웨이블릿 에너지 스펙트럼 기반의 관리도를 통한 화력발전소 발전기의 상태기반보전 방법에 대한 연구

Abstract

Purpose: Failing power-plant equipment can cause significant harm to both human life and expensive infrastructure. Hence, both effective process-monitoring and failure-prediction techniques are needed to prevent catastrophic equipment failure. Existing time-based maintenance-methods are not enough to prevent catastrophic failures. In this work, a condition-based maintenance (CBM) method, which can predict failures using wavelet-based process-monitoring methods, is studied. Methods: Both a wavelet spectrum analysis (with coefficients extracted from the wavelet transform) and a  chart (with a slope and intercept based on a linear trend profile) are used in this study. The results are then embedded into a  chart. Results: The proposed methodology in this study is validated using data signals coming from thermoelectric power plants. Overall, the proposed monitoring method has a higher prediction performance using breakdown signals than the traditional  chart-based maintenance method. Conclusion: The investigated failure-prediction method is not only more successful but also more specific when abnormal signal patterns are detected before failures. Therefore, using the proposed method for condition-based maintenance can help save resources and prevent human losses.