Neural Modeling and Control of Diesel Engine with Pollution Constraints

摘要

The paper describes a neural approach for modelling and control of aturbocharged Diesel engine. A neural model, whose structure is mainly based onsome physical equations describing the engine behaviour, is built for therotation speed and the exhaust gas opacity. The model is composed of threeinterconnected neural submodels, each of them constituting a nonlinearmulti-input single-output error model. The structural identification and theparameter estimation from data gathered on a real engine are described. Theneural direct model is then used to determine a neural controller of theengine, in a specialized training scheme minimising a multivariable criterion.Simulations show the effect of the pollution constraint weighting on atrajectory tracking of the engine speed. Neural networks, which are flexibleand parsimonious nonlinear black-box models, with universal approximationcapabilities, can accurately describe or control complex nonlinear systems,with little a priori theoretical knowledge. The presented work extends optimalneuro-control to the multivariable case and shows the ?exibility of neuraloptimisers. Considering the preliminary results, it appears that neuralnetworks can be used as embedded models for engine control, to satisfy the moreand more restricting pollutant emission legislation. Particularly, they areable to model nonlinear dynamics and outperform during transients the controlschemes based on static mappings.