Data dual-drive power grid fault prediction method, device and equipment under typhoon disaster

Abstract

The invention discloses a data dual-drive power grid fault prediction method, device and equipment under typhoon disasters. The method comprises the following steps: constructing a disaster-causing data set, equalizing the disaster-causing data set, constructing a dual-channel prediction model, and carrying out the prediction through the dual-channel prediction model. The method comprises the following steps: classifying multivariate influence factors of a power distribution network disaster condition under a typhoon disaster into static data and dynamic data, extracting features of the static data by using a feedforward neural network, extracting features of the dynamic data by using a long-short-term memory network enhanced by a multi-head self-attention mechanism, and finally fusing all the extracted features by using a linear layer to obtain a power distribution network disaster condition prediction model. And establishing a mapping relation between the multivariate influence factors and the disaster condition of the power distribution network. The stability of the static data acting on the disaster condition of the power distribution network and the time-varying property and accumulation of the dynamic data acting on the disaster condition of the power distribution network are fully considered, and a power distribution network fault prediction model with higher accuracy and higher interpretability under the typhoon disaster is constructed.

Description

technical field [0001] The invention belongs to the technical field of power grid failure prediction, and in particular relates to a data double-driven power grid failure prediction method, device and equipment under a typhoon disaster. Background technique [0002] The impact of typhoon disasters is large and lasts for a long time. With the change of global climate in recent years, the proportion of tropical cyclones with typhoon and above intensity has been increasing, which has caused a huge threat to the normal operation of power transmission and distribution networks in coastal areas. Compared with the transmission network, the number of equipment in the distribution network is larger, the aging of the equipment is serious, and it is more susceptible to natural disasters such as typhoons. Therefore, in view of the destructiveness of typhoon and the vulnerability of distribution network, it is necessary to study the fault prediction method of distribution network under t...

AI Technical Summary

Problems solved by technology

However, the existing data-driven model only considers the relationship between each influencing factor in each time section and the fault of the distribution network, and does not take into account the cumulative effect of some factors on the fault of the distribution network.

At the same time, the SMOTE algorithm used in the current research has a certain degree of blindness and randomness in the selection process of the sample synthesis object, and the quality of the generated minority class samples is poor. When the cost-sensitive learning method determines the penalty coefficient of each class, it needs to be based on the model. Performance repeatedly adjusts parameters, and the direction of adjustment is more subjective

Images