Innovative Diagnostic Methods in The Control of The Combustion Process

The purpose of the scientific article is to develop a model control circuit for complex combustion control. Two approaches to neural adaptive control have been developed and tested successfully. The considered circuit takes into account the multi-input and multi-output nature of the combustion process. Using complementary information, optical signals based on descriptor vectors for flame surface area and contour length proved the tracking properties of the system. The scientific novelty of the article is that two MRAC systems have been developed and compared. The first of these used a non-optical, measurement-based set of input vectors, respectively quantifying the secondary air flow, fuel flow and vectors describing respectively the chamber exhaust temperature recorded at the first measurement point. The second circuit uses the secondary airflow control signal and selected flame image descriptors.