Research on the Optimal Output Power of Proton‐Exchange Membrane Fuel Cell Based on Improved Model Predictive Control Strategy

To achieve optimal power output of proton‐exchange membrane fuel cell (PEMFC) under load variation and noise disturbance, this article proposes a recursive subspace model predictive control based on Laguerre function (RSMPCL) strategy. First, a fractional PEMFC identification model is established, and an optimal power setter is introduced to provide an optimal power value corresponding to the current load. Second, during the predictive process, a subspace predictor is updated in real time where a Householder transformer is applied to decrease the computation time of the frequent QR decomposition, and a variable forgetting factor is introduced to enhance the sensitivity and accuracy of predictive control. In the control process, a Laguerre function is employed to transform the problem of computing the input increment at the next step into calculating the coefficients of the input variables. This transformation reduces the computation time of the control quantities. Finally, load and disturbance tests conducted on a dSPACE semiphysical simulation platform indicate that compared with constrained subspace model predictive control, the proposed RSMPCL strategy can help the fuel cells achieve the optimal power point with control accuracy, robustness, and calculation speed. The optimal power setter is designed to obtain the optimal output of the system under the influence of a certain load current and temperature. The recursive subspace model predictive control based on Laguerre function strategy proposed in this study is able to track the optimal output performance well in controlling the fuel cell output, which improves its response speed and robustness.