| Popis: |
Power load forecasting plays an important role in the power grid, which affects the decision-making of power production and consumption. However traditional forecasting methods cannot effectively model power load data with time series characteristics. To solve this problem, a short-term load forecast based on whale optimization algorithm (WOA) and long short-term memory (LSTM) neural network is proposed, whose input is historical load data, temperature, humidity and time constant. At the same time, optimization and improvement have been made to solve the premature problem of WOA. The results show that, compared with the traditional LSTM neural network, this method improves the prediction accuracy by 1.78% and reduces the root mean square error by 1.4184, which is suitable for short-term prediction of power load. |
