电工技术学报2024，Vol.39Issue(23)：7319-7330,12.DOI:10.19595/j.cnki.1000-6753.tces.231208
换流阀冷却系统均压电极垢层沉积声学检测方法
Study on Acoustic Detection Method for Sediment on Grading Electrodes in Converter Valve Cooling System
摘要
Abstract
Sediment on grading electrodes is one of the main causes of failures in converter valve cooling systems.Traditional maintenance methods mainly involve periodic disassembly and inspection,which are quite blind and can lead to issues such as system leaks from repeated disassembly,severely delaying the maintenance progress of the converter valve system.Most existing research relies on water quality monitoring or other indirect detection methods,which cannot directly reflect the sediment situation.Acoustic detection,however,offers advantages such as high sensitivity,fast detection speed,and strong versatility.Therefore,this paper proposes an acoustic detection method for sediment on grading electrodes in converter valve cooling systems. Firstly,based on the principles of acoustic detection for grading electrode sediment,an acoustic detection device was designed.This device achieves synchronized control of the sound wave excitation and reception ends,utilizes cumulative averaging for noise reduction in digital signal processing,and transmits the echo signals to a PC for feature extraction and analysis.Secondly,an acoustic detection platform for grading electrode sediment was built.This platform verifies that the designed device can effectively drive the transducer and receive echo signals from the grading electrode and its sediment.It also tests the device's interference resistance in noisy environments,confirming its ability to reliably filter out low-frequency noise,high-frequency spurious signals,and random spikes,thus obtaining clean echo signals.The platform was used to perform acoustic detection on seven groups of grading electrodes with different sediment thicknesses,analyzing the differences in echo signals under sedimented and non-sedimented conditions. Further,using waveform factors(Sf),peak factors(Ip),kurtosis factors(K4),pulse factors(Cf),peak factors(Ce),and skewness factors(K3)as characteristic indicators for echo signals,it was found that Sf,Ip,K4,Cf and Ce had a coefficient of variation of less than 10%in the non-sedimented condition,indicating good stability.In the presence of sediment,these factors decreased with increasing sediment thickness,while K3 was generally negative without sediment and positive with sediment.Finally,by calculating the variation rates of echo feature values in relation to the non-sedimented conditions,it was found that ΔSf,ΔK4,ΔCf and ΔCe had a strong correlation with the sediment thickness,with the Boltzmann fitting function's determination coefficients(R2)exceeding 0.96,indicating that these measures can reliably characterize changes in electrode sediment thickness.Based on the sensitivity relationship between feature quantities and sediment thickness,a prediction strategy for sediment thickness was proposed,effectively predicting sediment thickness at different stages of sediment. From the experimental analysis,the following conclusions can be drawn:(1)The designed acoustic detection device for grading electrodes can detect sediment and has advantages of strong portability and interference resistance,showing good application potential in practical valve cooling system inspections.(2)Among the used characteristic indicators,K3 can be used for qualitative analysis of sediment presence,while Sf,Ip,K4,Cf and Ce are suitable for quantitative analysis of sediment thickness.(3)ΔSf,ΔK4,ΔCf and ΔCe show a strong correlation with the sediment thickness,and the Boltzmann fitting function effectively characterizes the thickness change process,providing guidance for the maintenance of grading electrodes.关键词
换流阀冷却系统/均压电极/垢层厚度/声学检测/特征指标Key words
Converter valve cooling system/grading electrode/sediment thickness/acoustic detection/characteristic indicators分类
信息技术与安全科学引用本文复制引用
高兵,黄驰,黄晨浩,王帅,杨明智..换流阀冷却系统均压电极垢层沉积声学检测方法[J].电工技术学报,2024,39(23):7319-7330,12.基金项目
国家自然科学基金资助项目(52007013). （52007013）
