The authors present a multiobjective optimisation design approach to improve the performance of the ALSTOM Benchmark Challenge baseline controller. As the gasifier process is complex and non-linear, with a high degree of cross coupling of the variables, manual tuning of the controllers is difficult. The use of a multiobjective optimisation method allows the simultaneous tuning of multiple controllers, seeking a trade-off between the different performance objectives. A further example using the same multiobjective optimisation method shows how the performance of the baseline controller can be improved by the addition of extra proportional controllers to reduce the fluctuation due to a sine-wave pressure disturbance. This control structure is then simplified using a genetic algorithm search procedure, to find the structure with the lowest error.
[1]
A. Farag,et al.
MULTIVARIABLE PID-CONTROLLER DESIGN FOR A GASIFIER PLANT USING PENALTY-BASED-MULTI-OBJECTIVE GA
,
2004
.
[2]
Andrew Pike,et al.
Alstom Benchmark Challenge II on Gasifier Control
,
2006
.
[3]
S. Daley,et al.
Optimal-Tuning PID Control for Industrial Systems
,
2000
.
[4]
Kenneth Alan De Jong,et al.
An analysis of the behavior of a class of genetic adaptive systems.
,
1975
.
[5]
David E. Goldberg,et al.
Genetic Algorithms in Search Optimization and Machine Learning
,
1988
.
[6]
Thomas F. Coleman,et al.
Optimization Toolbox User's Guide
,
1998
.