NEURO-FUZZY CONTROL OF THE LUMBER DRYING PROCESS

Abstract

Background. The problem of control over the process of lumber drying is considered. The quality of drying is determined by the modes of operation of power plants that provide heat supply to the drying chamber and the parameters of the moisture content of the dried sawn timber. Recently, in many works, the process of drying sawn timber is considered as an optimal control problem, in which the material to be dried must achieve the specified state by its properties in a minimum time. Materials and methods. To determine the modes of high-quality optimal control and effective change of these modes in the process of drying control, it is necessary to have at each moment of time the exact values of the parameters of the model of the controlled object. These values cannot be accurately determined using measuring instruments. Results and conclusions. Thus, the process of optimally managing the drying of lumber involves uncertainties. To eliminate the problem of uncertainties in the work, it is proposed to use the mathematical apparatus of fuzzy sets to describe them, which, in the process of fuzzification of variables, will translate the undefined values of the model parameters into linguistic terms with certain membership functions. To obtain control actions based on the analysis of linguistic variables, it is proposed to use a neuro-fuzzy control system with Tagaki– Sugeno–Kang logical inference based on the ANFIS neural network, which implements optimal control of sawn timber drying based on the rule base set by the developers of the control system.

Keywords

• optimal control
• lumber drying
• neuro-fuzzy control