DESIGN AND MODELLING OF AN AUTOMATED SYSTEM FOR QUALITY CONTROL IN FLEXIBLE POLYURETHANE FOAM PRODUCTION

Abstract

Flexible polyurethane foam is one of the classes of cellular plastic from mixture of different chemicals, this cellular plastic are used in producing materials; such as pillows, furniture foam mattresses, carpet underlay, cushioning material used in automobile. Due to the larg area of its application, production in large quantity and maintaining quality has become a huge task for the producers. Preproduction processes is cumbersome and most times result inhuman error which affects the quality of foam to be produced. Controlling the rate of flow of each of the chemical (proportion of mixture) is one of the great tasks in producing flexible polyurethane foam because variation from the set mixture proportion brings about change in the quality and increased wastage. This work is on modelling of an automated system that generates the chemical formulation needed for a specified foam density and foam mass to be produced. The model consists of three blocks which are determination of chemical proportion and rate of flow, flow control loop development and performance evaluation. The model generate the proportion of mixture of chemicals and uses it to determine the rate at which each chemical should flow base on the production time with the required pump speed. The generated rate of flow serves as input to the flow control loop developed. The flow control unit was modelled using a PID controller and the pump was modelled using a form of Riccati equation and a flow meter was used as the feedback element. Two sample t-test was used to compare the output of thefirst stage of the model with real life data, the result obtained shows no significant difference as t statistic for polyol, TDI and water were less than the t critical value of 2.22814. Time and frequency domain analysers were used to examine the stability of the system. The performance and robustness of the model with PID controller shows 0.307 seconds rise time, settling time of 1.18 seconds and overshoot of 3.66%.The Nyquist plot for the flow control loop was stable as the plot crosses the real axis in the left hand plane of the plot and there was no encirclement at -1 + j0 point.