Managing Loosely Coupled Networked Control Systems with External Disturbances
Abstract:
One of the urgent tasks for wastewater reclamation plants (WRPs) is to develop a cost effective infrastructure upon the existing facilities to meet the new regulations on the concentration of total ammonia nitrogen required by USEPA by 2014. To overcome the challenges and meet the urgent needs in a timely and cost effective manner, our research is to extend the legacy systems with available technologies, such as sensors and wireless networks, and provide real-time, on-line monitoring and process control to minimize energy demands and carbon footprint associated with nutrient control. We are the first to study cyber-physical systems as a loosely coupled networked control systems. One of our to-be-deployed systems can help a WRP to save millions of dollars on electricity bill.
We have developed an aquaponics-testing system that can be used for testing our large scale system, a wireless sensor networking system for testing the sensors and communications, the bifurcation analysis of the system, a ZIMO system for recover the zigbee signal under the wireless interference, and collected and analyzed the historical data. For the evaluation of protocols designed for this project, we also developed a wireless sensor network testbed composed of more than 200 sensor nodes and several DO sensors. This infrastructure will remain a research facility for other projects. Response times of cyber-physical systems for water resources management have a wide range. Data for those systems are currently collected weekly and response times, which play an important role in water quality in the CAW, are often several days. We also developed needed models for waste-water processing and used the ten-year data from local Stickney WRP to evaluate the effectiveness. We also successfully addressed several theoretical challenging questions such as efficient periodic data query, and top-k ranked elements query in multihop sensor networks. These results have been published in related conferences and journals. In the past year, we have focused on
- measuring water quality, and air usage statistics, environments’ impact on these statistics,
- review the existing models for design and operation of biological wastewater treatment systems, and design our own model based on the historical data collected from the WRPS,
- use historical data analysis, WRP process modeling, and D.O. field measurements to investigate the possibility of energy saving,
- Bifurcation analysis, a key method for investigating the behavior of a complex system through varying important parameters and observing the steady-state response.
- designing methods for optimal spectrum access scheme in a multi-hop network where each device can dynamically switch its channel
- coexistence of cross-technology wireless communications using MIMO and CR technologies,
- Research and Education Activities:
- In the academic year of 2012 to 2013, we organized weekly group meetings which were attended by all PIs and students working in this project. During these meetings, we discussed research challenges and system requirements, design, development, and deployment. We focused on the analysis of the historical data from the MWRDGC system, and the challenges posted to different disciplines. Our preliminary study (emulation study based on historical data) showed that a new type of cyber-physical system will lead to more efficient and effective water resources management. We specifically focused on the following aspects in the past year
- design our own model based on the historical data collected from the WRPS,
- use historical data analysis, WRP process modeling, and D.O. field measurements to investigate the possibility of energy saving,
- For bifurcation analyses we study the impact of influent flow rate, influent concentration of wastewater components, airflow rate into the tank and temperature. In addition, we also evaluate the effects of various sizes of storms through altering the storm water reaching the plant.
- The coexistence of WiFi and Zigbee technologies, the challenges posted to protecting the Zigbee signal under the WiFi interferences.
- A new channel access scheme for multi-hop wireless networks, where the channel quality is a random variable for each wireless node. The channel quality could be stochastic. The goal is to design a strategy that can learn the channel quality and access the best channel for each node.
- System integration. We have implemented several different components such as sensor networks that we deployed to get real-time data and the analysis and predict model.
- We offered a new graduate level course on 'Cyber Physical Systems: Algorithm and System Foundations' in Fall 2013 semester. The course was well attended. Several external speakers will be invited to this class to present the current state of the art on several aspects and applications of Cyber-Physical Systems.
We trained several PhD and MS students. These students have been working on the project: performing data analysis and building various models based on the historical data, collecting data using the sensors and sensor networking developed, designing hardware for the system, and designing software for learning and accessing the wireless channels with predictable services. Significant Results:
- We have achieved the following results in the past year in these research activities
- We collected DO data using the sensor networking we instrumented. To support the study, we also collected more densely sampled data using specially designed sensor networking frame. We went the Calumet WRPs several times (during different seasons) to collect the data.
- After developing and validating a comprehensive model for the Stickney wastewater plant in previous years of the project, we have focused on obtaining maximum understanding of the static and dynamic behavior of the plant under various operating conditions as a way to develop an expert knowledge database that will facilitate the development of an intelligent autonomous agents network for the monitoring and control of the plant.
- Steady-state simulations were used to explore the possible energy savings for aeration in different influent scenarios. Generally, results from these simulations suggest that about 50% energy savings could be realized while still meeting the effluent requirements. The Calumet WRP typically operates with excess aeration. Based on eight different steady-state operating scenarios, the aeration rate could be decreased by at least 41% and as much as 67%. With a safety margin, a minimum 30% savings will be tested in different perturbations. Aeration rate plays a role in maintaining process stability when dealing with influent perturbations. To investigate the potential savings in aeration, the magnitude, interval time, and duration time of stormwater flow will be used for further simulations.
- We designed and implemented a protocol, called ZIMO, which can assure harmony coexistence of wifi and zigbee signal. We also are working on ensuring harmony coexistence when the number of antennas is less than the number of interfering signals. We also designed and implemented a multi-channel access protocol with zero-regret for multi-hop wireless networks.
Activated Sludge Model
In 1983, the International Association on Water Quality1 (IAWQ, formerly IAWPRC) formed a task group, which was to promote development, and facilitate the application of, practical models for design and operation of biological wastewater treatment systems. The first goal was to review existing models and the second goal was to reach a consensus concerning the simplest mathematical model having the capability of realistically predicting the performance of single-sludge systems carrying out carbon oxidation, nitrification and denitrification. Although the model has been extended since then, for example to incorporate more fractions of COD to accommodate new experimental observations to describe growth and population dynamics of flow forming and filamentous bacteria to include new processes for describing enhanced biological phosphorus removal, the original model is probably still the most widely used for describing WWT processes all over the world. Due to its major impact on the WWT community it deserves some extra attention and it can still be considered as a ‘state-of-the-art’ model when biological phosphorus removal is not considered. As a comparison, the fourteen process equations of the UCT model were reduced to eight in the ASM1 whereas the number of state variables was only reduced by one (from fourteen to thirteen). An evaluation of the two models revealed more or less identical predictions under most operating conditions when the models had been properly calibrated.
- State Variables – COD Components in ASM1
- Readily Biodegradable Substrate
- Slowly Biodegradable Substrate
- State Variables – Nitrogen Components in ASM1
- Particulate Organic Nitrogen
- Soluble Organic Nitrogen
- Ammonia
- Nitrate
- State Variables – Suspended Solids Components in ASM1
- Heterotrophic Biomass
- Autotrophic Biomass
- Inert Particulate Matter
- GPS-X modeling
The second research approach is WRP modeling. GPS-XTM (Hydromantis, 2013) will be used to simulate the plant based on historical data. The model will serve several purposes after calibration:
- The model will be used to develop detailed concentration profiles in the aeration tank. These profiles include ammonia, D.O., and biomass. These profiles will provide valuable information on the utilization efficiency in the current aeration tank, energy saving possibilities, and a general picture of plant response to influent scenarios.
- Another function of the model will be to simulate how influent perturbations influence plant performance. Changes in stormwater flow, temperature, and loading can lead to changes in the process. By using the model it should be possible to develop corresponding control strategies to maintain process performance throughout these perturbations.
- The model will also be used to investigate energy savings options, quantify the plant resilience, an assess how to balance the potential savings while maintaining resilience in the face of perturbations.
The model will be used to test sensor temporal and spatial resolution. Major procedures of this investigation are:
- Identify target profiles in different influent scenarios based on k-means cluster analysis;
- Explore the optimum sampling temporal resolution by evaluation of the minimum accumulative RMSE value;
- Simulate typical perturbations and evaluate the accumulative RMSE values;
- Determine the optimum number of sensors by testing the 95% CL difference of accumulative RMSE value between two adjacent numbers of sensors;
- Validate this number of sensors by testing historical storm events in different influent scenarios, and the effluent quality and energy savings will be evaluated;
- With consideration of plant resilience, different target profiles will be tested for this sensor network to evaluate whether this sensor network can meet all permit requirements, save energy, and also keep plant resilience.
Bifurcation Analysis:
After developing and validating a comprehensive model for the Stickney wastewater plant in previous years of the project, we have focused on obtaining maximum understanding of the static and dynamic behavior of the plant under various operating conditions as a way to develop an expert knowledge database that will facilitate the development of an intelligent autonomous agents network for the monitoring and control of the plant. This network of agents when coupled with the network of sensors and the physical plant constitutes the essence of the cyber-physical system under study. In this phase of the project we have used bifurcation analysis to obtain this knowledge. A manuscript based on these results has been prepared and will be submitted for publication in the coming weeks. Bifurcation analysis is a key method for investigating the behavior of a complex system through varying important parameters and observing the steady-state response. For most complex systems, this reveals multiple steady states at a fixed value of the bifurcation parameter. In the case of a chemical process, bifurcation analyses enable the creation of a road map for the management of the process, so as to avoid dangerous operating conditions (or steady states) and to find an optimum setting for the corresponding parameter. For an aeration tank in Stickney WRP, the best candidates for bifurcation analyses are influent flow rate, influent concentration of wastewater components, airflow rate into the tank and temperature. In addition, it is possible to evaluate the effects of various sizes of storms through altering the storm water reaching the plant. Potentially, airflow rate is the most critical parameter among other candidates, which can greatly reduce operating costs and increase sustainability. We carried extensive studies of bifurcation on the impact of various variables. From our study, we can see that the mass transfer coefficient can be greatly decreased to half of its typical value without causing a disturbance in any of the species concentrations. This would translate to 4640 m3/d/zone of air usage, which promises more than 55% savings on air and therefore energy.
Coexistence of WiFi and Zigbee communication:
Recent studies show that WiFi interference has been a major problem for low power urban sensing technology ZigBee networks. Existing approaches for dealing with such interferences often modify either the ZigBee nodes or WiFi nodes. However, massive deployment of ZigBee nodes and uncooperative WiFi users call for innovative cross-technology coexistence without intervening legacy systems. In the past year (2012 to 2013) we mainly investigated the WiFi and ZigBee coexistence when ZigBee is the interested signal. Mitigating short duration WiFi interference (called ash) in long duration ZigBee data (called smog) is challenging, especially when we cannot modify the WiFi APs and the massively deployed sensor nodes. To address these challenges, we propose ZIMO, a sink-based MIMO design for harmony coexistence of ZigBee and WiFi networks with the goal of protecting the ZigBee data packets. The key insight of ZIMO is to properly exploit opportunities resulted from differences between WiFi and ZigBee, and bridge the gap between interested data and cross technology signals. Also, extracting the channel coefficient of WiFi and ZigBee will enhance other coexistence technologies such as TIMO (proposed by Dina Katabi from MIT).We implement a prototype for ZIMO in GNURadio-USRP N200, and our extensive evaluations under real wireless conditions show that ZIMO can improve 2.6× to 3.8× throughput for ZigBee network, and 1.1× to 1.9× for WiFi network as byproduct in ZigBee signal recovery. This project is led by Prof. XiangYang Li, from Computer Science Department, Illinois Institute of Technology, Prof. Shangping Ren from CS department, IIT, Professor Paul Anderson from the Department of Civil, Architectural, and Environmental Engineering and Professor Fouad Teymour from the Department of Chemical and Biological Engineering. This project is collaborated with the Metropolitan Water Reclamation District of Greater Chicago, which manages the Chicago Waterway System.
Submitted by Xiangyang Li
on
Abstract:
One of the urgent tasks for wastewater reclamation plants (WRPs) is to develop a cost effective infrastructure upon the existing facilities to meet the new regulations on the concentration of total ammonia nitrogen required by USEPA by 2014. To overcome the challenges and meet the urgent needs in a timely and cost effective manner, our research is to extend the legacy systems with available technologies, such as sensors and wireless networks, and provide real-time, on-line monitoring and process control to minimize energy demands and carbon footprint associated with nutrient control. We are the first to study cyber-physical systems as a loosely coupled networked control systems. One of our to-be-deployed systems can help a WRP to save millions of dollars on electricity bill.
We have developed an aquaponics-testing system that can be used for testing our large scale system, a wireless sensor networking system for testing the sensors and communications, the bifurcation analysis of the system, a ZIMO system for recover the zigbee signal under the wireless interference, and collected and analyzed the historical data. For the evaluation of protocols designed for this project, we also developed a wireless sensor network testbed composed of more than 200 sensor nodes and several DO sensors. This infrastructure will remain a research facility for other projects. Response times of cyber-physical systems for water resources management have a wide range. Data for those systems are currently collected weekly and response times, which play an important role in water quality in the CAW, are often several days. We also developed needed models for waste-water processing and used the ten-year data from local Stickney WRP to evaluate the effectiveness. We also successfully addressed several theoretical challenging questions such as efficient periodic data query, and top-k ranked elements query in multihop sensor networks. These results have been published in related conferences and journals. In the past year, we have focused on
- measuring water quality, and air usage statistics, environments’ impact on these statistics,
- review the existing models for design and operation of biological wastewater treatment systems, and design our own model based on the historical data collected from the WRPS,
- use historical data analysis, WRP process modeling, and D.O. field measurements to investigate the possibility of energy saving,
- Bifurcation analysis, a key method for investigating the behavior of a complex system through varying important parameters and observing the steady-state response.
- designing methods for optimal spectrum access scheme in a multi-hop network where each device can dynamically switch its channel
- coexistence of cross-technology wireless communications using MIMO and CR technologies,
- Research and Education Activities:
- In the academic year of 2012 to 2013, we organized weekly group meetings which were attended by all PIs and students working in this project. During these meetings, we discussed research challenges and system requirements, design, development, and deployment. We focused on the analysis of the historical data from the MWRDGC system, and the challenges posted to different disciplines. Our preliminary study (emulation study based on historical data) showed that a new type of cyber-physical system will lead to more efficient and effective water resources management. We specifically focused on the following aspects in the past year
- design our own model based on the historical data collected from the WRPS,
- use historical data analysis, WRP process modeling, and D.O. field measurements to investigate the possibility of energy saving,
- For bifurcation analyses we study the impact of influent flow rate, influent concentration of wastewater components, airflow rate into the tank and temperature. In addition, we also evaluate the effects of various sizes of storms through altering the storm water reaching the plant.
- The coexistence of WiFi and Zigbee technologies, the challenges posted to protecting the Zigbee signal under the WiFi interferences.
- A new channel access scheme for multi-hop wireless networks, where the channel quality is a random variable for each wireless node. The channel quality could be stochastic. The goal is to design a strategy that can learn the channel quality and access the best channel for each node.
- System integration. We have implemented several different components such as sensor networks that we deployed to get real-time data and the analysis and predict model.
- We offered a new graduate level course on 'Cyber Physical Systems: Algorithm and System Foundations' in Fall 2013 semester. The course was well attended. Several external speakers will be invited to this class to present the current state of the art on several aspects and applications of Cyber-Physical Systems.
We trained several PhD and MS students. These students have been working on the project: performing data analysis and building various models based on the historical data, collecting data using the sensors and sensor networking developed, designing hardware for the system, and designing software for learning and accessing the wireless channels with predictable services. Significant Results:
- We have achieved the following results in the past year in these research activities
- We collected DO data using the sensor networking we instrumented. To support the study, we also collected more densely sampled data using specially designed sensor networking frame. We went the Calumet WRPs several times (during different seasons) to collect the data.
- After developing and validating a comprehensive model for the Stickney wastewater plant in previous years of the project, we have focused on obtaining maximum understanding of the static and dynamic behavior of the plant under various operating conditions as a way to develop an expert knowledge database that will facilitate the development of an intelligent autonomous agents network for the monitoring and control of the plant.
- Steady-state simulations were used to explore the possible energy savings for aeration in different influent scenarios. Generally, results from these simulations suggest that about 50% energy savings could be realized while still meeting the effluent requirements. The Calumet WRP typically operates with excess aeration. Based on eight different steady-state operating scenarios, the aeration rate could be decreased by at least 41% and as much as 67%. With a safety margin, a minimum 30% savings will be tested in different perturbations. Aeration rate plays a role in maintaining process stability when dealing with influent perturbations. To investigate the potential savings in aeration, the magnitude, interval time, and duration time of stormwater flow will be used for further simulations.
- We designed and implemented a protocol, called ZIMO, which can assure harmony coexistence of wifi and zigbee signal. We also are working on ensuring harmony coexistence when the number of antennas is less than the number of interfering signals. We also designed and implemented a multi-channel access protocol with zero-regret for multi-hop wireless networks.
Activated Sludge Model
In 1983, the International Association on Water Quality1 (IAWQ, formerly IAWPRC) formed a task group, which was to promote development, and facilitate the application of, practical models for design and operation of biological wastewater treatment systems. The first goal was to review existing models and the second goal was to reach a consensus concerning the simplest mathematical model having the capability of realistically predicting the performance of single-sludge systems carrying out carbon oxidation, nitrification and denitrification. Although the model has been extended since then, for example to incorporate more fractions of COD to accommodate new experimental observations to describe growth and population dynamics of flow forming and filamentous bacteria to include new processes for describing enhanced biological phosphorus removal, the original model is probably still the most widely used for describing WWT processes all over the world. Due to its major impact on the WWT community it deserves some extra attention and it can still be considered as a ‘state-of-the-art’ model when biological phosphorus removal is not considered. As a comparison, the fourteen process equations of the UCT model were reduced to eight in the ASM1 whereas the number of state variables was only reduced by one (from fourteen to thirteen). An evaluation of the two models revealed more or less identical predictions under most operating conditions when the models had been properly calibrated.
- State Variables – COD Components in ASM1
- Readily Biodegradable Substrate
- Slowly Biodegradable Substrate
- State Variables – Nitrogen Components in ASM1
- Particulate Organic Nitrogen
- Soluble Organic Nitrogen
- Ammonia
- Nitrate
- State Variables – Suspended Solids Components in ASM1
- Heterotrophic Biomass
- Autotrophic Biomass
- Inert Particulate Matter
- GPS-X modeling
The second research approach is WRP modeling. GPS-XTM (Hydromantis, 2013) will be used to simulate the plant based on historical data. The model will serve several purposes after calibration:
- The model will be used to develop detailed concentration profiles in the aeration tank. These profiles include ammonia, D.O., and biomass. These profiles will provide valuable information on the utilization efficiency in the current aeration tank, energy saving possibilities, and a general picture of plant response to influent scenarios.
- Another function of the model will be to simulate how influent perturbations influence plant performance. Changes in stormwater flow, temperature, and loading can lead to changes in the process. By using the model it should be possible to develop corresponding control strategies to maintain process performance throughout these perturbations.
- The model will also be used to investigate energy savings options, quantify the plant resilience, an assess how to balance the potential savings while maintaining resilience in the face of perturbations.
The model will be used to test sensor temporal and spatial resolution. Major procedures of this investigation are:
- Identify target profiles in different influent scenarios based on k-means cluster analysis;
- Explore the optimum sampling temporal resolution by evaluation of the minimum accumulative RMSE value;
- Simulate typical perturbations and evaluate the accumulative RMSE values;
- Determine the optimum number of sensors by testing the 95% CL difference of accumulative RMSE value between two adjacent numbers of sensors;
- Validate this number of sensors by testing historical storm events in different influent scenarios, and the effluent quality and energy savings will be evaluated;
- With consideration of plant resilience, different target profiles will be tested for this sensor network to evaluate whether this sensor network can meet all permit requirements, save energy, and also keep plant resilience.
Bifurcation Analysis:
After developing and validating a comprehensive model for the Stickney wastewater plant in previous years of the project, we have focused on obtaining maximum understanding of the static and dynamic behavior of the plant under various operating conditions as a way to develop an expert knowledge database that will facilitate the development of an intelligent autonomous agents network for the monitoring and control of the plant. This network of agents when coupled with the network of sensors and the physical plant constitutes the essence of the cyber-physical system under study. In this phase of the project we have used bifurcation analysis to obtain this knowledge. A manuscript based on these results has been prepared and will be submitted for publication in the coming weeks. Bifurcation analysis is a key method for investigating the behavior of a complex system through varying important parameters and observing the steady-state response. For most complex systems, this reveals multiple steady states at a fixed value of the bifurcation parameter. In the case of a chemical process, bifurcation analyses enable the creation of a road map for the management of the process, so as to avoid dangerous operating conditions (or steady states) and to find an optimum setting for the corresponding parameter. For an aeration tank in Stickney WRP, the best candidates for bifurcation analyses are influent flow rate, influent concentration of wastewater components, airflow rate into the tank and temperature. In addition, it is possible to evaluate the effects of various sizes of storms through altering the storm water reaching the plant. Potentially, airflow rate is the most critical parameter among other candidates, which can greatly reduce operating costs and increase sustainability. We carried extensive studies of bifurcation on the impact of various variables. From our study, we can see that the mass transfer coefficient can be greatly decreased to half of its typical value without causing a disturbance in any of the species concentrations. This would translate to 4640 m3/d/zone of air usage, which promises more than 55% savings on air and therefore energy.
Coexistence of WiFi and Zigbee communication:
Recent studies show that WiFi interference has been a major problem for low power urban sensing technology ZigBee networks. Existing approaches for dealing with such interferences often modify either the ZigBee nodes or WiFi nodes. However, massive deployment of ZigBee nodes and uncooperative WiFi users call for innovative cross-technology coexistence without intervening legacy systems. In the past year (2012 to 2013) we mainly investigated the WiFi and ZigBee coexistence when ZigBee is the interested signal. Mitigating short duration WiFi interference (called ash) in long duration ZigBee data (called smog) is challenging, especially when we cannot modify the WiFi APs and the massively deployed sensor nodes. To address these challenges, we propose ZIMO, a sink-based MIMO design for harmony coexistence of ZigBee and WiFi networks with the goal of protecting the ZigBee data packets. The key insight of ZIMO is to properly exploit opportunities resulted from differences between WiFi and ZigBee, and bridge the gap between interested data and cross technology signals. Also, extracting the channel coefficient of WiFi and ZigBee will enhance other coexistence technologies such as TIMO (proposed by Dina Katabi from MIT).We implement a prototype for ZIMO in GNURadio-USRP N200, and our extensive evaluations under real wireless conditions show that ZIMO can improve 2.6× to 3.8× throughput for ZigBee network, and 1.1× to 1.9× for WiFi network as byproduct in ZigBee signal recovery. This project is led by Prof. XiangYang Li, from Computer Science Department, Illinois Institute of Technology, Prof. Shangping Ren from CS department, IIT, Professor Paul Anderson from the Department of Civil, Architectural, and Environmental Engineering and Professor Fouad Teymour from the Department of Chemical and Biological Engineering. This project is collaborated with the Metropolitan Water Reclamation District of Greater Chicago, which manages the Chicago Waterway System.