Process requirements-based adaptive PWM for improved efficiency of machine tool feed drives

Feed drives of machine tools are usually controlled by cascaded PI controllers combined with Pulse Width Modulation (PWM). Based on the strictest performance and accuracy requirements, the drive control system is configured, often necessitating a high PWM-frequency. While improving the control bandwidth, the energy efficiency and lifetime of the power electronic decrease significantly due to switching losses. Analyzing the production process performed by the machine tools reveals that the requirements must only be met in critical process states, e.g. when the final surface is manufactured. Much laxer requirements are acceptable in other states. Therefore, a significant amount of energy could be saved if the minimal required PWM frequency would be used in each system state. To exploit this unused potential, an algorithm deciding on the frequency and a control system allowing to change the frequency during runtime are prerequisites. Both are addressed by the presented research. To decide on the frequency, an algorithm based on process knowledge is proposed. The key challenge of a control loop with a non-constant frequency are the parameters. An efficient way to derive these is presented. The validity of the parameter calculation as well as the energy savings are experimentally validated.