Electric Load Forecasting with Wavelet Denoising and EMD-ARIMA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current power system load prediction methods, such as ARIMA, Kalman filtering, and deep learning algorithms, face challenges in accurately modeling non-stationary and nonlinear load data due to noise interference and data preprocessing limitations, leading to reduced prediction precision and efficiency.
Innovation Solution
A method and system that combines wavelet noise reduction and empirical mode decomposition with autoregressive integrated moving average (EMD-ARIMA) to preprocess and model electric system load data, optimizing ARIMA models using Akaike and Bayesian information criteria for improved prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If wavelet noise reduction and EMD preprocessing are applied to load data, then prediction accuracy is improved, but data processing time and computational complexity increase
Solution Approach 1:
The patent applies wavelet noise reduction and EMD decomposition as preliminary preprocessing steps before ARIMA modeling. By performing noise reduction and decomposition in advance, the model receives cleaner, stationary data that requires less complex modeling, ultimately reducing the overall computational burden and improving prediction accuracy without excessive time cost
Solution Approach 2:
The patent segments the load data into multiple components through EMD decomposition, separating different frequency and amplitude characteristics. This segmentation allows each component to be modeled independently with simpler ARIMA models, reducing the overall computational complexity compared to modeling the entire non-stationary signal at once
2Device complexity
If ARIMA model is used for non-stationary load data, then computational simplicity is maintained, but prediction precision deteriorates
Solution Approach 1:
The patent transforms non-stationary load data into stationary components through wavelet noise reduction and EMD decomposition before applying ARIMA. This preliminary transformation enables the use of simple ARIMA models on data that now meets the stationarity assumption, achieving both computational simplicity and improved prediction precision
Solution Approach 2:
The patent changes the parameters of the input data by transforming it from non-stationary to stationary through preprocessing. The load data parameters are modified through wavelet thresholding and EMD decomposition, making them suitable for ARIMA modeling while maintaining model simplicity
3Duration of action of stationary object
If deep learning algorithms like LSTM are applied to load prediction, then long-term prediction capability is improved, but training time increases and convergence speed decreases
Solution Approach 1:
The patent uses simpler, computationally cheaper ARIMA models instead of expensive deep learning models. By preprocessing the data to remove non-stationarity, the patent achieves adequate prediction performance with much lower computational cost and faster execution, making the solution more practical for real-time applications
Solution Approach 2:
The patent changes the approach by transforming the data parameters through preprocessing rather than using complex model parameters. The non-stationary load data is converted to stationary components, allowing simple statistical models to capture the essential patterns without requiring extensive training
4Productivity
If raw load data is used directly for prediction without noise reduction, then processing speed is maintained, but prediction reliability deteriorates due to noise interference
Solution Approach 1:
The patent applies wavelet noise reduction as a preliminary step to remove noise and outliers from load data before prediction. This preprocessing improves data quality and prediction reliability while maintaining reasonable processing speed through efficient wavelet thresholding algorithms
Solution Approach 2:
The patent converts the harmful effect of noise in load data into a benefit by using wavelet analysis to identify and remove noise components. The noise that would normally degrade prediction reliability is transformed into removable artifacts through decomposition, leaving only the meaningful signal for prediction
Data Source
AI summary
A method and a system of predicting an electric system load based on wavelet noise reduction and empirical mode decomposition-autoregressive integrated moving average (EMD-ARIMA) are provided. The method and the system belong to a field of electric system load prediction. The method includes the following steps. Raw load data of an electric system is obtained first. Next, noise reduction processing is performed on the load data through wavelet analysis. The noise-reduced load data is further processed through an EMD method to obtain different load components. Finally, ARIMA models corresponding to the different load components are built. Further, the ARIMA models are optimized through an Akaike information criterion (AIC) and a Bayesian information criterion (BIC). The load components obtained through predicting the different ARIMA models are reconstructed to obtain a final prediction result, and accuracy of load prediction is therefore effectively improved.