Model predictive control of gas turbine engines

摘要

A method of controlling a system comprising:(a) providing a system model M which is in explicit form and which predicts one or more system performance variables y from one or more system control variables u, such that y(k+1|k) = M ∘ u(k+1|k), where k is the present time,(b) at time k, measuring or estimating the present values of the performance variables y, and then using a deterministic algorithm to optimise a cost function:where:N is selected such that the cost function J is convergent,k+i is the time in the future after the passage of i of the time increments,u(k+i|k) and y(k+i|k) are restricted in the cost function J to values of the respective control variables u and performance variables y which are compatible with predetermined system operational constraints, andfor each time increment i = 1...N, values for F(u(k+i|k),y(k+i|k)) are calculated using the system model M to determine y(k+i|k) for a plurality of respective discrete values of u(k+i|k), the cost function J being optimised by identifying the combination {u(k+1|k),u(k+2|k),...u(k+N|k)} of respective discrete values u(k+i|k)(c) selecting values for the control variables u which correspond to u(k+1|k) in the identified combination {u(k+1|k),u(k+2|k),...u(k+N|k)};(d) at time k+1, controlling the system with the selected values for the control variables u; and(e) repeatedly performing steps (b) to (d), time k+1 becoming the new time k after each performance of step (d).