Demand Response Baseline Load Forecasting Based on the Combination of Time Series and Kalman Filter

Jun Dong

doi:10.11648/j.epes.20190803.11

Abstract

The customer baseline load is an important reference for the industrial and commercial users to participate in the demand response project, and is affected by various factors such as the environment and user electricity usage. In order to improve the accuracy of the baseline load forecasting of industrial and commercial users, a demand response baseline load forecasting model based on time series and Kalman filter combination is proposed. The marginal contribution rate of the single forecasting model to the combined model is obtained by the Shapley value method, then gets optimal prediction results. The case results show that the Kalman filter model has higher prediction accuracy in the period of stable load fluctuation, and the ARMA model has higher prediction accuracy in the period of large load fluctuation, and the combined prediction model combines the advantages of both models and reduces the single model is affected by the time factor in the prediction process, which improves the overall prediction accuracy and expands the scope of application.

Highlights

Demand response (DR) is widely recognized as a key technology to improve the flexibility of power system and the reliability of power supply, which is a more economical and environmentally friendly method than traditional ones. [1] As an effective means of power demand side management, replacing supply-side energy with user-side resources using price signals and incentive mechanisms to guide users to optimize power usage is of great significance in mitigating grid pressure and maintaining safe operation of the grid
In price-based DR, users change their power consumption according to the price of electricity, which means, when the power system approaches the peak value, the price of electricity rises, and the user automatically reduces the demand for electricity or shifts it to off-peak hours, which serves to cut the peak and fill the valley, helping the power system to run stably; In the incentive-based DR, the user needs to sign a response contract with the DR implementing agency, and the DR implementing agency will reward and punish users according to the user's contract performance
A large number of foreign research on the baseline load, and due to the relative lag in demand response in China, the current focus is on short-term load forecasting of power, [5,6,7] And the researches on the baseline load mainly take a page from overseas researches, summarize the results and improve the existing methods. [8,9,10] Foreign calculation methods of user baseline load are relatively rich, [11,12,13] New

Summary

Introduction

Demand response (DR) is widely recognized as a key technology to improve the flexibility of power system and the reliability of power supply, which is a more economical and environmentally friendly method than traditional ones. [1] As an effective means of power demand side management, replacing supply-side energy with user-side resources using price signals and incentive mechanisms to guide users to optimize power usage is of great significance in mitigating grid pressure and maintaining safe operation of the grid. The new baseline load calculation formula avoids user speculation by assigning a weighted coefficient of the previous day and the same day; New York ISO established a baseline load calculation method for a recent demand response project, through selecting 5 days with the highest average daily load as reference data from historical data in the past ten days, and the prediction accuracy of the method is relatively low, but easy to operate, compared with other ISO, The California ISO calculates the 24-hour load on the event day by selecting the average of the three highest load days in the 10 days prior to the event date (excluding holidays and other event days); There is a similarity between PJM and New York ISO in baseline load calculation. In the state of data loss, the time series contains various components that affect the system, accurately reflects the characteristics of the future development trend of the system, improves the weakness of Kalman filter tracking failure in the case of data loss, and obtains the optimal prediction result

Kalman Filter Model

ARMA Model

Optimization Combination Prediction Model

Case Study

Conclusion