Research On Material Level Prediction And Control Of Sand Mill Based On Multi-sensor Information Fusion

Vertical sand mill is used for ink grinding.It is the key equipment in ink production.The material level of the mill affects the quality and production efficiency of ink.Therefore,it is necessary to precisely control the material level during the grinding process of the mill.In the process of ink grinding,due to the large vibration amplitude of the mill,it is not easy to directly detect the material level,so the traditional contact measurement method is not suitable for the detection of the material level of the mill.And the traditional material level control relies on the operator’s experience and requires manual adjustment,so that the material level in the mill cannot be stably maintained at a reasonable height.This requires a more intelligent material level prediction and control method to ensure a reasonable height of the material level.This paper proposes a method of using multi-sensor information fusion to predict the material level and control it through fuzzy PID,which improves the production efficiency of ink on the basis of improving the quality of ink production.In the process of material level prediction,multi-sensor soft sensing technology is used to detect pressure,temperature and current related signals,and wavelet transform is used to perform three-layer decomposition to obtain eigenvalues as the input of information fusion.Considering the excellent performance of BP neural network in dealing with nonlinear problems,a three-layer BP neural network model was constructed to predict the internal material level of the sand mill.BP network is easy to fall into local optimum when searching for optimization,while genetic algorithm and particle swarm algorithm can improve its searching ability by improving the parameters of network weights and thresholds.The results of the two optimization algorithms are compared,and the genetic algorithm is finally selected to construct the GA-BP model for material level prediction.In the process of material level control,the material level is guaranteed to fluctuate within the optimal state range through forward-looking control.The deviation of the predicted material level and the optimal material level and the change degree of the deviation from the multi-sensor information fusion are used as the input of the fuzzy control,and the decision-making is inferred according to the control rule table obtained by the operator’s experience,and the output change of the PID gain is obtained.The initial value gets the final PID gain value.The frequency of the frequency converter calculated by the PID algorithm is used to control the speed of the feed pump of the mill system to maintain a reasonable height of the material level.This paper designs the mill system according to the functional requirements of the system,selects the controller,touch screen and other equipment,and finally builds an experimental platform to verify the feasibility of the control strategy.The lower computer adopts S7-200 Smart PLC as the controller,completes the configuration and module design on the corresponding platform,realizes the control of hardware equipment such as lubricating motor,mill,feed pump,discharge pump,cooling water pump,etc.,and receives Data from sensors such as current,temperature,pressure,etc.The upper computer uses Win CC to design the human-computer interaction interface for real-time monitoring and control of the field status.The experimental results show that when the material level is abnormal due to external interference or internal system wear,the system will automatically adjust to keep the material level dynamically at a reasonable height and achieve optimal control of system parameters.