Intelligent prediction method and system for residual life of urban sewage pipeline

Abstract

The application discloses an intelligent prediction method and system for the residual life of urban sewage pipelines, and relates to the technical field of intelligent operation and maintenance, comprising the following steps: based on the extracted features, a five-tuple is constructed, an observation likelihood model containing a likelihood function and a Gaussian space kernel function is defined, the grid anomaly probability is calculated, the multi-source conflict is detected through prime number identification, the resolution strategy is selected based on the conflict rarity, and the pipeline health state probability tensor is output; the heterogeneous flow data is uniformly modeled through multi-source data fusion with conflict awareness, the multi-source conflict is detected through prime number identification and is adaptively resolved, and the four-dimensional health tensor field containing the anomaly probability and the confidence is generated; the static attribute and the dynamic health degree are combined, the residual life posterior distribution is obtained through the Weibull model and the Bayesian update, the inference stability in the complex scene is realized, and the prediction accuracy and the decision support capability are improved.

Images