Probabilistic Decentralised Control and Message Passing Framework for Future Grid

Abstract

<pre>In this paper, we propose a unified probabilistic decentralised control and message passing framework for real time control of the electrical grid which enables the development of the future smart grid. The key elements of the proposed framework are the design of local randomised controllers and probabilistic message passing methodology which enables the coordination between the designed local controllers to account for optimisation considerations on the system operation. Within the proposed framework, the electric grid is decomposed into a number of control areas. Additionally, since the frequency is the ubiquitous grid state variable, representing the balancing of active power generation and consumption, the proposed framework is demonstrated based on local Load Frequency Control (LFC). Simulation studies involving a six-area power system and three interconnection schemes are carried out to illustrate the applicability and effectiveness of the proposed approach.</pre>

Decentralised Fully Probabilistic Design for Stochastic Networks with Multiplicative Noise

“…Instead of the conventional deterministic control methods, stochastic control methods have more superiorities, considering the inherent randomness of complex systems. Among the existing stochastic control methods, Decentralised Fully Probabilistic Design (DFPD) [15], [16] has been applied in many recent publications to address the remained control challenges of complex systems [2], [15], [17], as it offers a specific form of the randomised optimal controller. For instance, [2] has implemented DFPD for real-time control of the electrical power grid, while DFPD has been applied in [17] for a class of complex stochastic systems with switching modes.…”

“…Among the existing stochastic control methods, Decentralised Fully Probabilistic Design (DFPD) [15], [16] has been applied in many recent publications to address the remained control challenges of complex systems [2], [15], [17], as it offers a specific form of the randomised optimal controller. For instance, [2] has implemented DFPD for real-time control of the electrical power grid, while DFPD has been applied in [17] for a class of complex stochastic systems with switching modes. However, most of the relatively published papers have only considered additive noises, neglecting the effect of other form of noises, such as multiplicative noises.…”

“…To fix all the aforementioned issues, in this paper, the conventional DFPD will be extended by considering the multiplicative noises, thus being more suitable for the application of complex systems. Compared with the extended FPD method presented in [18], the newly proposed control method in this paper is designed in a decentralised manner, while compared with other existing DFPD methods [2], [15], [17], the proposed method in this paper has no longer limited to additional noises. Moreover, instead of single multiplicatiove noises in [18], a serious of multiplicative noises are considered in this paper for representing multiple sources of disturbances.…”