Application of wavelet transformations on microcontrollers for monitoring and optimising energy systems in industrial conditions

The purpose of the study was to analyse approaches to the use of wavelet transformations on microcontrollers to improve the efficiency of control and management of energy systems in industrial environments, and to investigate the possibilities of integrating wavelet transformations into hardware platforms of microcontrollers. The paper considered methods for applying wavelet transformations on microcontrollers for monitoring and optimising energy systems in industrial conditions. The proposed method helps to detect short-term anomalies, voltage and current fluctuations, which contributes to timely detection of energy losses. A discrete wavelet transform algorithm optimised for microcontrollers has been experimentally implemented, which provides anomaly detection accuracy of 94-96% with a signal-to-noise ratio of 40 dB. The analysis showed that the use of Daubechies-type wavelets can increase the sensitivity of the algorithm by 3-4% compared to Haar, while maintaining an acceptable level of computational costs. Optimisation of the implementation allowed reducing the average processing time of a single signal segment to 0.1 seconds on the STM32 microcontroller. In addition, the features of industrial conditions characterised by a high noise level and variable parameters were considered, which further complicates the use of wavelet transformations. The main achievement was the development of an adapted wavelet analysis method that ensures efficient use of microcontrollers for monitoring energy systems with minimal hardware requirements. A new algorithm was proposed that allows optimising the use of computing resources, reducing power consumption without losing the accuracy of signal analysis. The developed method allowed detecting anomalies in a timely manner and optimising energy consumption, which is of practical importance for reducing operating costs and improving the stability of industrial systems