IMPROVING THE OPERATIONAL RELIABILITY OF A FRANCIS TURBINE EQUIPPED WITH ANONLINE MONITORING SYSTEM

Hydropower plants play a key role in stabilizing the operation of the power grid. However, restoring network stability forces plants to constantly change operating modes. As a result, equipment components are subjected to atypical cycles of fatigue loads and vibrations, which provoke defects in the units and shorten their service life. Traditional methods of technical condition control (offline monitoring) often prove ineffective for continuous tracking of system operability. This problem can be solved with the help of online monitoring systems, which continuously record signals and provide information on the current state of the machine. Key points: 1. For strain gauges, the finite element method (FEM) is used to determine the optimal installation location; 2. Vibration monitoring is an effective method for detecting malfunctions in hydraulic machines; 3. Pressure sensors are usually installed in the penstock, spiral case, on blades, draft tube, and in the stay-vane-less region; 4. A non-destructive testing (NDT) method such as acoustic emission (AE) is based on measuring elastic waves that arise due to the initiation and growth of defects in the material; 5. Real-time monitoring of defect initiation and growth processes is a challenging task for hydro units. The physical parameters for monitoring the condition of hydropower plants, including those at the Dnipro and Dniester cascades in Ukraine, are recorded by a system of appropriate sensors and converted into electrical signals. Factors such as sensor sensitivity, measurement frequency, and signal attenuation characteristics significantly affect the quality and reliability of data collection. Signals from sensors may be distorted by numerous factors. To extract relevant information from the input signal, specialized tools and methods are applied. Reliable machine learning algorithms for defect detection can become an additional tool for supporting predictive maintenance, which is highly relevant, as it allows eliminating up to 75% of cases of equipment failure, reducing downtime by 30%, and significantly cutting maintenance costs.