Prediction method for industrial load demand response adjustable potential

The invention belongs to the technical field of power demand response short-term load prediction, and particularly relates to an industrial load demand response adjustment potential-oriented prediction method based on a bidirectional long short-term memory (LSTM) network prediction method process, which is used for supporting a power grid to carry out industrial renewable energy consumption. Load prediction is crucial to balance power supply and demand and new energy consumption. At present, most load prediction based on machine learning belongs to a one-way LSTM network, the prediction accuracy and precision are limited, and therefore the load prediction method is improved through the two-way LSTM network. For an industrial load prediction scene, industrial power consumption equipment and a production process flow are considered, and the adjustable potential of different industrial production typical factories is predicted based on historical power consumption data. On the premise of considering regional new