an Application to Water Quality Monitoring in Malawi
Contents
1. Distance m 0 20 40 60 50000 packets Broadcast mA 12 5764 12 4254 12 3494 12 5623 Unicast mA 17 7225 17 8313 17 8194 17 7995 100000 packets Broadcast mA 25 0977 24 8324 24 9410 24 8964 Unicast mA 35 6845 35 5227 35 7539 35 7750 tances to discover the power consumption of the Sun SPOT using different output powers Using a certain amount of packets transmitted between two Sun SPOTs we tested both unicast and broadcast transmissions Note that since unicast transmission has an inherent ACK retry mechanism it could consume more power We conducted the experiment by having the received packet carrying the battery information of the transmitter in order to make the receiver able to calculate the power consumption during transmission We conducted a first experiment to discover the effect of distance on power consumption Theoretically under same output power the distance has no effect on power consumption To prove this the same amount of packets were transmitted with maximum output power under different distances Om 20m 40m 60m We decided to use the maximum power in order to reduce the chance of packet loss The results presented in Table II below reveals that the Sun SPOT consume almost the same power at different distances We condcuted another experiment to compare the power consumption for different output power Based on the first experiment we conducted only the experi2. 0 4 0 5 Oxy Tur 14 06 0 04 0 05 0 5 0 7 Con Tur 13 99 0 04 0 05 0 5 0 7 rately or together The results presented by Table I reveal a uniform pattern where when 1 using each of the probes alone consumes less power than when two or more probes are used 2 some of the probes such as Oxy and Con require higher voltage 14 24 14 25 V than others such as Tur 14 V but using higher current 0 03 0 04 A and consuming more power 0 5 0 7 W compared to 0 4 0 5 W and 3 the combinations of two or more probes lead to different performance patterns in terms of voltage current and power C Extending the lifetime of the WQWSN using a wake up mechanism The wake up mechanism is designed for saving energy We conducted a second set of experiments with and without wake up mechanism to evaluate the energy savings resulting from the wake up mechanism in a system setting where the level converter is powered by the free range Sun SPOT while the basestation Sun SPOTs power is supplied by the computer so that only the power consumption of the 90 FLT meter and the free range Sun SPOT are taken into account The experiments were run by having the system firstly runing without the wake up mechanism for 4 days by taking 60 measurements per day and thereafter having the same system runing with wake up mechanism for another 4 days also by taking 60 measurements per day For the experiment using wake up mec
3. device on weekly bases to get correct results II THE WQWSN IMPLEMENTATION The main actor in the system is the wireless sensor node During a measurement it 1 wakes up 2 fetches the latest data from the 90 FLT and 3 passes the data to the SunSPOT base station Fetching data is the process of communicating with the 90 FLT sensor It includes waking up the sensor by sending a specific string via the serial connection reading data from the sensor and turning the sensor off by sensing another string Passing data to the SunSPOT base station on the other hand includes assembling data in a more readable format and sending it via radio to the host station The SunSPOT base station is connected via USB to the gateway as depicted by Figure 2 Data coming from the free range SunSPOTs reach the SunSPOT base station via a 805 14 wireless link The java code running on the base station connects to the MySQL database that runs on the gateway and stores the received data in the database The SunSPOT base station is also responsible for 1 sending the configuration files to the free range SunSPOTs containing the measurement times and 2 synchronizing the clocks With the challenging African environment in mind we implemented the software to tackle the two main problems encountered energy consumption and inter networking Following are the mechanisms we devel oped A Wake up mechanism In order to prolong the system lifetime we decided to tur
4. 12 when the battery is fully charged the device provides the capability to read data every half hour continuously for over 40 days The device is powered off between readings and only powered on for three minutes whenever a reading is due So the device can keep working for 40 24 x 2 x 180 345600 seconds which is around 4 days The experimental results depicted by Figure 8 show that after about 4 days the battery level decreases to a warning level of 5 6 volts This reveals that the data from the manufacturer s information is reliable Assuming that N is the number of measurements per day the device wakes up for 50 seconds each time and the 90 FLT meter consumes zero energy when it is off The device can 345600 last for soy days D Extending the lifetime of the WQWSN using power control Besides the wake up mechanism power control is another mechanism that provides the potential to extend the lifetime of a wireless sensor network This may be done by either reducing the power transmission to reach less neigbour sensors or increasing the power transmission when needed to extend a sensor s reach We conducted another set of experiments using a power adjustment application under different dis 90 FLT with wake up m Samples Fig 7 Impact of wakeup mechanism on SunSPOT E Oe Fig 8 Impact of wakeup mechanism on 90 FLT TABLE II DISTANCE VS POWER CONSUMPTION MAXIMUM POWER
5. 2 J Bartram and R Ballance Water Quality Monitoring A Practical Guide to the Design and Implementation of Freshwater Quality Studies and Monitoring Programmes Edited by Published on behalf of United Nations Environment Programme and the World Health Organization ISBN 0 419 22320 7 Hbk 0 419 21730 4 Pbk 1996 3 World Health Organization Guidelines for Drinking water Quality World Health Organization 3rd Edition Volume 1 ISBN 92 4 154696 4 Available http www who int mediacentre news releases 2004 pr67 en 2006 4 J Porter et all Wireless Sensor Networks for Ecology Bioscience 55 77 Pages 561 572 July 2005 5 Zhen Fang et all Ultra Low Power WSN Node with Integrated THP Sensor st IEEE International Conference onNano Micro Engineered and Molecular Systems Pages813 816 January 2006 6 J Goldman et all Distrbuted Sensing Systems for Water Quality Asses ment and Management Center for Embedded Network Sensing Papers Paper 107 http repositories cdlib org cens wps 107 February 1 2007 7 Alix2 board Available http www pcengines ch alix2c2 htm 8 SunSPOT world Available http www sunspotworld com 9 TPS Field Lab Analyzers for PH mV TDS Dissolved Oxygen Turbidity and ture http www tps com au products combination 90 fl htm 10 Optimized Link State Routing Available http www olsr org Conductivity Available 12 90 FLT User Manual Available http www tps com a
6. On the design of a Water Quality Wireless Sensor Network WQWSN an Application to Water Quality Monitoring in Malawi Marco Zennaro Athanasios Floros Gokhan Dogan Tao Sun Zhichao Cao Chen Huang Manzoor Bahader Herve Ntareme The Royal Institute of Technology Telecommunication Systems Laboratory Stockholm Sweden Email marcoz kth se Abstract More than one billion people lack access to safe drinking water in the world Providing a way to measure auto matically water quality will help tackle this problem This paper presents the design of a water quality measuring system and proposes a prototype implementation of a water quality wireless sensor network WQWSN as a solution to this challenging problem When applied to developing countries the design and implementation of such a system must take into consideration the difficult environment in which it will operate An application to water quality measurement in Malawi reveals the relevance of using our novel solution to mitigate two challenging issues energy consumption of the system and the inter networking problem I INTRODUCTION According to the United Nations in 2005 an estimated 1 1 billion people worldwide lack clean drinking water and 2 6 billion lack access to basic sanitation 1 Hence 2005 2015 was designated the International Decade for Action Water for Life Clearly the international community has recognized that water distribution must be carefully moni
7. challenging environment in which the network is going to be installed When designing a system to be deployed in a challenging environment like the African one one has to carefully take into consideration local conditions The two main contributions of this paper are Energy consumption Although energy consumption is a well adressed issue for WSN nodes 5 water quality mea surement adds another dimension to this issue since one has to take into account the contribution of the energy consummed by the water quality sensors in the overall energy consumption of the water monitoring system As low power sensors for water quality are not yet commercially availiable 6 we propose an energy consumption minization strategy where a wake up mechnaism that triggers sleeping wake up modes is used to reduce energy consumption Furthermore while most current development solutions assume that logging devices will en counter ideal scenarios in terms of power and connectivity this is not true in many Developing Countries where power is not stable We propose a low power gateway node that can save the measurements and make them available in a web application Inter networking The gateway is the component that en ables the inter networking between the 802 15 4 and ZigBee protocols used in a WSN environment with different other protocols such as WiFI WiMAX MPLS and others to allow the dissemination of the sensed information for further pro cessing and o
8. eSerial It is a stackable add on board for the SunSPOT motes providing a single RS232 connection The board requires an extension flat ribbon cable to the appropriate DB9 or DB25 connector It contains the identification serial Flash memory RS232 level shift circuit and activity LEDs An eSerial is shown in Figure xxx C The water sensor board layer The 90 FLT series E sensor 9 purchased by TPS in Australia made it possible to measure pH conductivity TDS dissolved oxygen turbidity and temperature All data can be saved and downloaded directly to a computer via the standard RS232 port Many instrument functions can also be controlled from the computer The 90 FLT water quality sensor is composed of a main unit and some sensor probes turbidity pH and dissolved oxygen The probes are the parts to be immersed in water while the main unit must be outside This device is big in size 230 x 140 x 100 mm and the cost is 3400 including sensors and cables The 90 FLT device needs to be calibrated before making any measurement In a standard calibration the sensors are put into a sample solution of known standard values for PH turbidity conductivity and TDS and then the meter is calibrated When all the calibrations have been successfully performed then the sensors are ready to be put into any unknown solution to make Fig 3 eSerial board k Fig 4 90 FLT sensor the measurement It would be a good practice to calibrate the
9. hanism the free range Sun SPOT and 90 FLT woke up for 50 seconds in each measurement in order to transfer the latest measurements via serial connection In each of the two experiments the base station was used to record the 90 FLT battery level and the free range Sun SPOT battery capacity for each measurement The results are depicted by Figures 7 and 8 which show the battery life of the free range Sun SPOT and 90 FLT meter with and without wake up mechanism in both of the experiments From this figure it is obvious that the wake up mechanism results in a lot of energy savings The free range Sun SPOT can operate in two energy modes referred to as wake up and deep sleep phases Assume that WPC is the power consumption per second of the free range Sun SPOT in wake up phase and SPC is the power consumption persecond of the free range Sun SPOT in deep sleep phase According to the experimental results WPC is 0 021627 mA s and SPC is 0 000308 mA s Assuming that N is the number of measurements per day the power consump tion per day is W PC 50 N SPC x 86400 50 N 1 06593 N 26 6112 mA As the fully charged battery capacity of the Sun SPOT is 720mAh the free range Sun SPOT can work for Losses 0I days As the curve of 90 FLT battery level is not strictly linear the battery life of 90 FLT in the system cannot be estimated directly from the data as it is the case for the Sun SPOT According to the 90 FLT manufacturer s information
10. ices is that it runs a Java Micro Edition Virtual Machine directly on the processor without an operating system The Sun SPOT system uses JavaTM technology to up level programming Each device comes with a 180 MHz 32 bit ARM920T core 512K RAM 4M non volatile Flash memory 802 15 4 radio and a USB interface Six analog to digital converter inputs and five general purpose I O pins can be used to add custom sensors and devices The internal battery is a 3 7V 720maH rechargeable lithium ion prismatic cell The battery has inter nal protection circuit to guard against over discharge under voltage and overcharge conditions The Sun SPOT drops into a power saving mode shallow sleep to reduce power consumption and extend battery life whenever all threads become idle Considerable power can be saved during shallow sleep even though it is still necessary to power much of the Sun SPOT hardware Hence the Sun SPOT can resume from shallow sleep without any latency and as soon as any thread becomes ready to run The shallow sleep power consumption mode is about 28ma with the radio off A SunSPOT kit comes with two free range Sun SPOT units and one base station unit The base station unit is thinner does not have a battery board communicates wirelessly with the free range units and streams the data via a USB connec tion The SunSPOTs were selected for their ease of use and available interfaces Connected to the SunSPOT is an additional board called
11. mate the monitoring of drinking water quality The general work flow of the system to be designed consists of 1 taking water quality samples at a pre defined time of the day 2 sending and storing sampled data in the gateway station 3 going to sleep afterwards and 4 waking up and repeating the previous steps These four steps may be mapped into a three layer system architecture depicted by Figure 1 which includes 1 a wireless Free Range fa unSPOT station SunsPoT gt Base S i i i 1 Gateway i i i sensor a Wireless Gateway Layer Fig 1 WQWSN system architecture Fig 2 The gateway sensor gateway layer 2 a wireless sensor node layer and 3 a water sensor board layer Following is a description of the single layers A The wireless sensor gateway layer The gateway is one of the most important components upon which the efficiency of the sensing activity of a WSN depends It collects all the information received from the motes in a database and makes this information available usually via a wireless network A gateway must have enough computing power to be able to run a database perform local calculation and communicate with an existing network but should be low power enough to run autonomously in the field If a gateway is to be used for WSNs in Developing Coun tries it is clear that the scenario requires a device designed around the following co
12. ments under 0 m distance The results presented in Table III below show that by reducing the transmission power the power consumption is reduced but not significantly This is due to the Sun SPOT which consumes lots of power in normal operation These TABLE III POWER CONSUMPTION VS TRANSMISSION POWER 50000 packets Transmission power Broadcast mA 32 12 3897 31 12 5764 100000 packets Transmission power Unicast mA 17 4098 17 7225 Broadcast mA 32 24 6764mA 31 25 0977 mA Unicast mA 35 0060 35 6845 include the power of running CPU memory control etc But broadcast consumes much less power than unicast since broadcast does not need MAC layer acknowledgement As a result the power adjustment using power control did not result in any benefits for our system the power used by the adjustment was much greater than the power saved The main power saving technology in this system would result from the use of the shallow and deep sleep modes whenever possible These modes can thus be used in a real life environment to increase the lifetime of the WQWSN network V APPLICATION TO WATER QUALITY MONITORING IN MALAWI In Malawi out of a population of 11 8 million only 62 95 urban and 58 rural have access to safe drinking water and 64 90 urban and 60 rural have adequate improved sanitation 13 Malawi Demographic Health Survey 2000 Although low priority has been given to water a
13. n off the deviced we didn t use We used the external eSerial board to switch off the 90 FLT sensor when not reading data It was quite easy to turn on and off the 90 FLT Sending a special string via the serial port was enough to power it on and off Furthermore we powered down the eSerial itself when not communicating with the 90 FLT In this way we powered just the SunSPOT s motherboard when in idle mode B Inter networking Based on Linux the gateway we developed is flexible enough to work in different scenarios and to provide a way to interconnect different networks For example 802 11 wireless It could work as a client connected to an access point which is then connected to the Internet Users could use a web browser to access the data via a web application Given possible long distances involved in the deployment we think that the best solution is to implement a mesh protocol such as OLSR 10 in the gateway In this way one deploys more gateways each connected to different SunSPOT networks and the gateways automatically form a network One of the gateways is con nected to the internet and traffic is automatically routed to it by OLSR Each gateway runs an instance of MySQL and the gateway connected to the internet hosts a database with data coming from all the databases This is done via MySQL replication mechanism 11 Replication is asynchronous by default the master gateway does not need to be connected permanently t
14. nd sanitation programs water quality has been a major cause of mortality especially for children under five year old The first implementation of our system will be at the Blantyre Water Board responsible for delivery of water supply and services in the city of Blantyre The Blantyre Water Board has two water intake stations Mudi dam is the smaller of the two intake stations and it provides approximately 10 of the water used in Blantyre The other station is located on the Shire River in Chikwawa district and accounts for 90 of the total water used in the city The catchment area of Mudi Dam also houses the main laboratories that monitor physical chemical and bacteriological parameters to control water quality The water that flows into Mudi Dam about 1 km2 xxx comes from two small rivers Namimba Stream and Mudi Stream When the Dam fills up at the end of the rainy season January March the overflow is drained into Mudi river which ultimately flows into the Shire River The water pumped from the Dam is first treated with a coagulant then passed through a set of pools for sedimentation sand filtration and finally chlorination After a site survey at the Chichiri premises we decided we would place three wireless sensors to monitor water quality one in the catchment area and two following the treatment process The distances between the pools is of about 100 meters while the catchment area is some 300 away The results of the analysis
15. nstraints low power consumption to run using solar panels or using batteries high storage capabil ities to be able to store data for a long period of time in case of remote deployments flexible connectivity to be able to connect to the network via wired or wireless connection using different topologies low cost to be suitable for deployments in Developing Countries and web based design to allow users to visualize the data from the WSN without installing specific software To develop our gateway we selecetd an ALIX 2 embedded Linux board 7 built by PC engines as depicted by Figure 2 It is based on the 500 MHz AMD Geode processor with 256MB of DRAM has two USB ports one Compact Flash socket two miniPCI slots and two Ethernet ports It is powerful enough to run a database and a web server We decided to use a 1G CF memory card capable enough to store the OS and sampled data It is compact lightweight has high performance and low power consumption It can be used as a wireless Access Point Client or mesh node Using one wireless card it consumes 4 4 W This value is low enough to be able to use a 20W solar panel and a 14 AH battery to make the gateway autonomous With a 1G Compact Flash card it can store months of measurements B The wireless sensor node layer The wireless sensor nodes were assembled from comercially available SunSPOT motes 8 with an additional board What distinguishes the Sun SPOT from comparable dev
16. o receive updates which means that updates can occur over long distance connections and even temporary solutions Depending on the configuration once can replicate all databases selected databases or even selected tables within a database We tested this solution and it works perfectly given the Open Source software used IV PERFORMANCE EVALUATION Using four sensor motes and one gateway we conducted different experiments to evaluate the performance of the newly proposed system and the impact of the wireless sensor network on water quality monitoring We considered different perform nace parameters These include 1 the power consumption and 2 the lifetime of the sensor network in different sensing scenarios While the power consumption is a parameter that measures the power utility of a wireless sensor device a wireless sensor network lifetime is expressed by the battery levels of its components A lower power consuming wireless sensor network is preferable to a highly power consuming wireless sensor network A wireless sensor network with long life batteries would be preferred to a short lived batteries since it provides a higher lifetime for the system monitored A Determining the lifetime of the WQWSN system We conducted a first set of experiments to determine the lifetime of the WQWSN by comparing the lifetime of the SUN SPOT and 90 FLT device to evaluate their impact on the lifetime of the system the lifetime of the system
17. r decision making A gateway that is intended to be deployed in challenging environments such as those of the African continent should be appropriately designed to meet stringent power communication and reliability requiments While PCs and laptops are currently being used as part of a gateway environment we propose a flexible gateway solution that builds upon an ALIX2 embedded Linux board from PC engines 7 Our gateway is compact lightweight has high performance and lower power and can be used as a wireless Access Point client or a mesh node in an OLSR based wireless network The remainder of this paper is as follows Section 2 presents the proposed WQWSN solution while the experimental re sults on power consumption and wireless sensor lifetime are presented by section 3 Section 4 reports on an application to water quality monitoring in Malawi Our conclusions and future work are presented in section 5 Il THE WQWSN DESIGN For the initial deployment of the system we are interested in studying fluctuations in pH dissolved oxygen temperature and dissolved oxygen Sensor data need to be gathered at inter vals of one hour and data collection has to be automated and the data easily retrievable at the end of the observation period The rest of this section describes the system architecture used to meet these requirements and the different components of the water quality monitoring system The ultimate purpose of the WQWSN is to auto
18. tored and controlled in terms of water quality and quantity Water quality can be thought of as a measure of the suitability of water for a particular use based on selected physical chemical and biological characteristics 2 To deter mine water quality scientists first measure and analyze char acteristics of the water such as temperature dissolved mineral content and number of bacteria Selected characteristics are then compared to numeric standards and guidelines to decide if the water is suitable for a particular use Standards and guidelines are established to protect water for designated uses such as drinking recreation or agricultural irrigation Stan dards for drinking water quality ensure that public drinking water supplies are as safe as possible Some aspects of water quality can be determined right in the stream or at the well The basic parameters to be measured include acidity pH dissolved oxygen and turbidity a measure of the clarity of the water 3 Analyses of individual chemicals generally Antoine Bagula Department of Computer Science Third Floor 18 University Avenue Rhodes Gift 7707 University of Cape Town Western Cape South Africa Email bagula cs uct ac za are done at a laboratory In many Developing Countries monitoring procedures are currently entirely manual based on sampling and subsequent analyses in water laboratories This often causes considerable delays in the monitoring process The quali
19. ty of the monitoring process would benefit from at least some basic parameters being monitored in real time in order to get early warnings that can trigger appropriate treatment Emerging technologies can be used to provide relatively inexpensive coordinated intelligent networks allowing a well coordinated and continuous monitoring of surface waters Environmental sensor networks are wireless sensor networks that may provide a key element for better understanding and managing the water cycle A typical wireless sensor network is composed of sensor nodes often referred to as motes which sense their environment and a wireless sensor base station that collects the information sensed by the motes for local and or remote processing Such networks have been proposed and or are being developed for a variety of environmental applications 4 but their deployment to measure water quality is still in its infancy This paper proposes an integrated sensor network for water quality monitoring and describes its application to water quality monitoring in Malawi In the newly proposed net work referred to as Water Quality Wireless Sensor Network WQWSN sensors become an integrated component of their aqueous environment throughout data collection and data are communicated between nodes and back to researchers remotely using wireless connections In addition to describing the development of the system we describe ongoing and future research related to the
20. u handbooks 90 FLTv9_01 pdf 13 Malawi Demographic and Health Survey 2004 Available http www measuredhs com pubs pdf FR 175 FR 175 MW04 pdf 11 MySQL Replication Available dev mysql com doc refman 5 0 en replication html
21. will be used both at the Water Board and at the Malawi Polytechnic so a web server publishing the results is a must From the connectivity point of view the Water Board has an ADSL connection and we are going to use that to connect to the Internet VI CONCLUSIONS AND FUTURE WORK This paper proposes the design of a water quality moni toring system and building upon the SunSPOT technology a prototype implementation of a water quality wireless sensor network WQWSN as a solution to the water quality monitor ing problem Using a three layer architecture we discuss the main features of the newly implemented system and propose its application to water quality measurement in Malawi The experimental results reveal the relevance of using our novel so lution to mitigate two challenging issues energy consumption of the system and the inter networking problem Besides Malawi the deployment of the WQWSN to mea sure water quality in the rural areas of South Africa Mozam bique Tanzania and the DRC has been planned as an exten sion to this work Its expansion into a WaterQualitySensorWeb WQSW is also a further step to the extensions of our proposed WQWSN prototype ACKNOWLEDGMENTS WQWSN is a project supported by SPIDER the Swedish Program for Information and Communication Technology in Developing Regions REFERENCES 1 United Nations Water a shared responsibility The 2nd UN World Water Development Report March 2006
22. will obviously depend on the lifetime of the shortest lived of the two devices Figure 5 and 6 report on the relationship between the battery lifetime of the 90 FLT device and the free range Sun SPOT in number of days and number of measurements per day These Figures show that the 90 FTL device has much longer life than the free range Sun SPOT So the system life depends on the life of free range Sun SPOT The usual sample frequency is 24 times per day In this situation the Sun SPOT can work for 14 days B Determining the power consumption of different water probes We conducted another experiment to determine the energy consumed by the system when using different probes sepa sooo 90 FLT battery level 4000F 3000F Battery level in mV 2000F 1000F 0 5 g 15 20 25 30 Measurements per day Fig 5 9OFLT lifetime 26 i 26 Free range SunSPOT life 3 2a 3 20 ees re a Fig 6 SunSPOT lifetime TABLE I POWER CONSUMPTION FOR OXY OXYGEN CON CONDUCTIVITY TUR TURBIDITY Voltage V Current A Power W None probe 14 3 0 03 0 04 0 4 0 5 All probes 14 03 0 04 0 05 0 5 0 7 Oxy Con Tur 13 97 0 04 0 05 0 6 0 7 Oxy Con Tur pH 13 97 0 04 0 05 0 6 0 7 Oxy Con Tur Redox 13 97 0 04 0 05 0 6 0 7 Con Tur 14 05 0 04 0 05 0 5 0 7 Tur 14 0 04 0 05 0 5 0 7 Oxy 14 24 0 03 0 04 0 4 0 5 Con 14 25 0 03 0 04 0 4 0 5 Oxy Con 14 33 0 03 0 04
Download Pdf Manuals
Related Search