RBF Neural Network Adaptive Control of Microturbine

Microturbine generator (MTG) system is a clean, efficient, low cost and novel energy supply system and it is widely used in a variety of power generation and industrial applications. MTG dynamics are often complex and vary with operating point and ambient conditions, although the relationship of input-output is very simple, therefore conventional control law adopted cannot achieve expectant result. A novel single neuron adaptive control algorithm is proposed in combining with PID, based on radial basis function (RBF) neural network on-line identification. Through adjusting control parameter on-line, excellent flexibility and adaptability as well as high precision and good robustness can be achieved. The algorithm has been applied in "100kW microturbine control and power converter system". The results of simulation are shown that the algorithm is very valid.