Direct heuristic dynamic programming method for power system stability enhancement

It is very important to take into consideration time delay in wide area power system stability; the design of PSS (Power System Stabilizer) should consider global optimization with WAMS (Wide Area Measurement System) time delay. Newly designed PSS should be insensitive to time delay and suppress internal low frequency oscillations. It is used as feedback signal and is real-time synchronous that WAMS signal shows. Power system is modeled with the consideration of time delay. LMI (Linear Matrix Inequalities) is used to solve the stability condition of time delay system. Based on the time-delay effect of the wide-area measurement signals, this paper redesigned the PSS with global optimization of power system. The attached two-area-four-machine system simulation illustrates that wide-area PSS designed by global optimization with the consideration of time-delay can limit internal low frequency oscillation with time-delay insensitivity, and improve the stability of power system. It implements global optimization of PSS with WAMS time delay stability.

A single machine-infinite-bus (SMIB) system including the interline power flow controllers (IPFCs) and the power system stabilizer (PSS) controller is addressed. The linearized system model is considered for investigating the interactions among IPFC and PSS controllers. To improve the stability of whole system again different disturbances, a lead-lag controller is considered to produce supplementary signal. The proposed supplementary controller is implemented to improve the damping of the power system low frequency oscillations (LFOs). Imperialist optimization algorithm (ICA) and shuffled frog leaping algorithm (SFLA) are implemented to search for optimal supplementary controllers and PSS parameters. Moreover, singular value decomposition (SVD) method is utilized to select the most effective damping control signal of IPFC lead-lag controllers. To evaluate the system performance, different operating conditions are considered. Reponses of system in five modes including uncoordinated and coordinated modes of IPFC and PSS using ICA and SFLA are studied and compared. Considering the results, response of system without controller shows the highest overshoot and the longest settling time for rotor angel at the different operating conditions. In this mode of system, rotor speed has the highest overshoot. Rotor angel in the system with only PSS includes lower overshoot and oscillation than system without controller. When PSS is only implemented, rotor speed deviation has the longest settling time. Rotor speed deviation in the uncoordinated mode of IPFC and PSS shows lower overshoot than system with only PSS and without controller. It is noticeable that in this mode, rotor angel has higher overshoot than system with only PSS. The superiority of the suggested ICA-based coordinated controllers is obvious compared with SFLA-based coordinated controllers and other system modes. Responses of coordinated PSS and IPFC SFLA-based supplementary controllers include higher peak amplitude and longer settling time compared with coordinated IPFC and PSS ICA-based controllers. This comparison shows that overshoots, undershoots and the settling times are reduced considerably in coordinated mode of IPFC based controller and PSS using ICA. Analysis of the system performance shows that the proposed method has excellent response to different faults in power system.

An efficient and most famous tool to enhance damping of the power system low frequency oscillations is the conventional widely used lead-lag Power System Stabilizer (PSS). To achieve the desired level of robust performance under transient situation, selecting a suitable design method for optimal tuning of PSS parameters is very important in multi-machine power system. Because, it is a multimodal and difficult combinatorial optimization problem, this paper presents a novel parameter automation strategy for Particle Swarm Optimization (PSO) called PSO with Time-Varying Acceleration Coefficients (PSO-TVAC). This optimization method has a strong ability to successfully control both global and local search in each iteration process for considerably increasing the probability of finding the global optimum solution. The PSO-TVAC algorithm is applied to optimal tuning PSS parameters problem in order to reduce the PSS design effort and find the best possible solution within a reasonable computation time. For this reason, the robustly selection of PSSs parameters is converted as an optimization problem based on the time domain-based objective function under different operating conditions. The robustness of the proposed method is demonstrated on a multi-machine power system in comparison with the classical PSO and conventional method based designed PSSs. It is shown through the nonlinear time domain simulation and some performance indices for a wide range of loading condition. The analysis of the results shows that the improved PSO-TVAC is not only very effective but also provides an excellent ability for damping low frequency oscillations and greatly enhance the dynamic stability of the power system. Moreover, the proposed PSO-TVAC algorithm is superior than that of the classical PSO one in terms of accuracy, convergence and computational effort. Key words: Power system stabilizer (PSS) design, particle swarm optimization, power system dynamics, time-varying acceleration coefficients.

Employing an inverse input/output mapped artificial neural network (ANN) as a controller, an ANN based power system stabilizer (PSS) has been implemented and tested, in the laboratory environment. Experimental test results are presented in this paper. The ANN is trained off-line using the data generated by an adaptive PSS controlling the generating unit under typical disturbances. Test results show that the proposed ANN PSS exhibits very good performance in damping power system low frequency oscillations and greatly improves power system stability.

Low-frequency oscillations in power systems sometimes lead to larger disturbances leading to instability. Addition of power system stabilizer (PSS) is very effective in damping the oscillations. The main purpose of the PSS is to increase damping of synchronous machines in the excitation loop. In this paper a new meta-heuristic algorithm termed Quasi-oppositional symbiotic organism (QOSS) search has been proposed for designing the PSS parameters in an optimal manner. The technique is tested on both single machine infinite bus system as well as a multi-machine system. Various operating conditions like loading effect and line outages are also considered for establishing robustness of PSS. The PSS parameters are tuned with the help of other optimizations like Symbiotic organism search, Genetic algorithm and Particle swarm optimization and are compared with those obtained with the application of QOSOS, to show the superiority of the proposed algorithm.

Approximate dynamic programming (ADP) has been widely studied from several important perspectives: algorithm development, learning efficiency measured by success or failure statistics, convergence rate, and learning error bounds. Given that many learning benchmarks used in ADP or reinforcement learning studies are control problems, it is important and necessary to examine the learning controllers from a control-theoretic perspective. This paper makes use of direct heuristic dynamic programming (direct HDP) and three typical benchmark examples to introduce a unique analytical framework that can be applied to other learning control paradigms and other complex control problems. The sensitivity analysis and the linear quadratic regulator (LQR) design are used in the paper for two purposes: to quantify direct HDP performances and to provide guidance toward designing better learning controllers. The use of LQR however does not limit the direct HDP to be a learning controller that addresses nonlinear dynamic system control issues. Toward this end, applications of the direct HDP for nonlinear control problems beyond sensitivity analysis and the confines of LQR have been developed and compared whenever appropriate to an LQR design.

In order to easily monitor the wide-area power system information covering many power producers, suppliers, and power companies, authors have been developed a PMU (phasor measurement unit) based WAMS (wide area measurement system). This WAMS momentarily gather information about the power oscillation affecting the stability of Japanese longitudinal power systems by signal processing of the power outlet phasor. This paper introduces the features of the PMU based WAMS, and an availability of the power outlet phasor is validated by comparing the measurements with actual power station. And as one of the applications, an estimation method of interconnected power system parameters from the monitored power oscillation information is proposed.

In this work, time-driven learning refers to the machine learning method that updates parameters in a prediction model continuously as new data arrives. Among existing approximate dynamic programming (ADP) and reinforcement learning (RL) algorithms, the direct heuristic dynamic programming (dHDP) has been shown an effective tool as demonstrated in solving several complex learning control problems. It continuously updates the control policy and the critic as system states continuously evolve. It is therefore desirable to prevent the time-driven dHDP from updating due to insignificant system event such as noise. Toward this goal, we propose a new event-driven dHDP. By constructing a Lyapunov function candidate, we prove the uniformly ultimately boundedness (UUB) of the system states and the weights in the critic and the control policy networks. Consequently, we show the approximate control and cost-to-go function approaching Bellman optimality within a finite bound. We also illustrate how the event-driven dHDP algorithm works in comparison to the original time-driven dHDP.

Nowadays, the increase in electrical energy consumption and power system restructuring have posed new challenges to the operation, control, and monitoring of power systems. In this situation, the supervisory control and data acquisition (SCADA) system is not enough to ensure power system security and stability. The SCADA is often unable to measure data of all buses simultaneously. In addition, the sampling rate in this system is not enough for some power system applications. Therefore, the information obtained from SCADA does not show power system dynamics properly. In order to improve the power system monitoring, wide area measurement system (WAMS) has been developed to overcome the problems of SCADA system. Phasor measurements units (PMUs) are the main part of WAMS system and it basically consists of three essential processes including collecting, transmitting, and analyzing data. WAMS receives obtained data via a high speed communication links. After data processing and extracting appropriate information, decisions are made to improve the power system performance. Efficient use of power system data to achieve a secure operation strategy is targeted using the WAMS system. Due to the effective role of WAMS and PMUs in the reliable operation of power system, it is necessary to study their concept and applications in this chapter. The history of PMU and its structure is presented in this chapter. In addition, the necessity of WAMS for future power system and its difference from SCADA system have been investigated in this chapter. Different algorithms and application of the WAMS are also introduced in this chapter which can be implemented to improve the performance of the future power system.

Memetic salp swarm algorithm (MSSA) is proposed in the present work to optimally set the gains of fractional power system stabilizer (FPSS) for damping angular speed variation in power system. The optimization is based on the objective of minimizing variation in rotor angular frequency subject to disturbance which is change in reference voltage applied to generator. A single generator system connected with infinite bus has been experimented in this work for sudden and random variations in reference voltage applied to generator. The power system stabilizer (PSS) is based on fractional lead-lag control actions and MSSA is employed to tune the controller gains. Eigen values are analyzed with proposed control and MSSA has been compared with PSO (Particle swarm optimization) and salp swarm algorithm (SSA). It is observed from system response that MSSA can effectively tune the PSS parameters for oscillation damping subject to critical disturbances like random reference voltage variations.

This chapter contains sections titled: Introduction Direct Heuristic Dynamic Programming A Control Theoretic View on the Direct HDP Direct HDP Design with Improved Performance Case 1��-��Design Guided by a Priori LQR Information Direct HDP Design with Improved Performance Case 2��-��Direct HDP for Coorindated Damping Control of Low-Frequency Oscillation Summary References ]]>

Improvement of low-frequency oscillation damping in power systems using a deep learning technique

As a result of the multiple interruptions induced by low-frequency oscillations in power systems, they may lead to system instability. This research employs the tunicate swam algorithm (TSA), a metaheuristic optimization approach, to design the power system stabilizers (PSSs) for the multi-machine power system networks. To increase system damping, a damping ratio-based objective function is proposed and a commonly utilized traditional lead-lag type PSS structure is adopted. The effectiveness of the TSA-based PSS architecture is illustrated by two systems: one with four machines and another with ten machines. The effectiveness and robustness of the proposed method are demonstrated using many performance tests. An additional measure of trustworthiness is provided by the comparison of the simulation results to the well-known particle swarm optimization (PSO) approach for swarm intelligence in this field.

With the wide application of synchronized phase measurement unit in power system, the wide-area measurement (WAM) system has enabled the use of a combination of measured information from remote location for improving transient stability. The paper focuses on a two-level hierarchical decentralized coordinated excitation control consists of multiple decentralized local fuzzy power system stabilizers (LFPSS) for each generator at the first level helped by a WAM based supervisory power system stabilizer (SPSS) at the secondary level for wide-area power system transient stability enhancement. In order to eliminate the inherent nonlinearities in the power system model and enable the model to be applicable to power system control, first, a direct feedback linearization compensator acting through the local excitation machine is proposed. Then, the T-S fuzzy model is employed to approximate the system model. Based on the T-S fuzzy model, the LFPSS is designed. And the SPSS is proposed to apply the remote signals from the WAM systems. For taking consideration of the time delays introduced by remote information’s transmission in WAM systems and overcoming their impacts on control performance, the authors develop a delay-independent H∞ robust control technique. Some new robust stability criteria in terms of LMI are derived by Lyapunov stability theory incorporating LMI techniques. Simulation examples demonstrate that the decentralized coordinated robust control has better transient stability performance in the face of transmission time delays, severe variations of operating point and faults in various locations.

The issue of designing a high performance controller to track a desired system trajectory is one of most important problems in control theory and practice. More recently, there has been a growing interest in the study of tracking control problem. In this paper, we discuss the design and stability properties of a special approximate/adaptive dynamic programming (ADP) method for a general multiple-input-multiple-output (MIMO) discrete-time nonlinear optimal tracking control problem. The direct heuristic dynamic programming (HDP) design algorithm is firstly derived by incorporating the PID control rule into neural networks (NNs). This design approach considers using not only the typical state variables but also their derivatives and cumulative sums as inputs to the controller output. It is therefore expected to retain PID controller properties with additional learning capability. Moreover, our nonlinear control problem is formulated under a general condition that system nonlinearity is unknown and therefore it introduces modelling errors for the controller design. By using a Lyapunov stability construct, we provide new results of uniformly ultimately boundedness (UUB) for the proposed PIDNN-based direct HDP controller in discrete-time nonlinear tracking setting with desired tracking performance.