Embedded digital system design for neural network based vector control of a single-phase inverter

Single-phase inverters are widely used to connect small-scale distributed renewable energy sources to the grid. However, unlike a three-phase system, designing the vector control for a single-phase inverter is more challenging. This paper presents a neural network based vector control technique for a single-phase inverter, as well as how to develop and design an embedded digital control system to implement the neural network vector controller by using Texas Instruments TMS320F28335 chip. The paper addresses circuit design for effective interface between the high power inverter system and the embedded microcontroller as well as how to utilize built-in interrupts to achieve efficient digital control implementation, with user specified sampling time, and generation of high-quality pulse-width modulated driving pulses. Moreover, an embedded software design was developed to implement the neural network vector control for the single-phase inverter. A laboratory scale single-phase inverter system was built to test and evaluate the embedded neural network microcontroller. The hardware experiment assessment shows that the microcontroller works well using the novel neural network vector control mechanism in challenging, variable, and distorted hardware experiment conditions.