Innovative Web System for Condition-Based Maintenance of Generators

A novel integrated diagnostic system has been developed at IREQ, the research institute of Hydro-Quebec, over the last eight years. The goal of this project was to build an integrated methodology for generator diagnostics using the results from on-site measurement toots, which would help the utility to make the transition from time-based maintenance to condition-based maintenance. The system makes use of information technology (Internet) to provide a new, modern and efficient way to produce a continuous classification of the condition of all generators of a fleet along with individual diagnostics for any unit at any time. The system computes actual measurements provided by plant personnel who transfer data files to a centralized server. The application calculates simple condition indexes from each of six on-line/off-line diagnostic tools and from visual inspections, and aggregates the results into a comprehensive global diagnostic for the selected generator. Plant and generator selection is done via a user-friendly interface displaying a simple rating of the results for every tool. The algorithm underlying the system generates a global diagnostic for any combination of tools, regardless of their number and selection. However, the level of confidence of the diagnostic will increase with the number of tools used for the diagnostic. In addition to the simplified integrated condition index values of an generators and the individual index for each tool, specialists can access and display the complete data for every measurement series. Each tool, some commercial, others developed at Hydro-Quebec, was selected based on its ability to characterize specific complementary aspects of the generator. Since the system was developed with an expandable modular approach, it will be possible in the future to add new diagnostic tools, not yet considered in the current version, without affecting the logic of the system. Hydro-Quebec has recently implemented this new system and is already noting improvement in maintenance practices [1, 2]. The ready availability of centralized, simplified information makes it possible for generator specialists and managers alike to assess the condition of any generator in a few minutes. Thanks to this new diagnostic system, technical and management staff can now work together with common information and in real time to optimize maintenance intervention on generators showing degradation. Thus, it is possible to plan any corrective action more effectively or request additional testing when doubts remain about active degradation mechanisms. At the same time, efforts in diagnosis and maintenance could be optimized by reducing the number of measurement campaigns for the vast majority of generators that are in good condition as revealed by their condition indexes.