Seddik Djouadi

Charalambos D. Charalambous, Umarbek Guvercin, Seddik Djouadi

Witsenhausen, in his seminal 1971 paper [1], introduced decentralized partially observable Markov decision problems (POMDPs), with multiple agents or controls operating under T-step delayed sharing information patterns. A fundamental problem in [1] is the identification of structural properties of optimal strategies that compress the information patterns into multiple information states. In this paper, we develop such structural properties of optimal strategies and associated dynamic programming (DP) equations, using the concept of decentralized sequential team equilibrium (a generalization of person-by-person optimality from static team theory). Within this framework, each strategy is assigned an individual value function conditioned on its delayed sharing information pattern, while the strategies of all other agents are held fixed. The resulting DP framework yields several new DP equations and characterizations of decentralized team equilibrium. Moreover, these DP equations exhibit fundamental properties analogous to those of centralized DP of POMDPs: the optimization in each agent's DP equations is performed over the agent's action space rather than over strategy spaces; each agent's multiple information states satisfy Markov recursions; and a separation principle holds. The DP equations reveal a structural compression property of optimal strategies: each agent compresses its delayed sharing information pattern into three components: 1) a private posterior distribution conditioned on the agent's delayed sharing information pattern, 2) a centralized posterior distribution conditioned on the common information shared by all agents, and 3) the agent's private information component. This structural result substantially extends Witsenhausen's Assertion 8 in [1].

Yichen Zhang, Alexander Melin, Mohammed Olama et al.

Generation portfolio can be significantly altered due to the deployment of distributed energy resources (DER) in distribution networks and the concept of microgrid. Generally, distribution networks can operate in a more resilient and economic fashion through proper coordination of DER. However, due to the partially uncontrollable and stochastic nature of some DER, the variance of net load of distribution systems increases, which raises the operational cost and complicates operation for transmission companies. This motivates peak shaving and valley filling using energy storage units deployed in distribution systems. This paper aims at theoretical formulation of optimal load variance minimization, where the infinity norm of net load is minimized. Then, the problem is reformulated equivalently as a linear program. A case study is performed with capacity-limited battery energy storage model and the simplified power flow model of a radial distribution network. The influence of capacity limit and deployment location are studied.

Yichen Zhang, Mohammed Olama, Alexander Melin et al.

Grid supportive (GS) modes integrated within distributed energy resources (DERs) can improve the frequency response. However, synthesis of GS modes for guaranteed performance is challenging. Moreover, a tool is needed to handle sophisticated specifications from grid codes and protection relays. This paper proposes a model predictive control (MPC)-based mode synthesis methodology, which can accommodate the temporal logic specifications (TLSs). The TLSs allow richer descriptions of control specifications addressing both magnitude and time at the same time. The proposed controller will compute a series of Boolean control signals to synthesize the GS mode of DERs by solving the MPC problem under the normal condition, where the frequency response predicted by a reduced-order model satisfies the defined specifications. Once a sizable disturbance is detected, the pre-calculated signals are applied to the DERs. The proposed synthesis methodology is verified on the full nonlinear model in Simulink. A robust factor is imposed on the specifications to compensate the response mismatch between the reduce-order model and nonlinear model so that the nonlinear response satisfies the required TLS.

Jin Dong, Teja Kuruganti, Seddik Djouadi et al.

Cybersecurity and privacy are of the utmost importance for safe, reliable operation of the electric grid. It is well known that the increased connectivity/interoperability between all stakeholders (e.g., utilities, suppliers, and consumers) will enable personal information collection. Significant advanced metering infrastructure (AMI) deployment and demand response (DR) programs across the country, while enable enhanced automation, also generate energy data on individual consumers that can potentially be used for exploiting privacy. Inspired by existing works which consider DR, battery-based perturbation, and differential privacy noise adding, we novelly consider the aggregator (cluster) level privacy issue in the DR framework of solar photovoltaic (PV) generation following. Different from most of the existing works which mainly rely on the charging/discharging scheduling of rechargeable batteries, we utilize controllable building loads to serve as virtual storage devices to absorb a large portion of the PV generation while delicately keeping desired noisy terms to satisfy the differential privacy for the raw load profiles at the aggregator level. This not only ensures differential privacy, but also improves the DR efficiency in load following since part of the noisy signal in solar PV generation has been filtered out. In particular, a mixed integer quadratic optimization problem is formulated to optimally dispatch a population of on/off controllable loads to achieve this privacy preserving DR service.

Yichen Zhang, Alexander Melin, Seddik Djouadi et al.

In this paper, a model reference control based inertia emulation strategy is proposed. Desired inertia can be precisely emulated through this control strategy so that guaranteed performance is ensured. A typical frequency response model with parametrical inertia is set to be the reference model. A measurement at a specific location delivers the information of disturbance acting on the diesel-wind system to the reference model. The objective is for the speed of the diesel-wind system to track the reference model. Since active power variation is dominantly governed by mechanical dynamics and modes, only mechanical dynamics and states, i.e., a swing-engine-governor system plus a reduced-order wind turbine generator, are involved in the feedback control design. The controller is implemented in a three-phase diesel-wind system feed microgrid. The results show exact synthetic inertia is emulated, leading to guaranteed performance and safety bounds.

Surour Alaraifi, Seddik Djouadi, Mohamed El-Moursi

Approximating regions of attraction in nonlinear systems require extensive computational and analytical efforts. In this paper, nonlinear vector fields are recasted as sum of vectors where each individual vector is used to construct an artificial system. The theoretical foundation is provided for a theorem in individual invariance to relate regions of attraction of artificial systems to the original vector field's region of attraction which leads to significant simplification in approximating regions of attraction. Several second order examples are used to demonstrate the effectiveness of this theorem. It is also proposed to use this theorem for the transient stability problem in power systems where an algorithm is presented to identify the critical clearing time through sequences of function evaluations. The algorithm is successfully applied on the 3-machine 9-bus system as well as the IEEE 39-bus New England system giving accurate and realistic estimations of the critical clearing time.

Charalambos D. Charalambous, Umarbek Guvercin, Seddik Djouadi

This paper develops a dynamic programming (DP) approach for decentralized stochastic optimal control problems with delayed sharing information patterns, which exhibits the fundamental Properties of classical DP of centralized partially observable Markov decision problems (POMDPs): the value functions and information states depend on the actions of the minimizing controls and not their strategies. This is achieved by invoking the concept of Person-by-Person (PbP) optimality, in which each control strategy is associated with a value function conditioned on its assigned delayed sharing information pattern, when all other strategies are fixed to their optimal responses. The value functions satisfy generalized and simplified DP equations. These are used to derive necessary and sufficient conditions for PbP optimality. The simplified DP equations are obtained by invoking the structural property that optimal strategies are separated and functionals of two information states: 1) a private a posteriori probability distribution based on the information pattern of the strategy, and 2) a centralized a posteriori probability distribution based on the shared or common information to all strategies, each satisfying a Markov recursion. The DP approach of this paper, settles a long standing open problem since the appearance of T-step delayed sharing patterns in [1, Section IV.G], in terms of generalizing the fundamental properties of classical DP approach.