UTILIZING BP NEURAL NETWORK APPROACH FOR DYNAMIC PARAMETER ADJUSTMENT IN DC ELECTRONIC LOAD SYSTEMS

Abstract

As the demand for electrical equipment in the era of new energy continues to surge, rigorous aging tests are essential prior to equipment deployment. Traditional performance testing methods are often susceptible to external factors, particularly temperature fluctuations, limiting their effectiveness. To address these challenges, DC electronic loads have gained prominence due to their compactness, portability, precision, and user-friendly attributes. This paper focuses on the constant current mode of DC electronic loads, which presents control challenges, including overshooting, in the presence of system interference and delay. The application of PID (Proportional-Integral-Derivative) control to DC electronic loads has shown promise, significantly enhancing control flexibility and maintaining accuracy within 0.1%. However, conventional methods for PID parameter setting are limited in their ability to adapt dynamically to real-world scenarios.

To overcome these limitations, this study employs a Backpropagation (BP) neural network to perform online, dynamic adjustments to PID parameters. Through extensive simulation analysis, the BP neural network PID control approach is demonstrated to be highly effective in enhancing system response speed, stability, and reducing overshooting. This innovative solution represents a significant step forward in optimizing the control of DC electronic loads for reliable electrical equipment testing