Abstract:Abstract New energy power generation is easily affected by climate, seasons, holidays and other factors, and has different fluctuation characteristics. Aiming at the problems of low prediction accuracy and weak generalization ability of power data with different characteristics, a new energy generation data prediction method based on wavelet transform and adaptive hybrid optimization is proposed. First, the power data is transformed by wavelets to obtain sequencing of different types of information. Then LSTM model and autoregressive moving average model are used to fit and predict different sequences. Finally, the prediction sequence is reconstructed to obtain the prediction result of the power data. This method takes the actual data of a power system as the data set, which not only has high prediction accuracy, but also has strong generalization ability.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the low accuracy and weak generalization ability in new - energy power generation data prediction. Specifically, due to the influence of factors such as climate, season, and holidays, new - energy power generation data has volatile characteristics, which makes prediction complex and full of uncertainties. In order to improve the prediction accuracy and enhance the generalization ability of the model, the author proposes a new method based on wavelet transform and adaptive hybrid optimization. ### Main Problem Description 1. **Low Prediction Accuracy**: Existing prediction methods have insufficient prediction accuracy when dealing with power data with different characteristics. 2. **Weak Generalization Ability**: Existing methods have poor adaptability to different types of power data (such as stationary and non - stationary sequences), resulting in weak generalization ability. 3. **Complex Data Processing**: Power data contains many non - stationary factors (such as trend terms and irregular terms), which pose higher requirements for prediction models. ### Solution Overview To solve the above problems, the author proposes a power data prediction method based on wavelet transform and adaptive hybrid optimization. The main steps of this method are as follows: 1. **Wavelet Transform**: - Decompose power data into different types of information sequences through wavelet transform. - Determine the number of decomposition layers according to the stationarity detection results to ensure that the characteristics of each sub - sequence are fully explored. 2. **Model Selection and Prediction**: - Use the LSTM model and the autoregressive moving average model (ARMA) to fit and predict different types of sub - sequences respectively. - For non - stationary sub - sequences, use the LSTM model for prediction; for stationary sub - sequences, use the ARMA model for prediction. 3. **Adaptive Optimization**: - Use the state transition algorithm (STA) to optimize the parameters of the LSTM model, including the number of neurons, learning rate, etc. - At the same time, optimize the order of the ARMA model to improve the overall prediction performance. 4. **Result Reconstruction**: - Recombine the prediction results of all sub - sequences to obtain the final power data prediction result. ### Method Advantages - **High Prediction Accuracy**: By decomposing data through wavelet transform and combining the advantages of the LSTM and ARMA models, the prediction accuracy is improved. - **Strong Generalization Ability**: The adaptive hybrid optimization method enhances the adaptability of the model to different types of power data and improves the generalization ability. - **Automatic Feature Extraction**: Wavelet transform can automatically extract potential information in data, reducing human intervention. ### Formula Representation The following are some key formulas involved in the paper: #### Wavelet Transform Formula Suppose the power data sequence to be decomposed is \( \tilde{C}_0 \), after \( n \) - layer decomposition, it can be expressed as: \[ f(\tilde{C}_0) = C_n + D_n + D_{n - 1}+\cdots+ D_1 \] where \( C_n \) represents the approximate component, and \( D_i \) represents the \( i \)-th layer detail component. #### LSTM Model Formula The state update formulas of the LSTM model are as follows: \[ a^{(t)}_i=\sigma(W_i h^{(t - 1)}+U_i x^{(t)}+b_i) \] \[ a^{(t)}_f=\sigma(W_f h^{(t - 1)}+U_f x^{(t)}+b_f) \] \[ a^{(t)}_o=\sigma(W_o h^{(t - 1)}+U_o x^{(t)}+b_o) \] \[ h^{(t)}=a^{(t)}_o \tanh(C^{(t)}) \] \[ C^{(t)}=a^{(t)}_f C^{(t - 1)}+a^{(t)}_i \tanh(W h^{(t - 1)}+U x^{(t)}+b) \] #### State Transfer Algorithm Formula The state update formulas of the state transfer algorithm are as follows: \[ x^{(k + 1)}=A_k x^{(k)}+B_k u_k \] \[ y^{(k + 1)}=f(x^{(k + 1)}) \]

Power data prediction method of new energy system based on wavelet transform and adaptive hybrid optimization

Short-term Prediction on Output Power of Direct-drive Wave Power Generation System

An Adaptive Weight Ensemble Learning Method for Probabilistic Wind Power Prediction

Forecast of short-term wind power based on a novel hybrid method

Short-term wind-PV-wave power probabilistic prediction based on an improved sequence-to-sequence attention model

A short-term wind energy hybrid optimal prediction system with denoising and novel error correction technique

Wind process pattern forecasting based ultra-short-term wind speed hybrid prediction

Wind power forecasting system with data enhancement and algorithm improvement

Hybrid Network Model Based on Data Enhancement for Short-term Power Prediction of New PV Plants

Photovoltaic power prediction based on hybrid modeling of neural network and stochastic differential equation

Short-term prediction of the power of a new wind turbine based on IAO-LSTM

A combination predicting methodology based on T-LSTNet_Markov for short-term wind power prediction

Ultra-short Term Prediction Model of Wind Power Generation Based on Hybrid Intelligent Method

Achieving wind power and photovoltaic power prediction: An intelligent prediction system based on a deep learning approach

Optimization scheme of wind energy prediction based on artificial intelligence

Ensemble optimization approach based on hybrid mode decomposition and intelligent technology for wind power prediction system

An integrated method with adaptive decomposition and machine learning for renewable energy power generation forecasting

A Hybrid Photovoltaic Power Prediction Model Based on Multi-source Data Fusion and Deep Learning

Forecasting residential electricity consumption using the novel hybrid model

To Predict the Power Generation Based on Machine Learning Method

A short-term wind power prediction model based on CEEMD and WOA-KELM