A data-driven park management service QoS monitoring method and system
Through wavelet transformation and standardization technology combined with DiPCA, PCA and KDA algorithms to extract the dynamic, linear and nonlinear features of campus equipment, the problems of incomplete data feature extraction and poor model interpretability in the prior art are solved, and the rapid and reliable fault detection and early warning of campus equipment are achieved.
CN116383751BActive Publication Date: 2025-09-05XIAMEN UNIV
2 Cites 0 Cited by
Patent Information
- Application Number
- CN202310453585.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-04-25
- Publication Date
- 2025-09-05
- Estimated Expiration
- 2043-04-25
AI Technical Summary
Technical Problem
The existing process monitoring and fault detection methods are incomplete in data feature extraction and poor model interpretability, especially in nonlinear dynamic processes, the monitoring performance is not efficient and reliable enough.
Method used
The monitoring data is preprocessed by wavelet transform denoising method and Z-score standardization technology, combined with DiPCA, PCA and KDA algorithms to extract dynamic, linear and nonlinear features, and fault detection is performed through Hotelling's T2 statistic and square prediction error statistic.
Benefits of technology
It achieves fast and reliable fault warning for park equipment, improves the accuracy and timeliness of fault detection, and has strong scalability and robustness.
✦ Generated by Eureka AI based on patent content.
Abstract
The present invention discloses a data-driven park management service QoS monitoring method and system, comprising: collecting observation data of various infrastructures, electrical equipment and their environments based on Internet of Things technology; preprocessing the collected data; extracting dynamic features, linear features and nonlinear features of the data set using a method combining PCA, DiPCA and KDA in series, so that abnormal changes in the data can be reflected in each subspace; and using Hotelling's T2 statistic and squared prediction error (SPE or Q) statistic and its upper control limit for fault detection. The present invention improves process monitoring performance by extracting multiple features, while improving the interpretability of the model, can provide users with reliable monitoring and fault warning results, and effectively improves the accuracy and efficiency of park fault detection and warning.
Need to check novelty before this filing date? Find Prior Art