Optimization of non‐steady‐state operation of reactors

Non‐steady‐state operation of reactors is approached from the optimal control point of view. To find the best type of input, dynamic programming is used. The results show that for some systems, the optimal input is oscillatory in nature as has been found experimentally by others. A modification to conventional dynamic programming is used to avoid the need for interpolation and thus allow high‐dimensional systems to be handled more readily. Three chemical engineering systems of different complexity were considered to test the computational aspect of the approach. In all cases, dynamic programming gives the optimum policy with respect to the period, cycle‐split and amplitude.