GRU-based model-free adaptive control for industrial processes

Many industrial processes have the characteristics such as large time delay, strong coupling or nonlinearity, which make them difficult to control. When exact process models are available, model-based control methods can achieve good performance. However, it is difficult to satisfy in many cases. This paper proposes a novel Model-free adaptive control scheme with an encoder–decoder structure based on gated recurrent unit (GRU) network and attention mechanism. The control objective is to make the process output track a known reference input. The controller does not need to know an accurate process model and can adjust the weight of each neuron of the neural network according to the error signal to achieve process control. The gating mechanism of GRU neural network enables the controller to take full advantage of the system’s history information. Lyapunov-based stability analysis is provided to guarantee the stability of the whole control system. Some process simulations and a Wood/Berry distillation column example show that only by adjusting a few parameters, the proposed controller can control a multivariable process with coupling or large time delay well.