پيش‌بيني توان راكتيو يك جبران‌ساز استاتيكي توان راكتيو در جبران‌سازي كوره قوس الكتريكي

The rocket power of electric arc furnaces is one of the most important measures in the steelmaking process. Because arc furnaces have adverse effects on the grid, such as harmonic output, flickering voltage, temporary and permanent voltage drop. On the other hand, the power of these furnaces in making steel is usually very high, and these side effects, of course, are significant. Therefore, it is impossible to use arc furnaces without compensation. Efforts to provide a method with accuracy and speed have always been desirable. Reactive power compensation will have the best and greatest effect when the reactive power required by the furnace is generated at any time, without interruption and delay by the static compensator of reactive power and injected into the network at the common connection point. But achieving this goal is not entirely possible; Because the static compensator itself reactive power and measuring devices are delayed. In this dissertation, a method based on reactive power prediction for this delay is presented, which is obtained by combining two methods of modeling time series and Markov chain. Initially, the initial data is modeled using time series methods and the next data is predicted. Then, using the Markov chain method, another prediction is made for the prediction error of the time series method. In the end, the final value of the predicted data will be equal to the sum of the two predictions made by the two methods. Finally, the proposed method is reviewed using the data collected from Mobarakeh Steel Complex and its results are displayed. The results are then eva‎luated and compared using three error eva‎luation criteria. It is also shown that the proposed method is more accurate than previous methods. It should be noted that although the proposed method has more steps, but this does not have a significant effect on the speed of its implementation; Because these steps are mathematically simple and linear relationships.

6

34976