Application Of BP Neural Network In New Product Development

This paper is based on the research method of new product development.Taking the AOTAI electric welding machine as an example,quantify and extract the perceptual images of users through the guidance of KANSEI theory,and use the advantage of BP(Back Propagation)neural network in dealing with non-mapping(non-linear)capabilities to establish a corresponding relation model between perceptual imageries and product design elements,to explore a reasonable quantification method of fuzzy perceptual image of users and establish an intelligent measurement model;using MATLAB neural network toolbox as the platform to conduct research on new product development and build a BP neural network for the development of new products of electric welder.In the new product development stage to predict the market results of new products,and facilitate the development of products that are more humanized and meet the needs of users.The research is based on the complex connections between neurons in BP neural network and the nonlinear way of transmitting signals in each neuron,and various relationships can be constructed between input and output signals,so it can be used as a black box model.The model and advantages of BP neural network and the reasons why it can be applied in the development of new products are analyzed and summarized.And in the whole new product development process,there is also a fuzzy black box process,so the BP neural network can be applied to the new product development process.At present,BP neural network has been applied in a variety of product form design processes,but it is still a blank in the field of electric welding machine research and application.As the electric welding machine is a specialized material welding tool,the usershave a clear and deep understanding on the long-term used welding machine.The form,color,and material of the product are clear,and the user information is representative.And there are many kinds of similar products,the information and data in the product sales are complete,so MATLAB neural network toolbox development platform can be used to build BP neural network model.Taking the known products as samples,through the computer autonomous learning,build the neural network model between input parameters and product design results,and form intelligent judgment.Finally,the trained BP neural network is used for the prediction of new product development factors and compared with the actual data.Within a certain range of errors,it can meet the needs of actual product development and production,and used to instruct the new product development of AOTAI welding machine.