Study of Condenser Spatial State Model Based on Dynamic Thresholding of Environmentally Adaptive Multi-Dimension Eigenvalues

¹ College of Automation Engineering, Shanghai University of Electric Power, Shanghai 200090, China
² China Nuclear Power Engineering Co., Ltd, Shenzhen 518052, China

^* Corresponding author: Sunshuyin@mail.shiep.edu.cn

Abstract

This study proposes an approach that takes into account the impact of marine environmental impacts on the multi-dimension eigenvalues and analyzes the condenser status based on multi-dimension eigenvalues dynamic threshold. Firstly, we get the multi-dimensional eigenvalues of the condenser Based on thermodynamic theory analysis. Then, utilizing seawater-temperature and seawater-level as environmental parameters, the partitioning of the condenser marine environmental region is optimized based on the CalinskiHarabasz index. Subsequently, the multi-dimension eigenvalues are used to calculate to get the multidimension eigenvalues dynamic thresholds and mean corresponding to the marine environmental regions by improved Gaussian mixture algorithm (GMM). Then, the multi-dimension eigenvalues of the condenser are used to construct the spatial state analysis model of the condenser by improving Euclidean distance and cosine distance. Real time multi-dimension eigenvalues and normal means corresponding to the marine environment region are utilized for spatial quantification analysis of the condenser running status, the running status’s average during the sliding time window is used as the standard to determine whether the condenser running status is abnormal. Dynamic threshold and spatial state model calculations are performed using real data under different marine environment regions, its demonstrated that the proposed method accurately identifies the condenser running status.