low voltage grid management units (dealing with energy balancing in a neighbourhood) up to top level grid management units that are in control of large regions of a certain country. The management units are connected using the e-balance communication platform that abstracts all protocols and a major part of the security issues from the management units and the applications running on top. This system architecture is in compliance with the Smart Grid Architecture Model proposed by the European standardisation committee CEN-CENELEC thus it can be adapted and replicated in other power systems and environments. The e-balance architecture is providing a kind of a blueprint for implementing ICT solutions for smart-grids. In the course of the project we realized two complementary instantiations of our architecture of which each one was selected to highlight a certain aspect i.e. energy balancing and grid resilience. Even though in both cases the architecture is the same, specific implementations such as low level communication protocols and high level applications are different. Page 5

6 e-balance instantation I: energy balancing An ICT platform to enable energy balancing at different grid levels. It is composed of several management units (mini-pc or virtual machines) deployed at the homes of customers/prosumers as well as at the low-voltage grid that communicate with each other in a hierarchical manner. The goal is to satisfy both prosumers and DSO s requirements with the intervention of energy retailers or aggregators as business enablers. At the home level, a Customer level Management Unit (CMU) is connected to existing smart-appliances (in our instantiation to Whirlpool and Mastervolt appliances) to schedule their energy profiles when it is more appropriate. In the following upper level, we can find the grid management units in the different power substations. In the secondary substations the low voltage grid management unit (LVGMU) is installed, which communicates with the CMU to balance the distribution grid, and with the upper level to check the grid status and combine the individual LVGMUs in a similar manner as the LVGMU does for the CMUs. Example of the e-balance architecture for energy balancing e-balance instantation II: grid resilience Innovative management platform for low-voltage (LV) and medium-voltage (MV) grids to increase the efficiency and resilience of a power system. The e-balance demonstrator for LV grids focuses on creating the means to enable the Distribution System Operator to: measure in near real-time consumption/production power flows (including distributed energy resources), detect and locate commercial losses (frauds), locate and detect faults, calculate energy losses and the quality of service (QoS) accurately. On the other hand, the MV solutions focus on optimising power flow, validation of optimised solutions, fault detection and location and self-healing. Page 6

7 e-balance is also An innovative approach to integrate demand response services at the neighbourhood level, avoiding the rebound effect of new peak load generation. Nowadays, demand response initiatives are based on dynamic prices, which allow users to select the most suitable periods to consume or generate electric energy. However, with current EU regulations that do not allow discrimination of prices between users, dynamic prices entails that multiple users will establish the same strategies to save money with lower energy prices, which creates new power peaks and managerial issues. On top of the platform, the e-balance s balancing solution follows the original approach of the University of Twente (named Triana), which suggests a negotiations process between grid operators and prosumers to flatten the whole electric energy profile. An advanced solution to enable controlling distributed energy resources. One of the e-balance partners (INOV) has developed and patented a dynamic control method of power injected into the power grid by distributed generators. This invention defines a method for dynamically controlling the power injected into an electric grid by micro- and mini-producers of energy (up to kw), which are connected to a branch starting at a transformation station. The related algorithms are implemented in a processing module that receives electric measurements (e.g. voltage, current, power ) from connected renewable generation sources and aims at optimizing the production considering technical and legal constraints. New business models to empower future smart-grid players. As a whole, smartgrids should provide means to manage energy flows, demand and consumption to achieve the maximum energy efficiency and satisfy users needs, extending the highlevel management of electric grids to households or appliances level. The e-balance project has analysed the roles and needs of smart-grid users in order to create a functional market for smart-grids: prosumers (consumers that generates electric energy), aggregators, energy suppliers, DSO and balance responsible parties are the main players of this new system. The integration of the aggregator role is the cornerstone of the proposed models, since it is based on trading with the available flexibility; a concept that must be defined, measured and calculated accurately to enable the business models. This integration entails changes and new business structures and procedures that have been widely discussed during the project. Research on social aspects to integrate the end-user perspective. The energy and its systems are part of the everyday life of people, which is the final end-user of all the developments and deployment. e-balance proposes new functionalities and services that can be very challenging for specific segments of the society. For this reason, the identification of potential barriers to adapt the system to people s requirements has been paramount to understand how people feels with these systems and how integrating effectively them at their homes. Page 7

8 The target readers of this guide book are involved in the value chain of energy services offered by e-balance: electricity market operators (distribution system operators, energy retailers, energy aggregators, balance responsible parties...), distributed energy producers, city authorities and smart-appliance manufacturers that request innovative solutions to turn power systems into smart-grids. A DSO interested in improving the energy management of low-voltage and medium-voltage grids can find different platforms to calculate losses, detect, prevent and locate faults, detect fraud and restore grid service automatically with self-healing mechanisms. In addition, one of the most innovative results (patented) is a power control algorithm for distributed energy producers of renewable energy. On the other hand, energy retailers, aggregators and balance responsible parties will discover an innovative energy management platform to enable demand response services at the neighbourhood level with power based scheduling. This platform aims at satisfying both prosumers and DSOs criteria in a consolidated business model, in which all the parties are benefited of the energy balancing. City authorities, policy makers and public institutions have the opportunity to learn about main technical, social and legal requirements to create smart-cities regarding the power system. For instance, a public lighting management system is presented to reduce maintenance costs. Finally, manufacturers of smart-appliances will learn how washing machines, dryers and power inverters have been integrated in the whole platform, defining new data structures that can be used. Page 8

9 Most of the technical and scientific information and resources that e-balance offers are included in deliverables developed by the consortium until July All of them are open, freely accessible and can be used without any particular limitations (except where otherwise noted), keeping the public dissemination policy that characterised the project. Find them on the e-balance website, under Publications Besides the freely available resources, you might also like to use or test services and products that should be purchased and customized from partners or rely on consultancy work. In particular: For information or further research projects on the energy balancing platform for neighbourhood and smart-grids write to: Dr.-Ing. Krzysztof Piotrowski, For information on the adoption and customization of the LV/MV grid management tools and control of distributed energy resources write to EFACEC: Paulo Delfim Rodrigues, Smart Grids Division Manager, For permission and information about the patent Dynamic control method of power injected into the power grid by distributed generators - PCT write to: Mr. Mário Serafim Os Santos Nunes, mario.nunes@inov.pt Otherwise, for other general and technical inquiries on e-balance exploitation you can write to info@e-balance-project.eu Page 9

10 e-balance was funded by the European Union Seventh Framework Programme (FP7/ ) under grant agreement n e-balance Consortium has worked by following the rules defined in the Consortium Agreement (CA). In particular, Section 8 (Foreground) of the CA regulates the ownership and the exploitation of knowledge and Intellectual Property (IP) generated by Consortium partners. The ownership of Intellectual Property Rights (IPR) and the protection of foreground are based on articles II.26-II.29 of the EC Grant Agreement (GA), with special provisioning for: Joint ownership; Transfer of foreground. Every e-balance partner is the sole owner of any knowledge developed by that partner. The joint ownership of the intellectual property is ruled by articles in the EC-GA, with additional provisions reported in the CA in the case in which no joint ownership agreement has been reached among the interested parties. The graphs and pictures edited in this guide book are provided by the e-balance consortium and several artists from FreeImages.com and Pexels.com. Page 10

11 .

12 Due to a move to a sustainable society and with government incentives to support this, renewable energy sources become popular. End customers take part in this development by installing (rooftop) photovoltaic (PV) panels to supply their own energy. Furthermore, due to environmental and economic reasons, more and more customers procure electric vehicles (EVs) and heat pumps. However, the PV production peaks around noon, while EVs and heat pumps have the peak consumption at the beginning of the evening. This development leads to a mismatch of production and consumption, an unbalance in the electricity grid. In the new situation, there will be two peaks: around noon (production peak) and in the evening (consumption peak). These peaks will be significantly higher than what was historically observed in the grid. They result in higher grid losses, congestion and a lower lifetime of grid assets. The DSO has to pay for grid losses, replace grid assets sooner than anticipated and needs to install more costly infrastructure. On the other hand, the renewable production is hard to predict, there will be higher peaks that need to be supplied by costly peak generators and to times when there is more supply than demand. Due to bad predictable circumstances, there is more risk for energy retailers, while the costs increase. Without resolving these issues, they lead to increased social costs (i.e., a higher bill for customers). These issues will be addressed within e- balance by balancing the grid. Whereas in the old situation the (flexible) production is adapted to match the demand, the current trend is to move to a situation wherein the demand is adapted to the (inflexible) production; this is called demandside management (DSM). Smart appliances are used that allow for such flexibility to balance the grid, e.g., EVs of which the amount and time of charging can be adopted by an ICT system to charge off-peak, or whitegoods that can be started such that their operation coincides with PV production. Page 12

13 To match supply and demand, a few challenges need to be overcome. Time-of-use pricing, dynamic pricing and similar tariffs are often proposed as possible solution to this problem. However, recent research shows that these approaches lead to higher peaks due to load synchronization and may at best only move the load peak [1]. As an alternative, capacity tariff is proposed [2], which means that customers are billed based on the maximum peak load. However, capacity tariffs may result in additional costs when there is no problem, e.g., when production and consumption match at a neighbourhood level. e-balance goes beyond capacity tariffs and related approaches by matching production and consumption at each hierarchical level. More precisely, the e-balance methodology aims to first balance the house loads, then it balances loads at a transformer level, then at neighbourhood level, etc. To accomplish this we needed to tackle several challenges for the control methodology and many details on how to address the challenges can be found in our deliverables and scientific publications that are indicated: Hierarchical nature: the system architecture should be hierarchical in nature and the algorithms need to function at each level of the hierarchy [3]. Cost effectiveness: there must be a relation between the costs and the optimization objective. Since the losses are quadratic within the current, this must be reflected in the objective that is collectively obtained [1]. Predictions: to match supply and demand properly, flexibility needs to be reserved for when it is needed the most. For this, predictions of supply and demand are needed. Prediction models are modelled with artificial neural networks (ANN) and are further explained in D5.2. Efficient algorithms: the e-balance algorithms are executed on small devices that are installed in houses and have low computational power. Because of this the algorithms need to be designed to be very efficient [4, 5]. Before testing the algorithms in the field, we combined simulations with load flow calculations to show that our algorithms are grid friendly, i.e., lead to less losses, reduce max. cable load, better power quality, longer lifetimes [6], etc. See D5.2 and [1] for details. The core idea behind our demand-side management methodology, called Triana, is explained in Section The Energy Balancing Algorithms. Page 13

14 To facilitate the methodology and enable customer support services (e.g. data access) it is crucial to design an ICT system with this in mind. This ICT infrastructure offers secure and abstract services that support the hierarchical system and provide access to the required grid components. The main challenges of developing such a system are: Privacy: data from customers are used and exchanged. Due to the sensitive nature of this data, access must be restricted based on stakeholders (Section ). Security: all data exchanged are sensitive and need to be kept confidential, in addition data integrity and authentication of its origin need to be ensured. Data is exchanged between different security zones and security measures such as firewalls and proxies need to be installed. Abstraction: the system supports services (balancing, data exchange, etc.) that require data exchange between different levels at the hierarchy. Abstract interfaces are needed to aid the implementation of such a system. More specifically, we need a middleware layer that offers locational transparency, data security, encryption, etc. Page 14

15 The energy management platform is the first application of the e-balance hierarchical architecture. It aims at balancing the electric energy of neighbourhoods with flexible loads as smart-appliances (appliances whose cycles can be rescheduled), PV power inverters (with the capability of power curtailment) and smart-storage devices (e.g. electric vehicles or electric batteries with control and management features). As mentioned before, the energy management platform has to deal with the technical and market challenges that are still issues to integrate demand response and demand-side management services in current electric grids. For instance, the build-up of appliances manufacturers offering smart solutions that support these applications demonstrates an adequate level of maturity. However, the lack of standard and open platforms to connect and integrate technologies from different vendors and the lack of a business model to convince both electric market players and consumers lead to multiple initiatives that are back where they started that have to rethink their approach from scratch and manufacturers are developing independent management services, which force end-users to install and use disaggregated solutions. In order to defeat the described challenges and barriers, the e-balance energy management platform has been developed as all-in-one ICT solution to integrate technologies from different vendors, considering business and social requirements, which allows an effective energy balancing framework where all stakeholders from prosumers to DSOs can perceive some benefit. This is composed of a communication platform that fulfils with the requirements of the Smart Grid Architecture Model proposed by CEN-CENELEC from the user-level to the DSO s control centre, considering privacy and security protocols. Users (prosumers) can interact with the platform through a specific GUI designed to satisfy their needs and raise their awareness on power management, which allows scheduling their appliances and knowing the improvements obtained through the balancing processes. Then, the system selects the most suitable planning for their home appliances to satisfy DSO s requirements (mainly flattening the energy profile and reducing power peaks). On the other hand, DSOs are able to quantify the peak reduction obtained, which reduces management and operational costs of electric grids. Page 15

16 2.2.1 THE ENERGY BALANCING ALGORITHMS To fulfil the aim of matching supply and demand on multiple levels, a coordination of loads is needed. Such coordination takes place in time by reserving flexibility for when it is needed the most. For example, when a high PV peak is expected battery charging may be postponed when PV production is low to reserve capacity for when the peak arrives. Such trade-offs require predictions that give profiles, i.e., the house load over time. Also coordination in space is needed, houses collaborate to adjust their load profiles such that supply and demand match and peak loads are mitigated. This coordination is based on planning and mathematical optimization. Based on the planning, devices are controlled in real-time, this step is called real-time control. This leads a methodology called Triana, which consists of the three steps described above: 1. Prediction of the consumption and production of energy 2. Planning when each appliance is used 3. Control of the appliances to achieve the planned behaviour Within e-balance we have developed prediction algorithms using artificial neural networks, planning algorithms using dynamic programming (see deliverable D5.2) and used FPAI to control the devices in the real-time control step (see deliverable D5.5). The energy balancing algorithms The energy balancing algorithms are based on the Triana approach and require the coordination between the low-voltage grid management unit (LVGMU) located within the secondary substation and every customer level management unit (CMU), which is connected to the smart-meter on the prosumers premises. Therefore, the algorithm is split and executed into both management units. The algorithm allows two different ways of balancing, selected by prosumers: local balancing to maximise energy generation (in Bronsbergen demo the PV panels) and neighbourhood balancing to flatten the energy profile, which has been one of the main goals in e-balance. Both options require accurate prediction provide in our case by neural networks algorithms, the former at the household level (poor accuracy due to the huge amount of uncertainties at home level) and the latter at the neighbourhood level (good quality predictions). Aggregation of energy demand prediction at the household level is not considered and recommended to apply the Triana approach unless new prediction methods or techniques improve the accuracy of predictions. The algorithms aim at fulfil the balancing objectives by rescheduling active smart-appliances. If users select local balancing the appliances will be scheduled preferably in solar peak production hours, while they select neighbourhood balancing the rescheduling process will avoid as much peak consumption hours. In addition, for the neighbourhood balancing users can select three different modes of cooperation with the DSO in order to schedule their smart-appliances: only reschedule is permitted, pausing when it is necessary and power curtailment is allowed. Page 16

17 All in all, the following figure shows how the algorithm works to schedule a smart washing machine. Energy balancing algorithm (rescheduling power profiles of smart-appliances) The algorithms have been implemented in Java and integrated in the e-balance middleware, which is described in following sections. The designed algorithm for neighbourhood optimisation requires the interaction between the CMUs and LVGMU and is composed of the following steps: 1. Get predictions (at LVGMU): prediction algorithm, based on neural networks, which provides 24- hour predictions of neighbourhood energy demand in a 96-value array format. 2. Calculation of priority factors (at LVGMU): this step normalises the aggregated energy profile with the prediction values obtained in step Calculation of available flexibility (at CMU): in this step the best starting point for every smartappliance available (scheduled) is calculated and sent back to the LVGMU. 4. Ranking of CMUs (at LVGMU): this step classifies CMUs according to flexibility. 5. Optimisation (at CMU): according to the order of the previous step, every CMU is requested sequentially to optimise the schedule of its smart-appliances and the output is the new energy profile, which will be used to update the aggregated energy profile. Page 17

18 2.2.2 INTEGRATING SYSTEMS AND SOLUTIONS In order to achieve the defined goals several aspects have to be addressed by the system and several issues need to be solved. Energy management algorithms, like, e.g. the presented balancing algorithms work well under two major conditions. The first is that the users are willing to offer flexibility to the system. This is a social aspect, which is very much related to the second and technical condition that users use devices, which allows them to offer the flexibility to the system. In other words the system is only able to work if these two conditions are satisfied, so users want to offer flexibility and they also have means to do so. Assuming the users are interested in using the e-balance system they need to use devices that allow the system to monitor and control these. There are different levels of control and monitoring provided by the smart devices, as well as different approaches to access the control interface of the device. The devices may be controllable from the local network, but may also require cloud connectivity. In order to support all available smart appliances the approach followed by the e-balance project is to access the devices via the API offered by the device vendor. Software device drivers specific for the smart device brand allow the e-balance system to access the API device and to include the supported devices in the control loop. The intention of the e-balance project was not to invent and develop new smart home appliances, but to provide a generic way to support these available on the market. What is new and important is that the e- balance system allows coordinated control for smart appliances of different vendors, what is currently often not supported. The appliances in each household are controlled by the household s respective CMU device. These devices have to support a diversity of possible network configurations at users homes. In general the CMU is installed behind the router so it is hidden from accesses from outside of the user s network. On one hand this is positive, because the CMU is then protected against most of possible attacks that require direct connectivity. But on the other hand, this set-up complicates the communication between the CMU and its parent LVGMU. This has to be realized using a cloud-like intermediate server. This server has to help the different management units to exchange their data (using the developed middleware), but it also helps to realize maintenance of these units, by distributing the software updates. It also supports the users in accessing their local CMUs by maintaining the local addresses of the CMUs so that the in-house access to each CMU is simplified. Similar, the LVGMUs have to be installed according to the rules defined by each DSO (or other stakeholder responsible for these units). These rules define the security and other requirements that can be very deployment specific. There are several options possible that influence the combination of the used e- balance system components and their configuration. For instance, the communication between the LVGMU and CMU can be realized in the Internet domain, but it may also be realized using DSO own data network parallel to the energy grid (e.g., using Power Line Communication), making the communication within the e-balance system independent from the user s Internet access. This section describes the means that have been developed for the e-balance system in order to make use of the available flexibility and it explains how the individual components are composed together to create the entire e-balance solution. It focuses on our Bronsbergen demonstrator deployment. Page 18

19 ICT Architecture Our concepts thrive when a lot of controllable connected devices communicate in a standardised way with the energy management system. The CMU should however be able to cope with this potentially large amount of devices. From the energy management point of view, only two physical connections were required in our demonstrator: the connection with the smart meter and a wireless LAN router that provides internet access. In the worst case, each amount of flexibility providing device needs to be connected to the CMU separately through the WLAN router or via a cabled solutions, for example, nearby sensors. In our case, the devices were connected to the participant s WLAN router, connecting to cloud services provided by Whirlpool and Mastervolt. In turn, our CMU connected to these clouds. The benefit this provides is that the number of connections to the CMU is limited and the number of APIs active can be minimised. This allows for a great many of devices being connected through cloud services, allowing for standard mini-pc hardware to run the CMU functionality. Direct communication between the CMU and LVGMU may only occur if it is initiated from the LVGMU due to Alliander cyber security requirements. This is however not in line with the required type of information exchange between the management units of the e-balance energy management platform. To overcome this, a change in communication is introduced into the set-up: asynchronous communication through a CMU-server, a proxy server that needs to be polled for information. A similar zoned approach is also deployed within the Alliander ICT infrastructure, allowing the LVGMU to be placed, in theory, at the secondary substation, interacting with sensors locally, while still being able to communicate with CMUs that communicate via the internet. However, by using this system, it is no longer important from a security point of view in which DSO security zone the LVGMU is placed. In case a e-balance architecture for energy balancing services measurement infrastructure is already in place, the measurement information can then just as easily be supplied to a virtualised LVGMU, limiting the devices in the secondary substation to measurement equipment only. In the Dutch case, smart meters communicate through integrated GPRS or CDMA modems. Hence, any methods on the LVGMU that require aggregation of smart meter data, needs to interface with the smart meter infrastructure, which is a virtual access point for the LVGMU. However, in case smart meters are communicating in a hierarchical manner through a local RF-mesh or PLC technology, it makes more sense to also have the LVGMU locally as demonstrated at the Batalha demonstrator. Page 19

20 Middleware Smart grids are complex scenarios where a large number of devices interact. The devices involved in the system range from small embedded devices such as the customer management unit in the households of the prosumers, to powerful servers (e.g. computers in the primary substation that the grid). Moreover, many different communication protocols are used such as low power wireless protocols in the wireless sensors deployed in the field and high-speed communication protocols such as TCP/IP over the internet. With such a diverse set of requirements, it is difficult for applications and developers to interact with each other. To overcome this issue and simplify the communication between devices a communication middleware has been developed in e-balance. A communication middleware is simply a software layer that runs on a device which is used to provide high level communication services to applications and developers regardless of the underlying platform and communication stack. The high-level communication services provided by the middleware not only simplifies the development of applications, but also establishes a common communication API that all applications need to use in order to exchange information with each other. This means, that the underlying communication stack and platform are transparent to the applications. In the e-balance project, the communication middleware is used to provide high-level communication services to the different management units in the system. It handles information exchange in the prosumer households (CMU) such as smart meter data and smart appliances configuration and schedule, in the different substations that control the grid and monitor/actuate in the households (LVGMU, MVGMU) and the control centres that have an overall picture of the status of the whole system (TLGMU). Deliverables D3.1 and D3.2 describe the general architecture of the middleware and the submodules it has. All applications running in each management unit use the same API, described in these deliverables, to exchange information regardless of the platform over which they run and the communication stack the use. The communication middleware in e-balance offers a platform-independent REST-based API that allows the devices to exchange and store information. Moreover, it provides a web-based GUI that is used to monitor the information exchanged, the security policies that applies to each piece of data and the configuration of the management unit. Implementation details about the e-balance middleware are described in deliverables D4.1, D4.2, D4.3 and D4.4. Page 20

21 Privacy and Security Technology The privacy and security means used in the e-balance system are realized as a two layer approach. First, the security (and cryptography) means are used to create secure and trusted infrastructure, where the middleware instances on individual management units exchange their data in a way that no external parties can interfere with this communication. Then, on top of that layer, privacy preserving protocol is executed where the individual system users define who can access their data in the system. The secure infrastructure is based on mechanisms located in the middleware but also on standard IT security means configured in the proper way on the management units (Linux machines). In the networking stack several mechanisms are used to protect the exchanged data against eavesdropping and modifications. The communication between the management units (and the buffering proxies) is protected by means of Transport Layer Security (TLS). The certificates of the individual management units (and buffering proxies) are used to generate session keys to encrypt the communication. Each management unit is identified by a unique identifier that is further confirmed by a certificate, signed by a trusted party the certification authority. The implementation of the certification authority is realized in Java and uses the X.509 certificates with ECC-256 signature using the Elliptic Curve Digital Signature Algorithm (ECDSA) with Secure Hash Algorithm SHA-1 hash function. Only authorized devices are allowed to join the secure group. And only communication with certified devices is allowed. In order to implement the privacy protocol each stakeholder is equipped with a user certificate that is used to identify the requests to access data. The services in the energy management platform on top of the middleware operate on behalf of a given stakeholder and each data access request is authenticated by a signature generated using the private key related to the user certificate. This way the access request sources are limited to the known system parties only. Implemented menu to specify middleware variable policies In order to define who indeed is authorized to access specific data stored in the middleware, the data owners specify policies for each of the variables (data items) stating which stakeholders and which services that run on behalf of these can access the specific variable data. The type of access is also specified, defining that the particular service of given stakeholder can read or write the owner s data stored in the variable. Additionally, delay between any two data accesses (and data timestamps) can be defined. This feature can be used to limit the temporal resolution of the obtained data sequence. Details on the security and privacy mechanisms can be found in D4.2 and D5.4 (available on the website). Page 21

22 Smart-appliances The e-balance users are able to schedule their smart-appliances in order to the balancing algorithm organises flatten the neighbourhood power profile. The experience obtained by managing data models and communication protocols between devices has facilitated the following guidelines and specifications that smart-appliances must have implemented to be compatible with new energy markets. Communication Unlocking flexibility in an optimal manner by means of DER production curtailment, advanced battery management and demand management requires the ability of involved devices to communicate with an energy management system. Technically, this energy management system can be part of a building s Building Management System, it can be stand-alone or part of an external (cloud) system. As with most communicating systems, either the devices should use the same language or the aggregating system should be able to speak many languages. From a usability and consumer point of view, standardisation of this communication is desired. Functionality and Control Closely related to communication is the ability of a device to teach the aggregating energy management system what kind of flexibility can be provided. For example, simply turning a washing machine cycle on or off is such a feature. However, a much more refined level of control is also possible, allowing for far more degrees of freedom in providing flexibility. It is therefore important that standardisation regarding unlocking flexibility facilitates both fine grained control as well as simple on/off control types of actions. At the same time, manufacturers of appliances suitable for demand side management need to be given the lead in safe-guarding the primary function of their devices. Users on the other hand should be provided with the ability to influence the applicable energy management system s control strategy in such a way that their desired level of comfort is maintained, while also educating them on the consequences and benefits of their choices. User interaction In the vision of the e-balance project, user interaction with the control system is limited to an initial or low frequency periodic configuration of the system. For example, new devices might require different control strategies that require user inputs or benefit or lifestyle changes merit a change of strategy. User interaction also includes data gathering and feedback however. The user should be provided with easy to understand information and control over their data. This includes control over privacy and security settings, allowing control over who has access to what data and for how long. In our vision this is not limited to commercial parties, but can also include actors in the local neighbourhood, allowing neighbours to compare notes. Smart meters and the energy market While an energy management system can be accompanied by local energy sensors, users connected to the energy grid already have a calibrated energy exchange measurement system in place: the smart meter. By integrating (local or market) access to the smart meter data, reliable data can be used for both building energy management as well as neighbourhood, regional or national energy management systems. Page 22

23 2.2.3 HOW TO DEPLOY THE SOLUTIONS Although the grid operator part of a metering cabinet is standardised, what the participant does with the remaining space is not. Some metering cabinets are in use as storage cabinets, others have all kinds of (telecom) electronics and associated wires installed. Hence, every installation will be slightly different. Figure 9 gives a nice impression of the installation of a CMU in a clean metering cabinet. Standard metering cabinets are not supplied with wall sockets in the Netherlands. Hence, care should be taken that solutions requiring power will also need the aid of a technician to install a wall socket. The numbers indicate the following components: 1. CMU 2. Power adapter for the CMU 3. USB hub, facilitating the connection of the smart meter P1 serial-to-usb converter cable 4. Power / Wall socket 5. P1 serial-to-usb converter cable 6. Smart meter 7. P1 port of this particular brand smart meter 8. Not shown: WLAN adapter attached to the USB hub Typical installation of a CMU in a participant metering cabinet, with component identification Deployment of household s elements to integrate balancing services As indicated in the ICT architecture section of this Guide Book, the physical connection of the CMU is only a part of the solution. When running a project like e-balance with parts of the demonstrator located at private property, the availability of flexibility providing devices should be ensured. This is not only supplying the devices to the participants, but also making sure that the configuration is correct and stays correct. In case the customer s WLAN network is used, as we did, you run the risk of having to repeatedly visit participants to reconfigure all devices in case of WLAN credential changes. We deployed Whirlpool washing machines and dryers and made use of previously installed Mastervolt PV inverters at the Bronsbergen demonstrator. The roll-out of the CMU and these devices and installation of (additional) wall sockets was performed by one and the same installation company. As a consequence, participants are not confronted with a lot of different people in their house and the project remains manageable. Finally, in our case, in order to provide flexibility with the Whirlpool devices, a smartphone app should be installed on a customer s device. Page 23

24 2.2.4 BRONSBERGEN EXPERIENCE The focus of the Bronsbergen demonstrator deployment was on energy balancing and security and privacy. The energy balancing algorithm used by the energy management platform service has proven the applicability of the e-balance platform (communication platform) and implements the related use cases as defined. An extensive evaluation of the energy balancing results was limited due to social factors that reduced the flexibility available in the Bronsbergen demonstrator deployment. Future projects could elaborate on that combination, working further on improved balancing algorithms, but as well as on improved social mechanisms related to the user involvement. Nevertheless, though the available flexibility shown during demonstration has not allowed us to test the entire potential of the balancing algorithms, simulation tests performed have been very promising to flatten the energy profile at neighbourhood or district level. This effect increases as users deploy more and more smart-appliances that are able to be rescheduled (flexible loads) and controlled. Therefore, the applicability of the developed platform and the proposed balancing algorithms can be extended to further projects related to energy management in different types of buildings with flexible loads (e.g. storage systems and smart-appliances). Performance of prediction algorithm (above) and effect of the rescheduling algorithm for a power profile (below) The communication platform provides the means for user data security and user privacy. The e-balance platform allows the user to control the access to the user data. However, the experience showed that this area is still a taboo for most of the users and thus, the presentation of the user interface to the implementation of this functionality should be realized more user friendly. A simpler GUI, but also a more compact way to define the security and privacy policies could be developed in further projects. The e-balance platform can be used in future grids as a base for energy management services. The implemented energy management platform for the Bronsbergen deployment including the proposed energy balancing algorithm has shown that the data exchange is working properly. The users have control over their data and the energy balancing service works. The available flexibility provided by the system users is managed properly and according to the user preferences to achieve the defined goal flattening the energy exchange profile of the household or the complete demosite. For broader application of the system it is necessary to work on the prototype to make it more mature with respect to the user interface (more user friendly GUI) as well as the set of available energy management services. Increasing the set of supported smart appliances would also be of advantage. Page 24

25 2.2.5 VISION FOR FUTURE DEVELOPMENTS OF ENERGY MANAGEMENT SYSTEMS: TRANSFORMING ACADEMIC SIMULATIONS INTO VIRTUAL ADVISORS As part of the Graphical User Interface, we proposed, though not implemented, a virtual advisor as part of the CMU or services accessible by the CMU. This virtual advisor can combine the energy exchange information gathered by the CMU and can give advice on the best energy management strategy or make suggestions with respect to the best appliances or storage devices to enhance the value of the system. This paragraph describes the analysis that such a virtual advisor can perform. In a house with local generation by PV-panels and battery storage, the e-balance system can optimize the demand such that the locally generated energy is also used locally, and the interaction with the grid is minimal. The power that is produced by the PV-panels during the day can be stored in the battery, and used during the evening/night. Figure below gives an overview of the possible energy flows within the house. Schematic picture of the energy flows within a house with local production and storage In the extreme scenario that a grid failure occurs, and the house is disconnected from the grid, the e-balance system can use the local storage and production to operate the house in an islanded mode. The duration of a grid failure is not known in advance, and may take some time from a couple of minutes to several hours. Whether the house can survive the grid failure on its own depends on: Duration of the failure; a longer failure will be harder to survive. Time of day of the failure; at night the PV-panels will not produce any energy, and all power must be supplied by the battery. Time of year time of year; in winter the PV-panels produce little extra energy that may be stored for night time. Reserved backup capacity of the battery; you can always keep the battery partially charged, specifically for the use to survive a grid failure. This will limit the balancing capabilities of the system, since less battery capacity is available for balancing. We have developed a model to analyse the probability the house will survive the grid failure, given a distribution of the random repair time. The model can also be used to give an advice on what battery capacity is needed for a household to survive most grid failures. Of course, when the battery is used for backup purposes it cannot be fully used for balancing and self-use. There is a clear trade-off between using the battery for earning money by providing balancing capabilities and using your own generated energy on the one side, and using the battery as backup energy system to provide energy in case of a grid failure on the other side. Page 25

26 With smart battery management strategies that adapt the backup level to the expected needs one can get the bests of both worlds. Figure below provides the user an indication of what battery size is needed per 1000 kwh of yearly demand to obtain a certain combination of self-use and survivability, for a household that has a yearly PV-production that matches the yearly demand. Self-use rate as function of the installed battery capacity for different survivability levels The self-use rate indicates what percentage of the locally generated energy is also used locally, either direct or via the battery. The survivability level indicates the chance that the energy supply is not interrupted when a grid failure occurs. For example, in case you want to have a 99% chance to not notice a grid failure, and have a self-use rate of 50%, you will need a battery of approximately 930 Wh per every 1000 kwh yearly demand. Thus, a household with a yearly demand of 3500 kwh, would need a battery of = 3255 Wh battery. (Note that, this capacity is the usable capacity. Often the batteries in home-energy systems are not used for their full capacity to prolong the battery life). It is important to know what the impact of the system is on your energy bill. In this computation we used the following assumptions based on [7]. We assume a system lifetime of 20 year. The feed-in tariff is set to 0.12 /kwh. This is the current feed-in tariff for Germany. We assume this tariff is kept for the full period. The mean grid electricity price is set to 0.34 /kwh. This is based on the current price, including taxes, of 0.28 /kwh, and the assumption of a 2% price increase per year. Note that, in this cost analysis we compare the different levels of survivability for an installed system. Thus, this does not include the investment costs to install the system. Figure below shows the costs for 1000 kwh (=1MWh) demand as a function of the battery capacity for various survivability levels. Costs for 1 MWh/y electricity as function of the installed battery capacity for different survivability levels. and more results can be found in the scientific papers [8], [9] and [10]. Page 26 When we take the scenario of the previous example, for the 930 Wh battery, and a 99% survivability level, the cost of every 1000 kwh on the energy bill every year will be around 110 euro. When the system would be used for self-use only (0% survivability) every 1000 kwh would cost 85 euro. So, in this scenario making sure you will survive a grid failure with a chance of 99% will cost you 35 euro per 1000 kwh demand per year. Our model can be customized to specific customer scenarios by changes the input demand and production profiles. Thus, we can provide specific numbers to provide the necessary insights for a customer how to set-up his home energy system. Details on the model

27

28 Presently, current ICT platforms available allow implementing better electric energy grids, meaning that they are better suited to face operational contingencies and fault occurrences, while providing a high service quality. The answer to this challenge is the smart grid, as it copes with those grid limitations, providing the grid system operator with resilience features towards reducing the likelihood of outages occurring, besides mitigating their duration and impact on people, businesses and industry. Moreover, the smart grid aims also at being operated at minimal cost, thus implementing energy efficiency mechanisms through the implementation of advanced features. Current smart grid technology goes as far as the MV grid goes, meaning that LV grids are normally excluded from participating in grid resilience features. Going beyond the state of the art, e-balance addresses both grid resilience and energy efficiency, through a set of new deployed features towards the LV and the MV grids. Regarding the LV grid, those features comprise further automation and monitoring solutions downstream the secondary substation, the typical last mile frontier of the smart grid. Within e-balance, RF Mesh communication technology is embedded in LV sensors towards deploying monitoring features serving the distribution grid and public lighting circuits. Moreover, the concept of the LV Grid Management Unit is proposed, merging the role of smart metering data concentration with the role of controller, thus enabling features as fault detection and location, power flows and losses calculation, assessment of key performance indicators, among others. Regarding the MV grid, those features comprise the calculation of optimal setpoints towards controlling switching primary equipment, capacitor banks and transformer tap changers, towards improving grid performance through mitigating power losses. Moreover, the concept of autonomous self-healing was introduced, serving adjacent primary substations towards grid resilience improvement and outage time mitigation. Page 28

29 Different stakeholders different perspectives mutual goals The end customer perspective: high quality of service, availability The regulator perspective: assess and incentivise the performance of the grid system operator The grid system operator perspective: differ investments, reduce OPEX, improve system reliability and provide high-level service to end customers The industry and academia perspective: provide applied knowledge and research towards providing cost-effective reliable technology e-balance goals grid resilience energy efficiency Development of a distributed architecture combining different grid level devices, from the higher hierarchical level Top Level Grid Management Unit, the primary substation level MV Grid Management Unit, the secondary substation level LV Grid Management Unit, the LV Sensor, the Smart Meter, among others. Several use cases were defined to address grid resilience at both MV and LV grid levels LV grid level Neighbourhood and RES power flows, losses calculation, fault detection and location, fault prevention, KPI calculation, fraud detection, energy management combined with a grid status traffic light. MV grid level Fault detection, isolation and service restoration Some use cases were defined to specifically address energy efficiency at MV grid level MV grid level Optimal switching of primary equipment, transformer tap changers and capacitor bank taps, coping with operational constraints and safety criteria Development of solutions aligned with the described use cases Deployment of a solution for the LV grid in two demonstrators in the region of Batalha Deployment of a solution for the MV grid in an offline demonstrator Page 29

30 Proposed solutions RF Mesh enabled LV sensors A development by INOV, comprising 3 different components: Mesh Gateway, LV sensors and PL Sensors. All components obey to the IEEE standard. The frequency of 868 MHz ISM band is chosen due to its lower attenuation in open space, when compared with the 2.4 GHz ISM band. For communication at the application level, the DLMS (Device Language Message Specification) protocol is used, supporting a number of OBIS (Object Identification System) codes. Page 30 The Mesh gateway main features are to allow reliable wireless communications with the LV and PL Sensors, using IPv6 protocol and RPL for the routing protocol. The Mesh gateway also communicates with LVGMU through an Ethernet interface, to receive measurement commands and report alarms received from the sensors.

31 The main features of the LV sensors are to perform measurements of LV lines, namely 3 phase voltage, current and power factor. They also are able to detect voltage and current faults and report them back to the Gateway. The LV sensors have three voltage inputs directly connected to the each phase of the grid node. They also have three current inputs connected via current transformers selected to cope with the maximum current magnitude. The main features of the PV sensors are to perform measurements of public lighting circuits, namely voltage, current and power factor, as well as report voltage and current fault of the public lighting circuit. For deployment and cost reasons in some deployments the PV and PL sensors were combined into a single box sharing the same radio module, creating a new PV+PL sensor. 3-phase sensor node (INOV) LV Grid Management Unit LVGMU A development by Efacec, made on top of its G Smart product, a Distribution Transformer Controller suitable for secondary substation automation. The main features developed include grid fault detection and location, public lighting s light bulbs failure detection and location as well as fault prevention, fraud detection and KPI calculation. A development by Efacec implemented on an embedded PC, as a complementary device to enhance G-Smart s functionalities. The main features developed include neighbourhood and RES power flows, as well as losses calculations for unbalanced three-phase grids. The LVGMU aims at being deployed at secondary substations operated by the DSO. Advantages of using a device like the LVGMU It plays the role of an RTU regarding the management systems places in the upper levels, as the MVGMU or the TLGMU. It participates in resilience mechanisms regarding the LV grid overall state and measurements awareness. Limitations of the LVGMU It deeply relies on smart metering infrastructures where their communication performance may affect the overall LVGMU behaviour. Depending on the regulatory framework, the LVGMU may not be able to participate in the curtailment of active power being injected by PV inverters, performing voltage regulation in response to their potential impact on grid voltage profiles. Sensor Gateway (INOV) G-Smart (Efacec) Embedded PC (Efacec) Page 31

32 3.2.1 HOW TO DEPLOY SOLUTIONS Three-phase LV grid sensors are installed in secondary substations, monitoring LV feeders immediately downstream the LV bus-bar. Moreover, this kind of sensors can also be installed in overhead circuits, serving selected grid points, namely those where the grid splits into different grid branches. Such LV sensors have three voltage inputs directly connected to the each phase of the grid node. Moreover, they also have three current inputs connected via current transformers selected to cope with the maximum current magnitude, as well as with over-currents. Where suitable, the LV sensors may also have a fourth pair of current and voltage inputs, meant to be used in public lighting circuits. Those sensors communicate with a Gateway installed at the secondary substation via the RF Mesh communications network, bridging the LV sensors with the LVGMU. The Gateway and the LVGMU communicate via an RJ45 Ethernet link. The application protocol is the DLMS/COSEM with IP routing. Besides LV Sensors, e-balance also foresees the use of 3-phase smart meters strategically placed in selected grid nodes, aiming at providing fault detection and location of potential frauds. The LVGMU is based on a combination of the G Smart a Distribution transformer Controller, or DTC, by Efacec, and an embedded PC, which enhances the computing capacity of the G Smart. The LVGMU provides a web-based GUI, which can be remotely reached, by a duly user. G Smart (distribution transformer controller), a LVGMU component Gateway for the wireless sensor nodes Embedded PC, a LVGMU component M Box I300, GPRS enabled 3-phase smart meter Wireless Sensor, for 3-phase LV feeder and public lighting circuit The deployment was carried out by EDP, where skilled staff has made all devices installation without affecting the daily life of the residential and small business end users. Page 32 The deployed system was fully tested, also with the intervention of EDP staff.

33 3.2.2 BATALHA EXPERIENCE The deployment of LV solutions in Batalha has demonstrated the following impact of results for future grid deployment: On one hand, the use of LV sensors brings further awareness of the LV grid state but their deployment incurs on a significant increase of investment by the LV distribution grid operator that must be evaluated deeply. The deployment of LV sensors inside secondary substation, serving all 3-phase LV feeders, should be coupled with the deployment of a smart-metering infrastructure and service quality and voltage curve data upon request. The combination of automation and monitoring features together with smart metering concentration services can be commercially explored towards improving grid resilience services. The outcome of alarm provision, fault detection and location, as well as fault prevention alarm are very well suited for outage and crew management, therefore with potential for addressing other corporate systems such as Outage Management or Maintenance Management. The proposed method for fraud detection has improved the accuracy for locating grid segments where frauds might occur. The proposed method for blown-up light bulbs within lighting circuits can be used either by LV grid operators or by municipalities, benefiting from early detection of contingencies, serving public service and enabling maintenance crews to respond. The power flows and losses calculations bring further insight of the LV grid, thus allowing the planning of engineers to better understand phase unbalances towards improving the grid performance. The implemented features initially thought for the low voltage management unit (LVGMU) can be applied in current Advanced Distribution Management Systems (ADMS), which corresponds to a LV branch of the top level grid management unit (TLGMU), thus inheriting important benefits from the developed features. Page 33

34 Proposed solutions MV Grid Management Unit MVGMU A development by Efacec, made on top of its SCADA/DMS solution (the SCATE X+ system) with the purpose of implementing self-healing features towards improving the distribution grid area comprised by adjacent primary substations. The solution addresses fault current indication data coming from fault detectors, as well as fault data coming from protection relays placed at primary substation level, and also from controllers coupled to MV grid reclosers. The solution also considers topology data regarding the MV grid area under the scope of self-healing. It may comprise several feeders from one or more primary substations. In the circumstance of the demonstrator, the MV grid comprises two feeders from the Batalha primary substation and another feeder from the adjacent Azoia primary substation. The MVGMU aims at being deployed at one primary substation operated by the DSO, integrating IED data from such primary substation and from other adjacent ones. Advantages of using a solution like the MVGMU. It provides self-healing features for a Distribution Grid Area (DGA) served by multiple feeders from one or more primary substations. The MVGMU may perform the self-healing features in automatic mode or in advisory mode. In automatic mode, once a fault is detected and the protection devices trigger trip orders to protect the MV circuits affecting a large number of customers, the self-healing feature responds automatically by identifying the faulty zone. Once identified, the self-healing feature automatically provides switching orders, which are sent to the remote terminal units and other controllers. These orders allow to performing an automatic isolation of the faulty area. Moreover, they allow carrying on with the service restoration, performed upstream and downstream the isolated faulty area. Page 34

35 In advisory mode, once a fault is detected and the protection devices trigger trip orders to protect the MV circuits affecting a large number of customers, the self-healing feature responds automatically by identifying the faulty zone. Once identified, the self-healing feature automatically provides switching orders, which are notified to the upper hierarchical level, namely at the TLGMU. At this higher level, a duly operator may validate and use, or even change such switching orders. Once concluded the validation process, from the TLGMU the operator may perform the final list of switching orders aiming at isolating the faulty area and to restoring service at the maximum extent possible, performed upstream and downstream the isolated faulty area. The performed self-healing feature is strongly linked to the restrict set of remote terminal units and other controllers that drive the switching actuators over the MV grid. As the DGA is confined to few feeders, the response time of the needed data acquisition and control orders is very fast and less prone to errors, therefore the algorithm performing very quickly if set on automatic mode. Limitations of the MVGMU As it is suitable for performing in automatic mode (its major advantage), this approach normally goes far beyond the traditional way DSO operate their own MV grid. Therefore, the deployment of a set of features as those proposed within the MVGMU need a very robust automation system, which strongly relies on accurate and high reliable ICT infrastructures. Top Level Grid Management Unit TLGMU A development by Efacec, made on top of its SCADA/DMS solution the SCATE X+ system with the purpose of implementing advanced voltage / reactive power control and energy efficiency features towards improving a wide distribution grid. The solution addresses data coming from all kinds of automation and control devices deployed along the MV grid. The solution also considers topology data regarding the MV grid area under the scope of grid efficiency improvement. It may comprise a significant amount of primary substations and their MV feeders, downwards to the corresponding secondary substations. In the circumstance of the demonstrator, the MV grid comprises several primary substations, comprising Batalha s and Azoia s, besides others. Page 35

36 The TLGMU aims at being deployed at a Dispatching Control Centre operated by the DSO, integrating IED data from the primary substations and from remote terminal units and other controllers deployed along the MV grid. Advantages of using a solution like the TLGMU It provides grid performance improvement features for a wide grid served by multiple primary substations The TLGMU performs the grid improvement features upon a duly user request. The solution provides a list of switching orders that, once performed, will improve the performance of the distribution grid, namely by mitigating losses. The duly user may tune the objective function behind the application engine. Moreover such user may define the set of restrictions, namely which kind of operable equipment is to be considered switchgear and switching actuators, transformer tap changers and capacitor bank actuators, which specific grid assets should be excluded from remote operation, exclusion or inclusion of manually operated assets, among others. Once triggered, the optimised power flow feature automatically provides switching orders, which are validated by the duly user, although he might change such switching orders. The correct order of performing those switching orders is assured, thus preventing asset damages and human injuries occurrence. Once concluded the validation process, from the TLGMU the operator may perform the final list of switching orders aiming at improving the grid performance, namely by mitigating losses. Limitations of the MVGMU The performed optimized power flows feature is strongly linked to the extended set of remote terminal units and other controllers that drive the switching actuators over the MV grid. As the grid comprises a significant extension of controllable devices, the response time of the needed data acquisition and control orders may affect the overall feature performance, as the grid is more prone to communication errors. Initial grid topology Final grid topology Example of an initial grid state and a final grid state after an optimized power flow and validation of optimized solutions performed Page 36

37 3.3.1 HOW TO DEPLOY SOLUTIONS The MVGMU is installed on a computer server suitable for harsh environments within the selected primary substation. The TLGMU is installed on a computer server, which should be deployed in a specific room of the dispatching control centre. Both computer servers must be installed in a room with controlled temperature set to not exceed 25 C. Both must be equipped with a local workstation for HMI purposes, as well as a communications front-end to establish the correct data path towards the fault indicators (if applicable), primary substation controllers, reclosing controllers, as well as general purpose remote terminal units. Both the MVGMU and the TLGMU must rely on a secure ICT infrastructure and on ancillary services backed up by interruptible power supply sources. In the figure below it is presented an example of a control centre of EDP Distribuição in Oporto, serving the MV and HV distribution grid of the northern and central regions of Portugal. The SCADA/DMS system of the control centre was installed by Efacec, which is based on its SCATE X+ solution. The OPF/VOS feature, one of the e-balance s outcomes, is suitable for such kind of SCADA/DMS system, this one corresponding to the Top Level Grid Management Unit, as defined in the e-balance architecture. Control Centre of EDP Distribuição in Oporto, Portugal Page 37

38 3.3.2 BATALHA EXPERIENCE The deployment of MV solutions in Batalha has demonstrated the following impact of results for future grid deployment: The implemented algorithm improves the service quality, as it reduces the downtime of affected end users, upon the occurrence of a severe fault triggering protection schemes. Moreover, the algorithm allows reducing the number of reclosing actions and circuit breakers definitive protection lock out, by reducing the ageing of the MV equipment assets due to their use. As positive consequence, the algorithm avoids overburdening control centre operators, as they can be kept dealing with other tasks while the selfhealing mechanisms perform automatically improving grid resilience upon fault occurrence. On one hand, the deployment of solutions as those designed and developed within e-balance targeted for the MVGMU, thus serving primary substations, is perfect to address very quickly fault contingencies in specific distribution grid areas. On another hand, the deployment of such solution spread along several distribution grid areas served by different primary substations may go through a significant investment. However, in order to deploy the MV solutions conceived within e-balance it is required a significant number of grid assets fully equipped with protection enabled circuit breakers, MV grid remotely operated switches and even reclosers, all combined and serving a MV grid area big enough to increase the ratio between self-healing flexibility versus investment. Such balance will be easier to perform if the served distribution grid area is heavily prone to faults. The self-healing features proposed can also be leveraged in terms of their fit, as they can be included at TLGMU level. This approach is more computing demanding, as the algorithms would address the complete MV grid and would also need to handle a more complex data communication infrastructure, therefore more prone to delays. Finally, the major outcome of the combined OPF/VOS features is the real mitigation of power losses, which is a valuable contribution towards energy efficiency. Any new grid topology deployment after executing the OPF/VOS is performed considering all grid assets and their restrictions, thus coping with asset and human security. As the correct order to perform the switching manoeuvres is performed automatically, such feature raises the confidence of the MV grid operator that no error while performing the manoeuvres will occur, thus avoiding any unstable or prone to failure grid condition. Moreover, the automatic provision of the switching orders sequence releases the control centre operators from identifying themselves the correct switching order and defining their execution list. Page 38

39

40 What do we propose as the e-balance platform business model? We propose an Energy Exchange Management which supports the interaction of the end user with the market and the grid. The e-balance platform provides two services: local balancing for neighbourhoods and grid monitoring, congestion management and resilience. What do we exactly offer? A communication platform and a set of tools which support information management throughout the grid system that connects its individual participants, mainly prosumers. Every participant/prosumer receives full knowledge on economic and technical information supporting the energy production and consumption in their neighbourhood. Prosumers participate only when they choose it for their comfort. The e-balance platform facilitates balancing and aggregation of energy flexibility by aggregators under current market rules. A new market actor like the aggregator s role is proposed as the supplier s functionality (the integrated model). The aggregators and DSO enable the system to manage the flexibility under demand response, which improves energy efficiency for the grid and enables the prosumers to obtain additional economic benefits. Why do we need it? The way in which energy is being used nowadays is changing all over the world. An increase in the number of renewable energy sources, a change in the demand profile, energy generation to meet individuals own needs, and striving to increase self-sufficiency, lead to the need for improving energy flexibility and application of active demand response (ADR) management mechanisms. Who is this offer addressed to? The platform and its full functionality are dedicated for the whole power grid from its lowest level like home users and their smart energy devices and appliances to the Day Ahead, Intraday, Balancing/reserves markets. It is directly addressed to prosumers defined in the system as end users. We have selected list of other stakeholders who can be also involved and interested in such system: electricity producers, Energy retailers, aggregators, DSO/TSO, Independent Service Providers (it can be ESCOs, weather forecast agencies or data providers etc.), Regulators, IED providers and manufacturers, Governments (national and local level). Page 40

41 Balancing business model Assumption: the flexibility market should exist in a country where the e-balance system will be rolled out. It means there is a need for such a service provided by prosumers to the grid. Depending on the aggregator s integration model - the prosumers have an appropriate contract with the supplier/aggregator for flexibility products delivery. This system proposes two flexibility products to the market provided by prosumers: Direct Energy Flexibility (DEF) a flex product dedicated to balancing market and for auxiliary services. Requests for aggregation can be initiated due to the lack of balance and appropriate bids are resulting from the balancing market and reserves market. Indirect Energy Flexibility (IEF) a flex product dedicated to optimize portfolio of a retailer/supplier/brp and provide a precise schedule of energy purchases by them on the DAM/IDM market or for the purpose of the DSO, as efficient management of the grid operations flexibility. The prosumer has economic benefits (see analysis and results) when he answers for the systems proposition to adjust his use of energy (shifting of use or limitation/saving). Such adjustment can be provided in every 15 minutes or faster. It depends on the aggregator s reactivity for the market signals Depending on this, the aggregator serves the aggregation service and can have different sources of benefit for himself and for the supplier the e-balance system owner/deliver sets up the business condition how to join to the system. Page 41

42 The system provider/the aggregator should propose pricing method /proposition how the prosumers will be incentivised. We have proposed: A dynamic price/incentive for the dependent flex product provided by the aggregation service. It means the price is dependent on the flexibility volume provided during each day A flat price/incentive for the independent flex product provided by the aggregation service. It means the price is clearing during the bids on the demand supply at the flexibility market The contract with the prosumers has to give clear rules how this system will work and how much money the user can earn depends on his/her behavior with electricity use. The DSO has benefits mainly from the second service: the grid monitoring and resilience. These are as follows: fault location, location of the service interruption, better quality of energy service providing. Other stakeholders like the IED manufacturers or ICT and data providers will have a possibility to develop their business as well if the demand of the new tools/services will be increasing. Page 42 Advantages and Limitations The need of the flexibility markets is substantial in such countries like Poland, the Netherlands and Portugal (see business cases). The flexibility products can influence on the higher level of energy efficiency, and grid sustainability. The flexibility market should be developed and forced by regulation. The users/prosumers should find this system really useful and profitable for them. Possibility to try out the functionality is crucial for increasing the interest. Without these factors such system is highly risky for market introduction.

43 New functionalities and new players Prosumer - produces the energy from his/her own installation of PV and also consumes the energy from the grid. It means that he/she draws power from the grid or feeds its surplus into the grid, depending on the local conditions. Prosumers can get and use a home battery or an electric car battery to save some extra energy when it is cheaper and use it when the energy from the grid is expensive. Aggregator - Unlocks and maximizes the value of demand side flexibility obtained from the prosumers. Aggregators bundle numerous small flexibility resources into a useful flexibility volume. Under the cooperation of BRP and appropriate suppliers they provide product flex to the market and provide such services. They need to endorse a relevant contract with all main players. DSO - Is the main facilitator of the e-balance system. The DSO provides suitable infrastructure and access to the data. DSO can also requests the prosumer to become active to facilitate grid congestion management. Supplier - Is responsible for purchasing energy on the market and supplying it to the users according to the result generated by the balancing and optimization mechanism. This party also supports the aggregator in the processes of a settlement and remuneration payment to the users who are active in the e-balance platform. BRP - Can use the energy flexibility to optimise its own portfolio, and trade it on the market. What we have done during the business model analysis Overall concept of the business with SWOT analysis, added value analysis and theoretical pricing method proposition. Lean business canvas analysis (BCA) with business cases for Poland and the Netherlands with market strategy, competitors analysis, risk analysis for the balancing service. Mathematical model for pricing method, and calculation for theoretical data. Results analysis and sensitivity analysis. Page 43

44 Page 44

45 Using the e-balance platform pays off for all the results of the simulation studies Really, benefits to all, dear user! It may be hard to believe so soon, but it is your active behaviour, the User, that the e-balance platform can bring benefits to all its users. First and foremost, it means here about creation a so-called flexible energy product that can be used in the energy markets. From the point of view of the User, the most important participants in the e- balance platform are: the power supplier from whom you are buying it; the aggregator you will work with when using the platform; and of course you - Active User. In certain situations, the energy supplier may find it worthwhile to ask you to use certain features of the platform, as this will provides him with a flexible energy product in the form of your suitably adjusted demand pattern and thus gives him additional revenues or allows to reduce his operating costs. From these additional resources, the power supplier will only take a share - the rest will share with other parties using the platform. The supplier does not know how to work with the users to create such a flexible product. As a Small User, you can only change your electricity needs to a small extent, not too noticeable on wholesale energy markets where you trade multiple megawatt-hours. This is why you need to collect the larger group here, so that the "demand profile change" you collect together is greater. This is the task for the aggregator. Page 45

46 So the supplier will give some of the benefits obtained in the flexible energy product to the aggregator, which will take some of it, of course finally he must also earn - but the rest would be spent on charges for active load adjustment - so-called incentives paid to related users - to you. Perhaps you will say - ha, they will take most of the extra resources and they will only give me some miserable pennies. Certainly not, they must give you a large enough piece of cake, otherwise you will react too weakly, reluctantly change your load or to a lesser extent. In this way, they may have a greater part, but a much smaller cake. To get you into action they just have to offer you high enough incentives. Of course, they cannot offer you too much. Total payments for all active users cannot exceed the benefits from selling a flexible product. So the magnitude of the incentives will be set to obtain some balance, to divide the cake in such a way that everyone gets the largest piece possible. Both from the point of view of the size of your share and the size of the whole cake. But how your active participation in the e-balance platform can be used? You may be financially encouraged, Dear User, to continually shift your load to other hours. For example, do some energy-intensive activities such as washing, ironing, or heating the water, during certain hours of the day. Such a permanent modification of the load profile of the User can be used by the energy supplier for instance to optimize the purchase cost of energy. Energy prices on the wholesale market change every hour. The difference in energy prices for various hours of the day ranges from a few to several tens of euros per MWh, depending on the day (e.g. on the Dutch and Polish markets). So if you decide to permanently shift your energy demand, from the hour when energy is more expensive, to the hour when it is cheaper, and the similarly will do several thousand users working with the aggregator, it can obtain annual revenues from a reduction in energy purchase costs, at the level of several hundred thousand euros - to be shared among the participants of the e-balance project. You may also be encouraged, Dear User, to postpone a bit longer or resign from some energy-intensive activities. For example, an aggregator will ask you to lower your normal demand for a given value at a particular hour of specific day. If you agree with this, together with other users working with the aggregator, a significant reduction in demand will be achieved for the energy supplier. Page 46

47 Such a reduction of demand can be sold at a high profit, bidding various types of reserve auctions in shortage energy periods, due to the low capacity or the constraints of the network. This reduction can also be used by the supplier to improve his own situation on the so-called balancing market. If any supplier lacks the energy to cover the needs of its customers, it must buy it on the balancing market. Energy prices on the balancing market tend to be on similar level comparing normal energy prices. However, sometimes (a dozen - twenty days a year) they can jump to a very high level, exceeding a normal price of up to one hundred euros (and more) per MWh. In this situation, (by reducing the yours and other users demand), the supplier will reduce its energy shortage and will not have to buy it on the balancing market, which again can give significant amounts of savings to share between the participants in the e-balance project. But what might be the scale of your benefits? It depends on your activity, Dear User, and how profitable the vendor will be able to sell a flexible product. Even if we take into account only the basic business operations described above, related to the optimization of energy purchase costs and the reduction of the need to participate in the balancing market, the results of performed simulation studies show that: With a high level of activity of you and other users, the amounts you earn may average around 75 per year. With lesser level activity on you and other users part, you can earn an average of per year. Of course, these amounts may be increased by revenues from subsequent business activities using a flexible energy product on the e-balance platform. And as you can see, the greater your activity, the greater the income of both you and all other participants in the project. Page 47

48

49 While creating new technology you cannot ignore the end-user perspective. If technology is to be introduced into market and truly adopted it must fit into people everyday life, taking into account consumer habits and attitudes. The e- balance system redefines the usage of energy in households at many different levels. Nowadays, electricity is perceived by consumers as somewhat unspecified and intangible - for example, it is not known whether it is a product or a service. Consumers generally do not fully know how much power they are consuming (it is easier for them to say how much they are paying for it). It is an abstract, immaterial thing, something that users take for granted. This source is also subjectively inexhaustible (in the common sense you cannot exhaust the electricity from the grid you can only be cut out from the source due to network failure). In the new paradigm, introduced by Smart Grids, the electricity is: Something that we consume in our household and what by the very fact of production (and storage) is constantly in our attention (such as baked homemade cake, which we ourselves "create", and then we observe how it decreases in successive days). Something we can sell to the network. So there is a possibility for different strategies (because the prices on the market are dynamic and different from the seasons / day we can sell or store our energy in anticipation of a more attractive offer). It involves engaging and thinking about these strategies and in effect uses mental resources. Something we can optimise by giving the autonomy of our household to smart-home management programmes or third parties (for example, a power provider that will turn on our washing machine when we are at work, at a time that will be convenient to him). This redefinition has fundamental importance and has been one of the greatest challenges in conducting social researches. We evaluated the concept of a technological solution that was very far from the everyday experience of ordinary consumers of energy. We had to map the potential barriers and obstacles to a future adaptation solution that was exotic and intangible to potential users. Page 49

50 Quantitative research conducted by the project has shown us that the introduction of smart-grid technologies will certainly entail addressing barriers and dislikes of various natures. We have realized that the concerns of potential users with regard to intelligent home energy systems focus on issues such as: Price / profitability - will the whole system be profitable from the users point of view? Will the savings outweigh the cost of investment in the infrastructure and how big will these savings be? Privacy - Will the system monitor the activity of individual devices at home in realtime without interfering with privacy? Who and to what extent will have access to private data. Will they be anonymized? Cognitive overload nowadays consumers are overwhelmed by the number of messages. Advertisers, media, applications and electronic devices (tablets, smartphones, etc.) are constantly struggling for their attention. In this way, another home appliance that needs attention is an additional nuisance. It can be said that for similar devices users pay in two currencies: one is money, the other - commitment and attention. Very often the cost expressed in the latter currency is more burdensome for consumers than the former. Loss of control over your own home / appliances - who actually owns the washing machine if the decision to switch it is made somewhere on the power supply's servers? If it is possible to remotely control all the hardware in a smart home with a smartphone, who will guarantee that hacker won t do it? Page 50

51 Which factors would discourage you from installing a home energy management system? So how to address these barriers properly? How to plan the system to avoid frequent trap designing the ideal technical solution which is overthrown by consumers? While focusing on the technical layer you make a risk of losing the final customers from sight their habits, customs, character, lifestyles. We wanted to avoid such situation, that s why we created the 4 personas of potential users. The basis for the creation was the segmentation of energy consumers in Portugal, Poland and the Netherlands. During the exploratory workshops we described in detail 4 people - along with their (fictitious) name and all the details. Thanks to that, we could use the common language in the consortium. The term "end-user" from the abstract concept was personalized into unique human beings, with blood and bones. Page 51

52 The four personas John - 54 years old traditionalist, avoiding innovation. Mike - a young man fascinated by new technologies. His main (often unspoken) motivation is to impress others. Page 52

53 Patrick efficiency oriented father of the family. Owner of the photovoltaic panels who wants to use them well. Paul - avoiding complex technology, oriented on measurable profit. Page 53

54 The personas accompanied us throughout the entire technical process. Thanks to that, we knew what and who we were designing for. However, during prototype testing at demo-site Bronsbergen, it turned out that even when designing a system for clearly defined potential users, it was impossible to predict everything. Here are some insights and tips to keep in mind when creating a system similar to e-balance. Privacy issues should be very clearly described. It is worth noting, however, that recipients do not perceive these issues spontaneously as threatening. They almost never perceive it as a real issue, assuming by default that the security level is proper and high. These issues become threatening only when users become aware of it. A large number of consumers are aware of the role of privacy in the use of new technologies, and the fact that data access is a kind of price for receiving various functionalities that enhance the quality of life. What people really care about is not the level of security, but the level and type of benefits. An important barrier to use smart devices may be the reluctance to leave them unattended at home, during its work. Some users are afraid of flooding the house due to a malfunction of the washing machine. Others concern the fire caused potentially by the dryer. This fear can be addressed with appropriate functionality (e.g. automatic SMS notification of hardware failure). Page 54

55 The design of the interface is crucial. Even if 99% of the system is cables, electronic boards and software, and only 1% is a graphical display, then this one percent means all the experience of people. Satisfaction of our users is largely determined by the way they interact with the functionalities. In our case, the barrier has been the mediation of interacting with smart devices through the application. Not every user has a smartphone, not everyone likes and wants to use it. The prototype solution for starting a washing machine with a local balancing function using a series of steps (interaction with a washing machine display and a smartphone application) has proved to be too difficult. Therefore, launching functionality should be as simple as possible and not harder than before. Interface is everything (let s repeat it once again) - it should be as simple and user friendly as possible; understandable not only for technical geeks. Design it in such way, that even your grandmother would understand it (and can handle it when she forgets the glasses!). During design, as a system developer, we tended to visualize the end-user as Patrick (one of our personas). During the testing the prototype in real-life scenarios, it has often been Paul (or his mother). Electricity meets many functions at home. This means that for a system such as e-balance there is no single user of the system in the household, but there are many of them. Sometimes it can lead to strange, paradoxical situations. For example, in one of the tested households - the husband was responsible for all technical issues (including use of smartphone applications) and wife was responsible for doing the laundry (and she was no smartphone user). The husband claimed that in terms of laundry functionality we should talk to his wife; wife - when she heard about that washing machine through the application on the smartphone - directed us to her husband. The specificity of the designed interaction made none of them possible to use the functionality of local balancing. Remember that a system like e-balance alters and redefines the life of an entire household. It affects existing patterns of behaviour, habits, and attitudes. Therefore it can be said that the end-user in many cases is not a single person but the whole household. Page 55

56

57 Technical aspects The initial expectations regarding the e-balance project considered that the outcome of the project would easily be integrated in the infrastructures run by EDP in Portugal and by Alliander in the Netherlands. The project s roll-out showed that the integration into the DSO s grids presents some difficulties and hard work is required to make that integration a success. For future projects comprising legacy infrastructures, here is a list of lessons learnt: The architecture of e-balance needs to address also a way to integrate different legacy smart metering infrastructures, besides the one addressed in the project. The DSO owning a smart metering infrastructure must participate in the project which was the case of e-balance, namely involving staff related to and fully aware of such infrastructure, from the project start phase. Do not trust that integration is easy, even using open systems. Often, status quo privacy and security rules prevent the deployment of what seems to be an easy and straightforward approach towards innovation. For demonstration purposes, the researchers suggest that a complete pack of smart metering components should be included in the framework of a project willing to add further features as regards to legacy smart metering systems. Regarding the inclusion of DER actors and components in a project comprising a demonstration, do not rely on their availability, especially if they were used some years ago. Within e-balance, researchers relied on an existing MV storage installation, which proved to be out of order. The lesson learnt here is that a preliminary effort must be done prior to the project application for funding. Moreover, budget for tuning and repairing such legacy systems at least part of them should be considered, as well as a specific task within an earlier project work package should also be considered. When involving householders as end users, always trust that each one is a separate case as regards to the others. Do not expect that all such users are familiar with smart phones or tablets as they may not even own them. Consider providing a low cost tablet within your budget. Moreover, do not rely that all end users are comfortable with technology. Some are, some are not. Consider providing a budget and human resources for providing training sessions for each householder, within the deployment phase. A specific task within the integration/deployment work package would be suitable. Finally, do not forget that age, gender and cultural backgrounds will strongly determine the kind of householders you have to deal with! Still regarding the involvement of householders, if you rely on their WLAN home platform, be aware that your solution might affect the daily use of internet at domestic level. As a recommendation, you should comprise a router in your budget. Moreover, be aware that bandwidth use is limited; therefore, care on how to use it at domestic level must be a subject you should pay attention to, namely during the specification phase within an early architecture work package. In order to reduce costs on computing devices, consider providing virtualization schemes to cope with computing needs, especially if one of the project consortium members is capable of providing a virtualised platform. Page 57

58 Even looking ahead towards achieving new architectural approaches and features, once your project copes with a demonstration involving existing legacy systems and platforms, be aware that their own characteristics may prevent the success of your endeavour. As a recommendation, your project must be future-proof, meaning that it has to cope with legacy systems if the project is meant to interface them or to be integrated with them. Spend some budget and time addressing this kind of integration issues. Consider very carefully in advance how solutions such as smart plugs and remotely operable appliances impact the proposed system and make sure that all proposed partners are aligned about the concepts that need to be deployed or developed. Challenges and constraints will always occur in your daily project routine. They are the drivers to further expand your creativity and your innovation skills towards succeeding on your project goals! Transferring demonstrator equipment and responsibilities from one project to another is not a big problem within the stakeholder/company that owns the demonstrator site. However, transferring dependencies and responsibilities of consortium partners from one project to another is. It is therefore very important to incorporate a true and proper demo-site plan that makes sure that (device) support goes beyond project lifetimes. It should at least make clear to other consortia what they can expect in terms of device availability so that they can perform a proper risk analysis or include additional partners into the consortium. Any smart grid solution provided should be accompanied by on-site support by experienced and knowledgeable consortium personnel or sub-contracts to ensure proper operation. If partners are unable or unwilling to provide such support, measures should be taken into account that minimise this involvement. For example, supplying an additional router attached to a customer s router or modem via wired connection or connect it to a dedicated project 4G modem. Deploying devices that are normally out of scope of consortium partners normal operational expertise, should be supported by experienced sub-contracted installation companies or suppliers and these companies should be involved from the start of the project. Projects involving real households and people should focus on mild forms of social engineering. This can make (potential) participants ready to provide good-will. This good-will can compensate typical research project issues (they are sympathetic to flawed imperfect systems) and can also provide willingness to stay at home for the installation or (re)configuration of demonstrator devices. Business aspects 1. Our proposal of the business model emphasizes the importance of good neighbourhoods. Only when the system provider will have bottom up initiatives starting at the level of the neighbourhood (i.e. the community of energy cluster) that will be interested in being environmentfriendly and efficient, the effects of the active response can be huge. This will be possible when such neighbourhood cooperates with the aggregator, and the aggregator becomes an integral part of the cluster. The success depends on the people, and that is why a real demo site for business model testing with realistic market condition can indicate whether our proposal proves successful. Page 58

59 2. The key success factor is that the e-balance system must fulfill its objectives and bring the promised benefits to its individual participants. This is only possible when the flexibility market will be created, the demand for flexibility products from prosumers will be significant and revenue generated for all parties will be significant enough to cover investment and implementation costs. 3. The proposed price model and the chosen way of distributing the system among users are the key of importance. Results of a sensitivity analysis indicated that the higher the demand for flexibility products and the active demand side (ADR) are, the higher the income of the respective parties will be, the prosumers in particular. 4. The role of the aggregator and its place on the energy market is crucial. This function can be performed by suppliers or independent aggregators. Their model of inclusion in the current market (integration models of aggregators) and the preparation of legal regulations will define their effectiveness Some key recommendations concerning market regulations and social conditions are presented in the conclusions that have been elaborated during the work on the business model, and they include: Countries, the EU members, should create an appropriate market environment which will provide a full range of flexibility functionalities and products, applying relevant regulations. New market models reflecting the real value of flexibility should be set up. All customer groups (and prosumers with self - production and consumption) need qualified background information about opportunities offered by the flexibility energy market for making their active flexibility available. e-balance end users need clear information about options and possibilities, cost and benefits of offers, program of incentives allowing them to compare markets opportunities, and services offered by different aggregators. The prosumers should learn new behaviours focused on flexibility of their energy consumption. The business model tests are needed in a real life demo site. Tax and levies reduction programs can provide initiatives and incentives. Any technological, ICT and regulatory barriers for prosumers (when we think about cooperation between prosumers and the DSO or other parties) and self -consumption which exist today should be lifted. Regulated prices for all consumers should be gradually lifted, and real time or time of use tariffs should be introduced (first tested in the real life domo site). Manufactures of smart appliances and dispersed small generation that can be a part of smart energy home systems will play a key role in creating the Demand Response market. The DSOs are regulated companies: they must be allowed to procure system flexibility services in all timescales and to recover their costs in an appropriate manner. The aggregator as a new entity should have a dedicated regulation how to procure and settle its own procedures with all other co-partners. Page 59

60 Social aspects While designing the e-balance system, we realized from the beginning how important it is to take into consideration the future customers perspective. Therefore, when planning a research scheme, we included both quantitative exploration (to describe the market, attitudes and habits of the population and to define the specific target groups) as well as tests of the prototype in respondents' homes. The process revealed us that despite the effort it is hard to predict the impact of a system such as e- balance on the functioning of a household and its individual members. Such systems fall deeply into the daily lives of households and redefine daily functioning. Therefore with today's knowledge, we can say that: If you are designing a system that affects so many aspects of the household, it is a good idea to carry out an ethnographic study first. In our case, the e-balance system has the ambition to coordinate production and consumption of energy and help to balance it locally. This affects basic housework in the household (like washing, drying clothes, washing dishes, using the oven, etc.). An ethnographic study would describe the practice of everyday life and show how the system can fit in. Otherwise you risk implementing a prototype that needs to re-orientate the habits of the family and introduce unnatural functioning. It is worth to put the effort on planned design of system-user interaction. By saying "interaction" we mean not only the interaction with the screen, but the whole user experience (e.g. the procedure for turning on the washing machine with the local balancing function). The interaction designing process should be interwoven with the evaluation of the project (designing processtestre-designingtestetc.). If the interaction with the system is not designed from the beginning, you have to use default and template solutions, designed for other purposes by other entities (for example, to turn on the washing machine on two devices simultaneously: the washing machine display and the app on smartphone). Even if 99% of the device is hardware and 1% is an electronic display, this display is responsible for the entire user's contact with the system. In fact, from user s perspective, display is the system. Therefore, separate resources should be reserved for tests of Graphic User Interface functionality. Page 60

61

62 Acknowledgements The e-balance consortium wishes to thank the support and cooperation received from end-users in the Netherlands, Poland and Portugal to carry out social surveys, which have been paramount to understand social aspects that must be considered to roll-out systems like e-balance. We would like to express our very great appreciation to owners and end-users of Bronsbergen in the Netherlands and Batalha in Portugal respectively, since without their willingness to participate in different workshops and the demonstrators the project results wouldn t have made sense. Finally, we would like to thank the Whirlpool and Mastervolt companies (latter through the cooperation with Increase project) for their assistance with the adaptation and installation of different smart-appliances that have contributed to enable flexible load to test the balancing algorithms and services. In memoriam of our colleague Wojciech Ciemniewski, who strongly contributed to make this project real with his intelligence and sense of humour. The members of e-balance consortium will never forget him. Wojciech promoting e-balance project in Warsaw (Scientific Picnic 2014) Page 62

63 Project consortium IHP GmbH Innovations for High Performance Microelectronics (project coordinator) INESC INOVAÇÃO INSTITUTO DE NOVAS TECNOLOGIAS EDP Distribuição Energia, S.A. Universidad de Málaga CEMOSA Centro de Estudios de Materiales y Control de Obra, S.A. (Dissemination manager and guide book edition) University of Twente Alliander N.V. Ośrodek Przetwarzania Informacji (National Information Processing Institute) University of Lodz / Management Faculty / Computer Science Department EFACEC Energia Lesswire GmbH Prof. Dr. Peter Langendörfer Im Technologiepark Frankfurt (Oder) (GERMANY) Prof. Augusto Casaca RUA DE ALVES REDOL Lisboa (PORTUGAL) Mr. Carlos Mota Pinto RUA CAMILO CASTELO BRANCO 43 inovgrid Lisboa (PORTUGAL) Dr. Manuel Díaz Avda Cervantes, Málaga (SPAIN) / Dr. Noemi Jimenez-Redondo Benaque Málaga (SPAIN) Prof. Gerard Smit DRIENERLOLAAN NB Enschede (NETHERLANDS) Mr. Peter van der Sluijs UTRECHTSEWEG AH Arnhem (NETHERLANDS) Jarosław Kowalski AL NIEPODLEGLOSCI 188B Warsaw (POLAND) Bozena Ewa Matusiak, PhD DSc. Prof Ul. Matejki 22/ Lodz (POLAND) Alberto Bernardo RUA ENG FREDERICO ULRICH MAIA PORTO (PORTUGAL) Mr. Rocco Mertsching RUDOWER CHAUSSEE Berlin (GERMANY) Page 63

64 References [1] Gerards, M. E. T., Toersche, H., Hoogsteen, G., van der Klauw, T., Hurink, J. L., & Smit, G. J. M. (2015). Demand side management using profile steering. In PowerTech, 2015 IEEE Eindhoven (pp : :6). USA: IEEE Power & Energy Society. DOI: /PTC [2] Anatole Boute, Promoting renewable energy through capacity markets: An analysis of the Russian support scheme, Energy Policy, Volume 46, 2012, Pages 68-77, ISSN , [3] Piotrowski, K., Casaca, A., Gerards, M. E. T., Jongerden, M. R., Melo, F., Garrido, D., Peralta, J. (2016). A hierarchical architecture for an energy management system. In The 10th Mediterranean Conference on Power Generation, Transmission, Distribution and Energy Conversion, MedPower 2016 (pp. 109). UK: IET Digital Library. DOI: /cp [4] van der Klauw, T., Gerards, M. E. T., Smit, G. J. M., & Hurink, J. L. (2014). Optimal scheduling of electrical vehicle charging under two types of steering signals. In IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe) (pp. 0122). USA: IEEE Power & Energy Society. DOI: /ISGTEurope [5] van der Klauw, T., Gerards, M. E. T., & Hurink, J. L. (2017). Resource allocation problems in decentralized energy management. OR Spectrum (2017) 39: 749. DOI: /s [6] van der Klauw, T., Gerards, M. E. T., & Hurink, J. L. (2016). Using demand-side management to decrease transformer ageing. In 2016 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT- Europe) (pp. 186). USA: IEEE Power & Energy Society. DOI: /ISGTEurope [7] J. Weniger, J. Bergner, T. Tjaden and V. Quaschning, Economics of residential PV battery systems in the self-consumption age, in 29th European Photovoltaic Solar Energy Conference and Exhibition, Amsterdam, [8] H. Ghasemieh, B. R. Haverkort, M. R. Jongerden and A. Remke, Energy Resilience Modelling for Smart Houses, in 45th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN), 2015, Rio de Janeiro, [9] M. R. Jongerden, J. Hüls, B. R. Haverkort and A. Remke, Assessing the cost of energy independence, in IEEE International Energy Conference (ENERGYCON), 2016, Leuven, [10] M. R. Jongerden, J. Hüls, A. Remke and B. R. Haverkort, Does your domestic photovoltaic energy system survive grid outages?, Energies, vol. 9, no. 9, Page 64

65

66 ISBN