Air pollution monitoring and management system UWEDAT G. Schimak & G, Dunnebeil Department of Information Technology and Telecommunication, ARC Seibersdorf research, 2444 Seibersor- Austria Abstract UWEDAT stands for a new environmental monitoring system based on Windows NT/2000, It is mainly used in the field of ambient air quality monitoring. The kernel of the system is designed and implemented in such a modular and flexible way, thus it is possible to customize this information system in any thinkable way. But the usage of UWEDAT is not only restricted to monitor air quality. The new UWEDAT system also plays an important role for example in Vienna s drinking water quality monitoring. Within the following chapters we describe the UWEDAT-NT System. First we focus on the UWEDAT- Kernel then we will give a brief overview about some UWEDAT-tools, like the data navigator built to manage structural metadata and the data inspection tool which incorporates new aspects of data quality assurance. 1 Introduction Since the beginning of the early 1980s we are engaged in building automated and computer aided environmental monitoring systems, During the last 20 years people have learned how important such systems are to reduce effectively air pollutants or other environmental pollutants, An impression, how tremendous the reduction of S02 or CO was, is given if we compare the annual air quality reports of the 80 s with them of the 90 s in the federal province of Upper Austria [1,2]. This is the result of permanent and consequent environmental monitoring. Information systems like UWEDAT (in former times translated as environmental data acquisition system based on the ability of computer system in 1980 were very helpful to support data collection [3].
352 Air Pollution X Today the scenery has widely changed so that monitoring systems have to take into account, beside all pollutant parameters, a lot of different federal, national and European wide regulations. That forces the computerized systems to be more than a pure data acquisition system. The new age of environmental monitoring systems have to be turned from stand alone systems into filly flexible and configurable environmental information systems [4]. In the sense of information technology, we had to fmd solutions so that such environmental information systems are flexible and portable enough to operate in Europe as well as in any other country of the world, e.g. Indonesia (UWEDAT- NT Indonesia). At the beginning of 1998 we started to redesign UWEDAT and completely rebuilt it (nearly from scratch) as a new information system for ambient air quality data. 2 UWEDAT kernel The overall structure is shown in Figure 1 below. In general from the station computer data is sent to the central host, using different kinds of telegrams. These telegrams are implemented within the communication server (UwedatCS). This part of the system is the so-called central network manager, which is responsible for handling all the communication to and fi-omthe measuring stations as well as the communication with the database. As mentioned before the principal idea behind this is that the UwedatCS transfers the measuring values from the measuring station outside into the host s database. Three separate server processes are attached to this very complex task: The FORMULA Server (UwedatFS): A high speed database interface to handle mass data (time series) The Limit/Threshold Server (UwedatGS): A monitor to check the measuring values against any kind (user definable) of limit violation.. The Message Server (UwedatMS): A service to process messages according to message datapoints, These message datapoints are established by the user or generated by the system. The Uwedat-Navigator (UWENAV, see Figure 1and Figure 3) is a tool which allows the user to parametrize the monitoring network, It is clearly stated, that before any data can flow into the database, the monitoring network has to be defined. All this information (i.e. structural metadata [5]) needed, like operator of the network, related hosts and monitoring stations as well as pollutant parameters (see Figure 3) is configured with this tool,
Uwedat-Structure Air Pollution X 353 2.1 Services Figure 1: UWEDAT-NT/2000 overall structure The FORMULA Server is one of the main services in our system. To overcome the troubles you have when writing mass data (like time series) into a data base or retrieving it from a database it was necessary to design a high speed mechanism. This mechanism performs blocked records with a fixed length instead of a record for each measuring value and stores it block-wise in the database. 2.1.1 Principles and methods The monitoring data comprising measuring values including all quality information (so-called status information) are put into or retrieved from the database using the FORMULA-Server, The implementation s principle is that the FORMULA-Server handles so-called FORMULA Expressions, FORMULA expressions are strings which get parsed, compiled and evaluated by the FORMULA-Server (see Figure 2), The result of such a FORMULA string is always a time serie, This pre-compiling on the one hand and the blocked storage mechanism on the other hand allow data retrieval of half hour mean values of six different parameters (including all quality information) over a whole year within a mouse click. All data are based on and stored as a special kind of mean value time serie in the host database, respectively in a special mean value archive (typically half hour mean values archives or ten minutes archives), All other, especially higher
354 Air Pollution X aggregated mean values are based on the measuring values stored in such a source archive, In general you use FORMULA expressions [6] to reference mean value archives (sources) at any time you want to retrieve data from the UWEDAT-System. A user hidden information, stored in an archive definition table, helps FORMULA to fmd the correct type of database archive to retrieve the requested measuring values. Figure 2: FORMULA principle The FORMULA Server which operates at the heart of the UWEDAT systems supplies the user with a library holding specialized functions for large time series. Data can be aggregated over time, either by speci~ing aggregated values for defined periods (e.g. half hour mean values) or by using grouping fictions (e.g. grouping by weekday). In addition data can be filtered according to time periods or values (e.g. all working days of a week). 2.1.2 Generic mechanism for flexible access to complex archives All time series data are stored in the UWEDAT system in tables called Archives. These archives are structured in a way to support our proprietary FORMULA query mechanism. Usually the original measuring value (fust measured value) and the checked and evaluated value, with the fill status information, are kept in these archives. Depending on the user requirements it is possible to define different kinds of archives for example for half hour mean values [HMW], ten minutes mean values [MM 10], for daily mean values[tmv] etc. Typically we base all our aggregated values (for example 3-hourly mean value) on the smallest available time base (typically half hour mean values or ten minutes mean values) if not configured otherwise. On a data request the FORMULA server looks up these archives and can decide according to the archive definition table in a generic way which type of archive
Air Pollution X 355 and building rules returned. it has to use and most important which result type has to be 3 Data management There are several tools available in the UWEDAT system (see Figure 1) which allows you to perform an easy and simple data management, We will briefly discuss the three main tools in this chapter: 3.1 UWENAV the UWEDAT navigator This tool is necessary for the efficient maintenance and parameterization of the environmental database. This user interface is based on ORACLE Forms. The idea behind the concept was to have an application available, that is close to the look and feel of MS-Explorer, That means, that we have on the left hand side of the user interface an hierarchical browser and on the right hand side the related properties to each of the left hand side items (see Figure 3), Thus, you click on an item in the tree window and you will get displayed all existing properties in the right window, 3.1.1 Hierarchically data management The browsing or editing (data entry) part of UWENAV can be switched into three different tree views, Depending on the mode set within the menu bar View the tree gets displayed ordered by all operators (main center as well as regional centers) or own measuring stations or byparameters and the parameters in the property window can be displayed, entered or updated. The icon in-front of each item in the tree shows, if a measuring station or a parameter is active or inactive (red crossed icon means deactivated parameter or station) Clicking on an icon or item expands or collapses the tree at the pointer location, On selecting all operators the tree expansion shows the related measuring stations of a monitoring center. In case of viewing all stations the expansion shows all related parameters of a station (datapoints). On selecting all parameters the expansion displays all measuring stations where this parameter is measured (equipped), 3.1.2 Managing site properties and parameter setting The browsing or editing (data entry) part of UWENAV can be switched into Clicking on an operator s icon or name its properties are displayed on the right hand side (see Figure 3), All information about an operator as well as it s
356 Air Pollution X description can be entered or updated here, From the view of user interface design we have consequently kept this structure to enter all properties and settings beginning with the operator s description down to each single parameter monitored in the UWEDAT system. Figure 3: Example for operator s properties (test data) Further important database parameters, which are managed with this tool are: Stations and equipment Components Communication types and devices Phone schemas Mean values Limits Message templates Data distribution list 3.2 UWEMON the UWEDAT monitor UweMon is an utility to visualize the situation of the air quality monitoring network and to achieve basic control tasks within the network. UweMon links to the database of a monitoring network and polls the tables for state changes or updates. These state changes are visualized in an hierarchical way analogous to the conilguration tool UWENAV.
Air pollutionx 357 The UweMon program runs on the central host server. It is connected to the Uwedat services like UwedatCS (Communication server), UwedatFS (FORMULA Server) and UwedatMS (Message Server). It allows to interact with the measuring stations of a monitoring network. Several control actions can be initiated, like: start communication with the actual selected station or reload data from station, The layout of the UweMon user interface is very similar to that of the UWENAV tool. All operators and stations of the operators are displayed in an hierarchical way, Differing from the UWENAV tool, UWEMON displays all activities within the network. Typically active messages or actual measuring values of parameters in a station are displayed. 3.3 UWEDK the UWEDAT data inspection& quality assurance tool The quality check and evaluation of measurement data is done with the data inspection tool, A user interface written in JAVA, which allows also to store quality assurance information for each value verification or manipulation. 3.3.1 Concepts The data inspection tool is a tool for interactive validation and evaluation of stored measuring values, The values that have to be checked are displayed on the screen of a Windows PC. The common principle is that all measuring values can be validated in a transparent way, For quality assurance reasons, changes of values must be documented by the user. Automatically stored are the username and the changes he made. Furtheron the user may comment on, why he did and how he made these changes. Additionally manipulated values get a separate identification flag. Uncertain values are marked as invalid and restored with a quality flag in the database, To facilitate the daily work it is important that so-called profiles can be defined. This is done within the data inspection manager, A profile is a kind of cotilguration file, in which so-called evaluation groups consisting up to six parameters (components) are grouped together, These profiles are stored in the database and are reloaded when opening a new session, The data inspection graph (see Figure 4) is the application to display a data inspection group for checking and evaluating the measuring values. 3.3.2 User interface functionality The user interface is split into the following parts: menu bar, icon bar, up to six group element boxes, a graphical display and a ruler, The graphical display shows the check state of the values along with the measurement values in two separate diagrams, the so-called checking diagram and the measurement diagram. The flmctionality is as following: The checking diagram: There are several lines displayed beside the check level and exactly above the measurement diagram to indicate how far the
358 Air Pollution X user is with the checking of the measuring values, If no lines are displayed or lines are partially missing then there is until now no checking or no checking with the indicated check level. The line color is related to the color of the group element, the y-axis and the measurement diagram. The measurement diagram: Shows the measuring values of the group elements in a step - polygon. The diagram color is related to the color of the group element. Undefined or invalid values are not displayed. Values between 50% and 90% availability get highlighted, The ruler: The ruler is the highly interactive element in the measurement diagram, If you click in the diagram a ruler appears. You may move this ruler around in the graphic. Parallel to the movements the value for each parameter is displayed in the second line of the group element box. Below the x-axis the ruler time is indicated, This ruler can also be shifled from value to value by using the arrow buttons on the keyboard. When a value table is opened the line position in the table corresponds to the ruler position in the graphic so that you don t have to scroll up and down to find the related information. The value table is a very powerful user interface within the graphical display of the data inspection tool. Here it is where the user get the time series displayed as value rows with all its status information like device error status, device status and mean value status. In this table the user may change or modify values, validity information and check levels. Each of such a modification triggers a quality assurance window where the user has to enter his rational,
4 Conclusion Air Pollution X 359 Recently we have equipped Indonesia with ten central hosts in their major cities and about 50 measuring stations spread all over Indonesian territory to fidfill their air quality monitoring task with in the Blue Sky Program of Indonesia. This was a big challenge as we had to internationalize our UWEDAT system and modularize it in a way that it can be used nearly at any location in the world. We think that the design and concepts of modularity in the UWEDAT system are good so that they will stand stable for the next ten years. The kernel (server) application can be separated fi-omthe user (client) applications. That means that we can deliver a kernel with interfaces and the clients may come from another companies (already done so in Switzerland last year). The interfaces are flexible and transparent so that we can integrate the UWEDAT system into other environmental networks or information systems having a similar structure. References [1] Oberosterreichische Landesregierung: 3, Me13bericht(3rd measuring report), Amt der Oberosterreichischen Landesregierung, Landesbaudirektion, Automatisches Luftiiberwachungsnetz Oberosterreich, Hrsg: Amt der Oberosterreichischen Landesregierung/LJA-Immissionsschutz, 1984, German. [2] Oberosterreichische Landesregierung: Jahresbericht 1999 (annual report), ~erwachungsbericht des oberosterreichischen Luftmessnetzes, Abteilung Umweltschutz, Umweltpriif- und ~erwachungsstelle des Landes Oberosterreichs, 1999, German. [3] Schimak, G, (1992), H. Humer, Air Pollution Monitoring, in: P. Zannetti (cd,), Computer Techniques in Environmental Studies IV, Envirosoft 1992, Portsmouth, September 1992, pp.637-652, Elsevier, 1992 [4] Denzer, R,, Schimak, G., Humer, H, (1993): Integration Problems in Environmental Information Systems, International Symposium on Engineered Sotlvvare Systems, ISESS 1993, May 1993, Malvem, USA [5] Denzer R., Guttler R,, Houy P., Schimak G., and Temsis-Consortium (1998): TEMSIS - A Translational System for Public Information and Environmental Decision Support in: Proceedings of the 31st HICSS - Conference, Vol. 7- Soi?ware Technology Track, pg. 7-451 (CD-ROM) Kona, Big Island, Hawaii, USA, Jan. 1998, edited by Hesham E1-Rewini published by IEEE Computer Society, USA [6] Humer, H, (1993): Uwedat-Formula oder effizienter Datenzugriff trotz relationaler Datenbank ilk Visualisierungszwecke, in: Denzer R., Schimak G., Haas W. (eds.), Visualisierung von Umweltdaten 1992, 3, Workshop, Schlofi Zen an der Pram Austria, Inforrnatik Aktuell, Springer, 1993, S. 9-16