Optimization for response time of solenoid valve through response surface methodology

Abstract: Solenoid valve driven by PWM (pulse width modulation) control signal is used as interface between electronic control and fluid flow in precision spraying, the response time of which has huge impact on the performance of precise spray system. Longer response time would reduce the accuracy of spraying and increase the overuse of chemical. A modified PWM technology (with much higher frequency for controlling the holding current) was applied and 3 parameters of PWM control signal containing driving voltage, delayed time and duty cycle were optimized by RSM (response surface methodology) to shorten the response time, which was composed of opening response time and closing response time. The optimization of parameters was carried out through the BBD (Box-Behnken design) with 3 factors and 3 levels. The critical factors (and their values) selected for the research were driving voltage (10, 12 and 14 V), delayed time (15, 40 and 65 ms) and duty cycle (5%, 15% and 25%). In order to better explore the influence of factors on each response phase of solenoid valve, the opening response time, closing response time and response time were discussed separately. The experiment was performed on an optimization experiment platform which consisted of human-machine control module, pressure supply module and PWM module. On the platform, the wave data that could reflect the relation between voltage of PWM control signal and instantaneous spray pressure near the nozzle were acquired through digital oscilloscope and then processed by Kalman filter. By analyzing the wave data, the opening and closing response time of solenoid valve with different parameters were calculated. After 15 runs of experiments under different parameter conditions, the mathematical regression models for the opening response time, closing response time and response time were built, respectively. Then the models were tested by the analysis of variance and statistical parameters of the model for response. The test indicated that the relations between response variables and independent variables were significant and the regression models were thought to be appropriate. Through the analyses of 3 tested models and response surfaces showing the effects of process parameters on response time, the effects of driving voltage, delayed time and duty cycle of PWM signal on 3 kinds of response time of solenoid valve were obtained. Firstly, the voltage had a significant effect on all response models and a positive effect on reducing response time. Secondly, the delayed time had a negative effect on reducing response time and the effect it brought was not so significant; in addition, longer delayed time could accelerate the opening phase but detain the closing phase. Thirdly, the duty cycle had a negative effect on reducing response time and had more remarkable influence on the closing response time than opening response time. Lastly, the interaction between driving voltage and duty cycle on closing response time was significant. After that, the optimal control parameter combination for the minimum response time of solenoid time was obtained by analyzing the regression model of response time, which was 12 V voltage for driving solenoid, delayed time of 15 ms and duty cycle of 5%. The predicted optimal control parameters were tested in the laboratory using the modified PWM control method and normal PWM control method which had stable voltage, respectively. The results showed that the measured value by the modified PWM control method was very close to the predicted value and 21.2% of response time could be reduced compared to the normal control method. The result indicates that the RSM is useful for optimizing the parameters of modified PWM control signal to improve the response characteristics of solenoid valve.