Development of Wireless Sensor Network for Dam Monitoring

Transcription

1 Journal of Information & Computational Science 9: 6 (2012) Available at Development of Wireless Sensor Network for Dam Monitoring Xinying Miao a,b, Jinkui Chu a,, Linghan Zhang a, Jing Qiao a a Research Center of Micro System, Dalian University of Technology, Dalian , China b School of Information Engineering, Dalian Ocean University, Dalian , China Abstract A Wireless Dam Sensor Network (WDSN) is presented in this paper. The WDSN consists of smart nodes and a sink node. The power saving strategies at various levels, from the network cluster architecture, to ZigBee communication protocol, the sensor nodes architecture, and the dynamic response mechanism are explained. The experimental results demonstrate the reliability of transmission, and show the maximal transmission distance. The demonstrated WDSN has the potential application to wide range of dam monitoring. Keywords: Dam Monitoring; Wireless Sensor Networks; ZigBee; Dynamic Response; Energy-aware Implementations 1 Introduction Extreme events can cause enormous damage to the health of the dams, such as earthquakes. Such damages can impose a serious threat to the safety of lives and economics. Nearly real-time structural monitoring of the dams can reduce the loss of human lives or properties by warning of hazardous dams and impending collapses, can also provide information to emergency response services. In addition to extreme events, dam environment undergoes gradual deterioration over its life span due to corrosion, fatigue, scour, etc. Therefore, periodic monitoring should be used to provide information as to the structural soundness of the dams over their operational lives. In general, dam monitoring systems are wire based, and the sensors are deployed at few critical points in the structure and connected to a central Data Acquisition (DAQ) module over a cable, generally a co-axial cable. The wired systems throw up a host of issues, and the primary problems are their installation and maintenance. Laying out the cabling is expensive and time consuming, which results from the large sizes of structures and the installed points which are generally hard Project supported by the National Basic Research Program of China (973 Program: No. 2011CB302105), the Fundamental Research Funds for the Central Universities (No. DUT10ZD104), and Liaoning Province Research Project (No , No and No. L ). Corresponding author. address: chujk@dlut.edu.cn (Jinkui Chu) / Copyright 2012 Binary Information Press June 2012

2 1610 X. Miao et al. / Journal of Information & Computational Science 9: 6 (2012) to reach. For example, installation time of a moderate size monitoring system can consume over 75% of the total system testing time with installation costs approaching over 25% of the total system cost [1]. To overcome the many disadvantages of the wired systems, uses of wireless technologies have been proposed for structural monitoring [1]-[5]. With the advent of low cost wireless technologies, such as Bluetooth, ZigBee/IEEE , etc., there has been considerable interest in Wireless Sensor Network (WSN) as a viable alternative to the wired systems. Bluetooth is constructed with a point-to-multi-point communication structure [6]. However, only at most eight nodes can be supported by a Bluetooth network. Therefore, Bluetooth can not form large and complex networks. While ZigBee, which is able to form large networks, can be used to solve this problem [7]. Theoretically, ZigBee network structure can connect over 65,000 nodes. In addition, ZigBee network has a character of low power consumption via multi-hop technology and has the option to self-organize the whole sensor network. In this paper, we propose a ZigBee wireless dam monitoring strategy requiring the appropriate choice of the wireless network topology, the specific wireless technology as well as a suitable protocol. 2 The Dam Monitoring Application In general, dams are quite large, and the monitoring systems are widespread with some of the sensors which are placed far from the central monitoring station, which results in the process of data acquisition difficult due to the limited range of the wireless links. Also, the amount of power required to transmit data over such long distances is quite large. In a dam monitoring system, the parameters that are being measured are generally seepage and displacement. Environmental variables like temperature, water level, and rainfall are also measured in order to get an accurate picture of the dam properties. As for any monitoring system, maintenance is an expensive part of the structural monitoring system. For a dam monitoring system to make economic sense, the maintenance cycles should be long, generally years. In a wired monitoring system, wiring has to be done at a great expense and power is the least of the problems. Whereas in a wireless monitoring system, the sensor units will have to depend on batteries to provide their power, therefore the sensor units have to be very energy efficient so that they can survive for one complete maintenance cycle using one battery power. From the discussion above, we obtain some important conclusions. For the wireless monitoring system, the sensor nodes are abundant and should consume the minimum amount of power, because they have to survive on battery power for long maintenance cycles. However, achieving large transmission range requires a significant amount of power, which is clearly in conflict with the requirement that the power consumption should be as small as possible. Resolving the conflict is one important challenge during the design of the wireless monitoring system. 3 System Overview In this paper, a Wireless Dam Sensor Network (WDSN) is presented, as shown in Fig. 1. The

3 X. Miao et al. / Journal of Information & Computational Science 9: 6 (2012) Internet or GPRS Detection region 1 Sink node Detection region n Cluster head Computer management center Smart node Fig. 1: Diagram of the wireless sensor network for dam monitoring network cluster architecture, which takes advantage of multi-hop and clustering, is adopted to lower the energy consumption. The system consists of a number of smart nodes (cluster heads and the nodes are collectively called the smart nodes), a sink node, and a computer management center. The smart nodes monitor the parameters such as temperature, water level, rainfall, seepage and displacement in the dam sections. Meanwhile, each smart node can be a relay node via which the wireless communication between the smart nodes and the sink node can be implemented. The sink node is responsible for creating and controlling the network, indicating the current situation, and alerting the external and emergency services. Alerting the emergency services may be implemented through the Internet or GPRS. Communication within the network is implemented using the IEEE standard, and data are transmitted over the ZigBee protocol stack [8]. 4 System and Component Architecture 4.1 Smart Node The block diagram of the smart node is presented in Fig. 2. The system consists of three modules, namely, sensor module, computation/communication (C/C) module and power module. The C/C module is the core of this smart device. It is responsible for most of the data processing tasks inside the node and managing the wireless communication links to and from neighboring nodes. The C/C module is based on JN5139 ZigBee module which is a communication module with an embedded microprocessor and has characterization of small size and low energy consumption in operating and stand-by/sleep modes. The sensor module commonly varies from applications. For example, in the dam monitoring system, the most commonly used sensors are temperature, water level, rainfall, seepage and displacement sensors. The power module plays an important role in the smart node, and high-capacity battery is always a good choice.

4 1612 X. Miao et al. / Journal of Information & Computational Science 9: 6 (2012) The physical implementation of the smart node is shown in Fig. 3. The size of the smart node is 75mm 45mm 50mm. This tiny node includes all the modules described above. The sensor module, the C/C module and the power module are connected together via extended interfaces. By means of this kind of structure, it is easy to assemble and disassemble the node. The extended interfaces make it possible to add additional modules to the node when the system is needed to upgrade in future. Four sensor interfaces are convenient for connecting to the optional sensors. Sensor module Displacement Seepage Water level Rainfall Temperature Computation/ communication module Power module Fig. 2: Block diagram of the smart node Antenna C/C module Sensor interface Battery case Fig. 3: Physical implementation of the smart node 4.2 Sink Node The WDSN sink node is a bridge between the detection regions and GPRS or Internet. As shown in Fig. 4, the sink node is based on the JN5139 ZigBee module and the MC55 GPRS module. This configuration is the minimum one required by the network coordinator implementation. Other modules provide additional functional capabilities, e.g., indicating the network current status and detection data with a MzL LCD or selecting the kind of data which can be shown on the LCD with key buttons. Being a network coordinator (data aggregation) device, the sink node can be easily adapted to various tasks depending on the needed functional capabilities. As an additional module that extends the functional capabilities, a PC has been connected to the sink node via RS232 interface. Because the sink node consumes a substantial amount of power and sometimes needs to be mobile for testing, two power supplies, namely, a stable power supply

5 X. Miao et al. / Journal of Information & Computational Science 9: 6 (2012) LCD Key button Computation/ communication module RS232 interface GPRS module Power module Fig. 4: Block diagram of the sink node and on-board batteries are adopted in the power module. The apparent size of the sink node is 60mm 48mm 40mm. 5 Network Operation As the application of the proposed network is to detect the dam safety in separate sections, we have chosen a cluster topology for the WDSN (Fig. 1) which provides energy-saving, secure and reliable operation. The data routing in the network is performed by the ZigBee protocol. Dynamic response focuses on lowering the power consumption in this paper, which is mainly indicated on two aspects as follows: (1) Frequent dam parameter readings should be collected when the environment is turbulent (say, every 10 minutes), but only infrequent readings (say, once a day) are needed when it is stable. (2) The smart nodes measure the dam parameters, alternating with the ultra low-power sleep mode. In the measurement mode, all components of the smart node are functioning except for the ZigBee transceiver. The measured values of dams are not transmitted to the sink node until they are over the predetermined hazardous thresholds, and can be saved in a memory chip onboard. These specified thresholds can be changed by reprogramming. Data transmission is only performed when the preset events occur, so most of the time the ZigBee module is in sleep mode and its energy consumption is insignificant. 6 Simulation and Experimental Results In our work, we assume some radio parameters shown in Table 1. As shown in Fig. 5, the sensor nodes are deployed randomly in the detection region, with the area of 100m 100m. We simulate that the sink node is located far from the closest sensor node, and at (x=50, y=160). The number of sensor nodes increases from 10 to 100. We also assume an energy loss due to transmission distance. Thus, to transmit message between the cluster heads and the nodes using the free space propagation model, the radio expends: E fs = le elce + lε fs d 2, (1) and to transmit message between the sink node and the cluster nodes using two-path model, the radio expends: N E mp = le elce k + le N DA k + lε mpd 4 tosink, (2) where k is the number of clusters, d and N are the distance and the number of the smart nodes respectively.

6 1614 X. Miao et al. / Journal of Information & Computational Science 9: 6 (2012) Table 1: Radio parameters Radio parameters value Data packet by each WSN node (l) 2000bit Transmitter and receiver electronics (E elec ) 50nJ/bit Data fusion electronics(e DA ) 5nJ/bit Transmit amplifier of two-path model(ε mp ) pJ/bit/m 4 Transmit amplifier of free space propagation model(ε fs ) 10pJ/bit/m Sink node Fig. 5: Simulation environment The typical WSN with multi-hop structure, the clustering structure, and WDSN with both multi-hop and clustering architecture have been compared on the base of the LEACH protocol energy consumption model proposed by W. R. Heizelman [9]. As can be displayed in Fig. 6, the power consumption of WDSN in this paper is smaller than that of the other WSN structures. An experiment to characterize the performance of the monitoring system has been conducted. The ZigBee WSN comprises a sink node and five smart nodes. A RST-A pressure transducer is connected to the smart node farthest from the sink node, and the other smart nodes act as the relay nodes. The distance between every two nodes is adjustable. We have focused on finding the reliability and accuracy of transmission by analyzing the values shown on the sink node which change with the weights loaded on the RST-A sensor. Then, the longest transmission distance can be found by increasing the distance of every two adjacent nodes. A comparison between display values and the weights is shown in Fig. 7. It can be seen that the values shown on the sink node achieve a higher satisfactory match with the measurements. In addition, the longest transmission distance is 300m which is longer than the distance of two adjacent sections of dam, and can meet requirements of data transmission.

7 X. Miao et al. / Journal of Information & Computational Science 9: 6 (2012) Power consumption (mj) Typical WSN WSN with cluster WDSN Number of the smart nodes Fig. 6: Power consumption analysis 900 Values shown on the sink node (g) Weights loaded on the RST A sensor (g) Fig. 7: Comparison between display values and the weights 7 Conclusion In this work we have developed a Wireless Dam Sensor Network (WDSN) that employs smart nodes and a sink node. Experiments made with RST-A pressure transducer have shown that the system is able to accurately transmit the detecting data. In addition, we have found that the maximal transmission distance is achieved when the distance of adjacent nodes is increased to 300m. We have also demonstrated that the optimization of data sampling and data transmission may significantly decrease the power consumption of the sensors. Besides, the ZigBee standard provides additional power saving for the on board transmitter. Due to its reasonable simplicity, wireless connectivity, and low power consumption, the WDSN can be deployed in a short time without entailing considerable maintenance cost. In addition, the use of Internet or GPRS technologies makes it easy to manage the network in real time. The demonstrated WDSN has the potential application to wide range of dam monitoring.

8 1616 X. Miao et al. / Journal of Information & Computational Science 9: 6 (2012) References [1] E. G. Straser, A. S. Kiremidjian, A Modular, Wireless Damage Monitoring System for Structures, Report No. 128, John A. Blume Earthquake Engineering Center, Department of Civil and Environmental Engineering, Stanford University, Stanford, CA, 1998 [2] V. A. Kottapalli, A. S. Kiremidjiana, J. P. Lyncha, E. Carryerb, T. W. Kennyb, K. H. Lawa, Y. Lei, Two-tiered wireless sensor network architecture for structural health monitoring, Proc. SPIE s 10th Annual International Symposium on Smart Structures and Materials, 2003 [3] M. J. Whelan, M. V. Gangone, K. D. Janoyan, R. Jha, Real-time wireless vibration monitoring for operational modal analysis of an integral abutment highway bridge, Engineering Structures, 31 (2009), [4] G. F. Qiao, G. D. Sun, Y. Hong, Y. L. Qiu, J. P. Ou, Remote corrosion monitoring of the RC structures using the electrochemical wireless energy-harvesting sensors and networks, NDT & E International, 44 (2011), [5] Y. Kohgo, I. Asano, Y. Hayashida, A. Takahashi, R. Towmezuka, On wireless pore water transducer for fill-type dams, Proc. the 4th International Conference on Dam Engineering - New Developments in Dam Engineering, 2004, [6] C. H. Lien, Y. W. Bai, M. B. Lin, Remote-controllable power outlet system for home power management, IEEE Transactions on Consumer Electronics, 53 (2007), [7] G. Song, F. Ding, W. Zhang, A. Song, A wireless power outlet system for smart homes, IEEE Transactions on Consumer Electronics, 54 (2008), [8] ZigBee specification, Documents.aspx, 2010 [9] W. R. Heinzelman, A. Chandrakasan, H. Balakrishnan, Energy efficient communication protocol for wireless microsensor networks, Proc. the 33rd Hawaii International Conference on System Sciences, 2000,