Gas Turbine Compressor Wash Performance Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for cleaning gas turbine compressors, such as water wash and hand cleaning, lack predictive capabilities, making it difficult to accurately assess performance improvements and power recovery, especially when grime is bonded to compressor parts.
Innovation Solution
A forecasting model is developed using sensor data from before and after wash operations to predict performance improvements and risk levels, accounting for statistical dispersion in results, allowing for better decision-making in maintenance operations like hand cleaning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If water wash or hand cleaning operations are performed on gas turbine compressors, then contaminant removal and performance recovery are improved, but the ability to accurately predict performance improvement and power recovery is degraded due to lack of predictive capabilities
Solution Approach 1:
The system performs preliminary actions by collecting sensor data before wash operations and using it to create forecasting models that predict performance improvement. This preliminary data collection and model creation enables accurate prediction of power recovery and performance gains before the actual wash operation occurs, resolving the contradiction between achieving performance recovery and maintaining predictive capability
Solution Approach 2:
The system implements feedback by continuously collecting sensor data from turbine operations, comparing actual performance against predicted performance from forecasting models, and using this information to improve future predictions. The feedback loop maintains predictive accuracy by incorporating real-world wash operation results into the forecasting models, addressing the lack of predictive capabilities while ensuring reliable performance recovery assessment
2Measurement precision
If comprehensive sensor data collection is performed before and after wash operations, then forecasting model accuracy is improved, but system complexity and data processing requirements increase
Solution Approach 1:
The system applies universality by using existing operational sensors for multiple purposes: monitoring turbine performance during normal operation and collecting data for wash operation forecasting. This multi-functional use of sensors improves forecasting accuracy without adding dedicated complex measurement equipment, as the same sensors serve both operational monitoring and predictive modeling functions
Solution Approach 2:
The system manages complexity by focusing on key performance parameters rather than collecting all possible sensor data. The forecasting models use selected critical parameters from sensor collections to achieve accurate predictions, transforming large datasets into meaningful performance indicators that maintain model accuracy while reducing data processing burden
Data Source
AI summary
In one embodiment, a method includes sensing first operations for one or more turbine systems in a fleet of turbine systems via a plurality of sensors disposed in the one or more turbine systems before a first wash operation. The method further includes sensing second operations for the one or more turbine systems via the one or more sensors after the water wash operation. The method also includes deriving at least one forecasting model based on the sensing first operations and the sensing second operations, wherein the at least one forecasting model is configured to predict a performance of a turbine system of the one or more turbine systems. The method additionally includes applying the at least one forecasting model to derive a predictive improvement in the performance for the turbine system.