A correlative approach to the integration of diagnostic features of gas piston electric units
Abstract
RELEVANCE. Gas-piston electric units with synchronous generators with excitation from permanent magnets are widely used in the oil industry and in small distributed power engineering. Continuous technical diagnostics are an effective means of increasing their operational readiness and reliability. The article examines the use of correlation dependencies of diagnostic features to increase the reliability of multiphysical diagnosis. THE
Purpose
To consider the possibility of reliable classification of defects in gas-piston electric units with synchronous generators with excitation from permanent magnets based on the analysis of correlation coefficients between vibration, voltage and magnetic flux. To investigate the use of the correlation approach as part of the digital shadow machine for integration into a single system of multiphysical diagnostics.
Methods
When solving this problem, mathematical modeling methods were used using Google Colab, PlantUML computer software packages, and object-oriented programming using the Python programming language.
Results
Experimental testing of the use of the correlation approach as part of the digital shadow on reference and real data confirmed the high sensitivity of the system to detect both single and combined faults. It has been established that this approach provides early diagnosis and increases the reliability of maintenance decisions.
Conclusion
The combined accounting of spectral and correlation characteristics within the digital shadow ensures strict, physically interpretable and statistically stable fault identification, which is confirmed by both computational experiments and empirical studies. This approach provides high sensitivity to complex failures. The results obtained confirm the expediency of using correlation analysis as a key element in the architecture of digital shadows and doppelgangers of multi-signal energy facilities.