The Research On On-line Thickness Measurement Of MDF Based On Artificial Optimized ELM

In the process of producing medium density fiberboard(MDF),the determination of the thickness value is of important meaning for stabilizing the basic physical properties of the fiberboard,reducing production energy consumption,and producing high-quality MDF.In the application of traditional MDF inspection system,the processing technology is hard to satisfy the dependability requirements,high efficiency,and automation.With my country’s revolutionary breakthroughs in artificial intelligence,bioinformatics,material manufacturing and other key technologies and their cross-integration have triggered a new round of industrial reform,and a new generation of information technology and manufacturing have been deeply integrated.Extreme Learning Machine(ELM)has superiority such as predictability,high precision,fast learning and calculation speed,and few parameters.Therefore,it has been widely used in mathematical calculations and system modeling in recent years.In this paper,extreme learning machine as the major approach,combined with the group intelligence optimization theory,research and practice of MDF thickness online detection method.The research mainly includes the following aspects:First of all,it briefly introduces the basic composition and control theory of the MDF thickness detection system,and stipulates the corresponding detection standards.The system meets the overall structural requirements and technical characteristics of the modern manufacturing process control system.Using the built detection system,a specific experimental plan for MDF thickness detection was designed,and three different thicknesses of MDF were selected as experimental materials to detect the actual thickness,and the measurement data was collected.Because there are many issues that influence the measurement results in the industrial production process,it is essential to enhance the veracity of the measurement results and carry out corresponding preprocessing to remove abnormal,secondary and redundant data.The particle filter algorithm is optimized by Harris Hawks Optimization(HHO),and the original data is preprocessed to eliminate abnormal data.Based on the preprocessed thickness data,the ELM regression method is used to compensate the detected thickness data.On the basis of the ELM error compensation model,a HHO algorithm improved by Tent mapping and opposition-based learning strategy is introduced.Named IHHO,it selects the input layer weights and hidden layer thresholds randomly generated by ELM,which overcomes the shortcomings of ELM easy to fall into local optimal solutions,retains its optimization mechanism on the basis of improving the algorithm performance,and the good performance of the online MDF thickness detection model is verified through comparative experiments.In summary,the MDF thickness online measurement system based on the intelligent algorithm optimized ELM developed in this paper runs stably and has high detection accuracy.The conclusions drawn have certain practical engineering reference value for MDF production and manufacturing.