Integrated time sequence analysis technology-based watt-hour meter measurement error prediction method

Abstract

The invention discloses an integrated time sequence analysis technology-based watt-hour meter measurement error prediction method. The method comprises the steps of carrying out data anomaly treatmentand data summary on metering errors and relevant data; dividing processed metering error data into a plurality of data sets according to time characteristics; constructing a time sequence model ARIMAand a neural network model, and carrying out training and evaluating; integrating the time sequence model ARIMA with the neural network model, and obtaining a final prediction result. According to the technical scheme of the integrated time sequence analysis technology-based watt-hour meter measurement error prediction method, the method is stronger in prediction capability compared with a traditional time sequence analysis method. Meanwhile, the time sequence model and the neural network model are combined to form a base model, so that the characteristics of a plurality of single models canbe effectively utilized. The prediction capability and the prediction accuracy can be improved.

Description

technical field [0001] The invention relates to the field of electric energy meter measurement, in particular to an electric energy meter measurement error prediction method based on integrated time series analysis technology. Background technique [0002] With the continuous development of the national economy and the continuous improvement of people's living standards, the annual electricity consumption is also increasing steadily. The accuracy of energy meter measurement is directly related to the economic interests of power companies and users. For a long time, the research on energy meter measurement errors by power operation units has mainly focused on after-the-fact maintenance and the analysis of factors affecting energy meter measurement errors. However, due to the limitations of data and analysis techniques, there is still little work on the prediction of energy meter measurement errors. The invention patent application with the application number 201610593828.1 d...

AI Technical Summary

Problems solved by technology

[0007] 1) The main use is the time-frequency characteristics of the data itself, without adding data with external influence characteristics, and the model lacks the expression of factors (and indirect factors) that affect measurement errors

[0008] 2) Classical time series analysis methods (such as ARIMA) have strict requirements on the statistical characteristics of the data itself, and the modeling process is cumbersome

[0009] 3) A single prediction model is weak in modeling and predicting complex measurement error time series data

Images