MODELING OF A PEM FUEL CELL POLARIZATION CURVE BY LOW-ORDER POLYNOMIALS FOR THE OUTPUT POWER CALCULATION ALGORITHMS

Abstract

The control and energy algorithms that govern the operation of a PEM fuel cell within an energy system must account for the nonlinear phenomena in the cell; it is often captured by the polarization curve. However, the modeling approaches used for simulation studies could be more suitable for the abovementioned algorithms due to their higher computational burden. That leads to the development of simplified approaches. Our study focuses on the algorithms used for the fuel cell output power calculation. Many of them are directed towards the maximum power point operation of the cell. We propose that the fuel cell representation for the algorithms can be used in the low-order polynomial form, thus decreasing the computational load compared to the other approaches. For the maximum power point prediction, we propose using the truncated form of polarization curve input data (ignoring the activation loss region). Our study has demonstrated that based on the same input data and MATLAB curve fitting commands, the 3rd-order polynomial provides the comparable RMSE 3.5…3.9 for the power curve approximation vs. 3.9 for the 5th-order polynomial representation. The value of the maximum PowerPoint is obtained with a 1.3 % relative error with the 2nd-order polynomial using the truncated input data compared to 1.2 % for the 5th-order polynomial representation.