Research on Calibration Method of Electricity Data Based on GA_SVM-KF

Building electricity data which is the basis of building energy consumption statistics comes from metering sensors scattered around the building, and the data obtained by metering sensors drifts partially. The cost of manually checking these data is very high. In response to this problem, this paper studies a method based on genetic algorithm (GA) optimization that combines Support Vector Machine (SVM) and Kalman filter (KF) to calibrate electricity consumption data. Training samples and test samples are obtained by preprocessing and normalizing the collected electricity consumption data. The support vector machine model is obtained from the training samples, and the electricity consumption data is predicted by the test sample data. Finally, the Kalman filter is used to track and calibrate the drift value to obtain the calibration electricity consumption data, and compare simulation experiments with other methods. The results show that the method has reached 91.88% of the agreed coefficient of regression prediction for the data, indicating that the algorithm is more accurate in calibrating the electricity consumption data, and has a certain guiding role in the verification of energy consumption data.