ESTABLISHING CRITERIA TO ENSURE SUCCESSFUL FEEDFORWARD ARTIFICIAL NEURAL-NETWORK MODELING OF MECHANICAL SYSTEMS

The emulation of mechanical systems is a popular application of artifi cial neural networks in engineering. This paper examines general princ iples of modelling mechanical systems by feedforward artificial neural networks (FFANNs). The slow convergence issue associated with the hig hly parallel and redundant structure of FFANN systems is addressed by formulating criteria for constraining network parameters so that FFANN s may be reliably applied to mechanics problems. The existence of the FFANN mechanical model and its stability during construction, with res pect to the error in the data, are analyzed. Also, a class of differen tial equations is analyzed for use with Tikhonov regularization. It is shown that the use of Tikhonov regularization can aid in FFANN data-d riven construction with a priori mathematical models of varying degree s of physical fidelity. Criteria to ensure successful FFANN applicatio n from an engineering perspective are established.