(504b) Robust Quasi-Decentralized Networked Control of Process Systems

The convergence of recent advances in sensor manufacturing, wireless communications and digital electronics has produced low-cost wireless sensor networks (WSNs) that can be installed for a fraction of the cost of wired devices [1]. The deployment of WSNs throughout the plant open up new avenues for enhancing and expanding the capabilities of existing control systems by collecting and broadly disseminating real-time information about the state of the plant units which in turn can be used to enhance the performance and robustness of the plant operations.

One of the key challenges when deploying a low-cost WSN for control is that of handling the inherent constraints on network resources, including the limitations on the computation, processing and communication capabilities. Other constraints such as limited battery power are also important when the WSN is deployed in harsh or inaccessible environments where a continuous power supply is not feasible. A tradeoff exists between the achievable control performance and the extent of network resource utilization. Specifically, maximizing the control performance requires frequent collection of data and disseminating it broadly to the target control systems. On the other hand, the limited resources of a WSN, together with the difficulty of frequent battery replacement in a plant environment, suggest that sensing and communication should be reduced in order to aggressively conserve resources and extend the lifetime of the network as much as possible. An approach to manage this tradeoff was presented in [2] where a quasi-decentralized networked control architecture that enforces close-loop stability with minimal cross communication between the constituent subsystems was developed. The main idea was to embed in the local control system of each unit a set of dynamic models that provide the local controller with estimates of the states of the neighboring units, in order to be used when state information is not transmitted over the network. Practical implementation issues such as the lack of complete state measurements were investigated and addressed in [3].

In addition to handling resource constraints, another important issue that must be accounted for in the design of networked process control systems is the presence of time-varying external disturbances which are always present in plant operations and, if unaccounted for, can degrade the networked closed-loop performance and may even lead to instability. Specifically, the presence of disturbances alters not only the stability and performance properties of the constituent subsystems but also changes the optimal rate at which these subsystems need to communicate to ensure the desired performance level. Motivated by these considerations, we develop in this work a robust quasi-decentralized networked control framework for multi-unit plants with tightly interconnected units that exchange output measurements over a shared resource-limited communication network. The objective is to robustly stabilize the plant near a desired steady-state in the presence of time-varying bounded disturbances while keeping the communication requirements between the local control systems to a minimum in order to reduce the unnecessary utilization of network resources.

The networked control structure consists of a collection of robust feedback controllers that enforce an arbitrary degree of asymptotic attenuation of the effect of disturbances on the closed-loop system in the absence of communication outages. Each controller is paired with a state observer that generates estimates of the local state variables from the measured outputs. The estimates are used to implement the local feedback control law and are also shared over the network with the other control systems to account for the interactions between the units. To reduce the exchange of information over the network without sacrificing stability, dynamic models of the interconnected units are embedded in the local control system of each unit to provide it with an estimate of the evolution of its neighbors when data are not transmitted through the network. The state of each model is then updated using the state estimate generated by the observer of the corresponding unit and transmitted over the network when communication is re-established. The closed-loop system is cast as a switched system and its stability and performance properties are analyzed leading to an explicit characterization of the interplays between the minimum allowable rate of communication between each control system and the sensors of its neighboring units, the accuracy of the embedded models, the choice of the control laws and state observers, the size of the disturbances and the achievable degree of disturbance attenuation. Finally, the theoretical results are illustrated through an application to a chemical plant example.

References:

[1] Akyildiz, I. F., W. Su, Y. Sankarasubramaniam and E. Cayirci, ``Wireless sensor networks: a survey", Computer Networks, 38:102-114, 2002.

[2] Sun, Y. and N. H. El-Farra, ``Quasi-decentralized Model-based Networked Control of Process Systems", Comp. & Chem. Eng., 32: 2016-2029, 2008.

[3] Sun, Y. and N. H. El-Farra, ``Quasi-decentralized State Estimation and Control of Process Systems Over Communication Networks", Proceedings of the 47th IEEE Conference on Decision and Control, pp. 5468-5475, Cancun, Mexico, 2008.