Yihui Zheng

Direct Output Voltage Control Strategy for STATCOM Based on Multi-model and Neural Network PI Controller

Aiming to deal with the voltage control problems and the limitations of the conventional PI controller in the Static Synchronous Compensator (STATCOM), a direct output voltage control strategy based on multi-model and neural network PI controller is proposed. This control scheme applied the multi-model and neural network technology to the PI controller to meet the accuracy and speed of the voltage control under different impact loads. Meanwhile, the neural network technology was used to tune the PI controller parameter values according to an optimal control law, which can meet the requirements of full range working conditions and optimality. Simulation experiments show that compared to the traditional PI controller, PI controller based on multi-model and neural network is proved to be better capable of adapting to the change of voltage with a higher compensating precision.

Multiple models adaptive feedforward decoupling controller using HDBD focus

In a multivariable system, when multiple parameters jump simultaneously, a multiple models adaptive feedforward decoupling controller using hierarchical Dimension-By-Dimension technology is presented to solve the problems such as too many models, long computing time and so on. To find the optimal parameter, it adopts one-dimension search in series instead of multiple-dimension search in parallel. During the one-dimension series search, a hierarchical structure is utilized to reduce the number of the models furthermore. According to the switching index the best model is chosen. At last the global convergence is obtained. In the simulation example, when compared with the conventional multiple models adaptive controller, it reduces the number of the fixed models greatly.

Multivariable direct adaptive decoupling controller using multiple models and a case study

In this paper, a multivariable direct adaptive controller using multiple models without minimum phase assumption is presented to improve the transient response when the parameters of the system jump abruptly. The controller is composed of multiple fixed controller models, a free-running adaptive controller model and a re-initialized adaptive controller model. The fixed controller models are derived from the corresponding fixed system models directly. The adaptive controller models adopt the direct adaptive algorithm to reduce the design calculation. At every instant, the optimal controller is chosen out according to the switching index. The interaction of the system is viewed as the measured disturbance which is eliminated by the choice of the weighing polynomial matrix. The global convergence is obtained. Finally, several simulation examples in a wind tunnel experiment are given to show both effectiveness and practicality of the proposed method. The significance of the proposed method is that it is applicable to a non-minimum phase system, adopting direct adaptive algorithm to overcome the singularity problem during the matrix calculation and realizing decoupling control for a multivariable system.

Reactive Power Optimization for Distribution Network Based on Chaos Guide Particle Swarm Optimization Algorithm with Gold Criterion

Voltage is an important aspect to measure the security of power system and reactive power can relatively exert great influence on the voltage level. So planning for reactive power is an important part of network planning. In this chapter a new algorithm called Gold Criterion Chaos Guide Particle Swarm Optimization (GCCGPSO) is presented in reactive power optimization for distribution. Firstly, a mathematical model of reactive power optimization for distribution network by capacitance is established. And the cost of system active power loss and investment in equipment is treated as the optimization objective. Meanwhile the node voltage and reactive power of generator is dealt with penalty function when they pass over the limitation. Then GCCGPSO is proposed. It adopts not only chaos algorithm with gold criterion to guarantee that the particles are not easy to fall into local optimum and search the same place, but also the Neighbor domain optimal item to promote the ability of choosing path. Finally, the result of the simulation shows that the algorithm is useful and has sound performance.

Reactive Power Optimization for Wind Power System Based on Adaptive Weights Flight Adjustment Particle Swarm Optimization

In recent years, the uncertain output of wind power has had growing effects on the regional power grid. Reasonable reactive power optimization can effectively improve the adverse effects of wind power. In this chapter, an Adaptive Weights Flight Adjustment Particle Swarm Optimization (AWFAPSO) is proposed for the reactive power optimization of wind power system. First, it established a mathematic model in which system active power loss will be treated as objective function, and adopted penalty function to process node voltage cross-border and generator reactive power cross-border. Then AWFAPSO was presented. Using variable inertia factor, it can locally regulate the flight speed of the particle which leads to finding the optimal solution effectively and adopting adaptive flight time to guarantee the flight convergence in general, thus preventing particles from oscillating near optimal solution in the late of conventional particle swarm. Finally, the simulation shows that reactive power optimized by AWFAPSO can effectively reduce the system loss and improve the node voltage level.

Three-Phase Four-Wire STATCOM Control Method Based on Neural Network PI Controller

In order to solve the neutral-point imbalance problem and to improve the control precision of three-phase four-wire STATCOM, this chapter focused on the three-phase four-wire STATCOM control method based on neural network PI controller. First by analysis of the voltage imbalance problem of the split capacitors in three-phase four-wire STATCOM, a neutral-point balance control method based on the zero-sequence current is proposed. Then in order to improve the control precision, the neural network PI controller is introduced into three-phase four-wire STATCOM. Finally, the neutral-point balance control and neural network PI controller are combined together to get the neural network triple close-loop control method. Simulation result illustrates that the proposed control method is capable of neutral-point balancing control in three-phase four-wire STATCOM and the control precision is higher than that of the conventional control method.

Short-Term Wind Power Forecasting Based on Lifting Wavelet, SVM and Error Forecasting

In order to improve the forecasting accuracy, a novel forecasting method using wavelet, support vector machine (SVM), and error forecasting technology is presented in this chapter. Firstly, it utilizes lifting wavelet method to decompose data to extract the data’s main characteristics. And then it establishes the SVM forecasting model and error forecasting model to realize the wind power load forecasting, relative error forecasting, and wind load data correcting. Finally, the actual data is adopted for simulation. The experimental results show that the method based on lifting wavelet transform, SVM, and error forecasting can improve the forecasting accuracy greatly. The test shows that the method used for the wind power load forecast is feasible and effective.

Analysis and Simulation of Cascade STATCOM Based on PAM Inverter

In order to achieve similar harmonic elimination effect as Pulse Width Modulation (PWM) method at a lower switching frequency and solve the problem of DC capacitor voltage unbalance, the Pulse Amplitude Modulation (PAM) method and the method of pulse exchanging circularly are proposed in this chapter. By solving the optimal objective function, the angles of switching point can be worked out. It makes the low harmonic performance and the total output voltage optimal. The pulse generators rotate each fundamental frequency cycle time in the pulse distribution, and 10 fundamental cycles (200 ms) are needed to complete one cycle pulse rotation mechanism, which effectively improves the condition of the capacitor voltage difference. Then a neural network PI controller is designed to tune the parameters of the PI controller timely. The results of simulation show the correctness of the proposed method. It can adjust the changes through the simulation.

Short-Term Load Forecasting Based on Fuzzy Clustering Analysis Similar Days

As to the short-term electric power load forecasting, its accuracy is affected by many uncertain influencing factors. To improve the forecasting accuracy, a novel method using Similar Days based on fuzzy clustering analysis is proposed in this chapter. Firstly it categorizes the weather factors as temperature, air pressure, wind speed, overcast day, rainy day, etc., and then together with week type and day type these factors form the influence items. According to the items above, fuzzy rules are applied to establish the mapping table to get the factors quantized. Next, the cluster technology is utilized to classify the content in the mapping table, and the similar days are chosen based on the clustering level, which is to reduce the numbers of samples and accelerate the speed of selection. Secondly, aiming to eliminating non-gaussian noise contained in the similar days’ power load, lifting wavelet transform is adopted to extract the low sequence components. Finally a Least Squares Support Vector Machine (LS-SVM), which is optimized by particle swarm optimization algorithm, is designed to predict the low-frequency part while mean square weighted method is used to predict the high-frequency part. The simulation results show that this fuzzy clustering similar days method is effective.

Multiple models direct adaptive controller using dimension-by-dimension technology

For a non-minimum phase system, a multiple models direct adaptive controller using Dimension-by-Dimension technology is presented. The multiple models are constituted with controller directly, which lower the calculation and avoid the ill-condition matrix solution. The multiple models are composed of multiple fixed controllers, one free-running adaptive controller and one re-initialized adaptive controller. The fixed controllers are derived utilizing system prior information directly and guaranteed to cover the whole region in which the parameter changes. To solve the problems such as too many models, long computing time and so on, Dimension-By-Dimension technology is proposed. It adopts one-dimension optimization method in series instead of multiple-dimension optimization method in parallel to reduce the number of the controllers greatly. At last the global convergence is obtained. In the simulation example, when compared with the convenient multiple models adaptive controller, if the same number of the fixed models is adopted, system transient response are improved effectively.