Advanced intelligent voltage control systems coordinating main and local power systems

Abstract

This paper describes technology required for advanced intelligent voltage and reactive secondary control systems enabling co-ordination of voltage control between main and local power systems. Firstly, the paper comments on the inadequate behavior of voltage and reactive power control systems, such as up-and-down fluctuations, which sometimes occur during offpeak load times, and explains the reasons for voltage fluctuations. It then explains that the true nature of the EMS of power system control is based on an autonomous distributed control system. Based on such recognition, the paper proposes the concept of advanced EMS systems for co-ordination of voltage and reactive power control. As background, the paper briefly describes two previously developed intelligent component functions, namely, flexible fuzzy zones and feed foreword control to prevent voltage fluctuation, and on-line acquisition of voltage characteristics by neural network technology. The paper describes local demand forecasting one hour in advance by the NN method. This method is also applicable to the forecast of the total demand of a power system.Index Terms---Voltage and reactive power control, coordination of main and local power system control, intelligent control systems, fuzzy expert, neural network, control of power systems

A local demand forecast by NN method for distributed autonomous voltage and reactive power control system

“…The NN intelligent architecture of acquiring voltage and reactive power characteristics can be designed as shown in Fig. 4 [5,10,11]. Table 1 shows an example of voltage and reactive power characteristics estimated by the developed NN method.…”

“…Local voltage and reactive compensation equipment should be provided with a more advanced intelligent control function based on distributed autonomous control theories. This paper briefly presents the concept of technological innovation for a local demand forecasting method [5,6,10]. Figure 2 shows a comparison of the control block diagrams between conventional and proposed AVQC systems [11].…”

“…The new demand forecasting NN model, shown in Fig. 5, was developed as a base model to extend the feasible time range 0.5 to 1.0 hour in advance [10].…”