Multi-layer Modeling,Assessment And Optimal Decision For Mechatronic Product Low Carbon Design

Greenhouse gas emissions are the main cause of the increasingly serious global climate problem and how to reduce emissions has become a common challenge facing human society.According to statistics,carbon emissions from the manufacturing account for more than 80%of the country’s total emissions,which have a significant impact on global warming and climate change.Made in China 2025,the National 14th Five-Year Plan,carbon emissions peak and carbon neutrality target,etc.,put forward higher requirements for the green development of manufacturing industry.China has the world’s largest import and export volume of mechatronic products,such as automobiles,construction machinery,energy equipment and so on.Mechatronic products are characterized by various types,large amounts,large manufacturing consumables,and high energy consumption.In recent years,with the release of regulations and rules related to product environmental footprint declaration in developed countries,more than 10%of export products will be involved in green trade barriers.The total amount of green trade barriers is up to 300 billion dollars every year,posing a huge challenge to Chinese enterprises,and reducing the environmental impact of mechatronic products is of great significance to the development of enterprises.Nearly 80%of the product’s performance and its impact on the ecological environment are determined during the design and development stage.Therefore,integrating low carbon performance requirements into product design and development stage and carrying out low carbon design of mechatronic products have strategic importance for the low carbon development of the manufacturing industry.At present,the low carbon design research of mechatronic products still has the following problems:(1)The product life cycle design scenario is diverse and the correlation is complex,and the lack of integrated expression of the whole life cycle design information limits the product life cycle design;(2)The information in product design stage is relatively lacking,and the inventory data is difficult to collect,which cannot support the assessment of carbon footprint of product design schemes;(3)The optimial decision of the low carbon design scheme is a process of constraint instantiation.At present,design constraints are diverse and focus on their own attribute constraints.There is a lack of analysis and research on correlation constraints among information.In addition,multiple constraints and alternatives make up a complex design space,which reduces the solving efficiency of low carbon design schemes;(4)The uncertainty of the life cycle design information leads to the doubt about the credibility of the carbon footprint results of the design scheme.Therefore,it is necessary to further improve the parameters that have a great influence on the results to reduce the uncertainty.Based on this,the modeling of product low carbon design,the prediction and evaluation of carbon footprint,the optimial decision of the low carbon design scheme,and the uncertainty analysis of product were proposed from the perspective of the life cycle of products,so as to support the development of the life cycle low carbon design.The main research content of this thesis is as follows:(1)A design scenario-based multi-layer modeling method for product life cycle design scheme is proposed.Function,structure,design features and manufacturing process are integrated to build a multi-layer design model for product life cycle design scheme.The complex correlation of design information in the life cycle is studied,and the relationship between design information at different layers is described systematically by using tree structure.The concept of deign scenario is identified and defined,which represents the life cycle activities of multi-layer model design information,and the design scheme information of product life cycle is comprehensively and systematically expressed.By extending the information of each node of the design model through the concept-knowledge theory,the product design space is generated to represent all the alternatives of the product,which lays the foundation for the life cycle evaluation and low carbon optimization of the product.(2)A carbon footprint prediction method for product life cycle design scheme is proposed.The case database of product design scheme is constructed based on ontology to uniformly store and manage the design scenario information and carbon footprint information,which reduce the complexity of design scheme retrieval and calculation,and provide sample support for the prediction of product carbon footprint.Based on the self-learning characteristics of neural network and the efficient iterative characteristics of differential evolution algorithm,a carbon footprint prediction and evaluation method integrating differential evolution and back propagation neural network is proposed.Relied on the sample training set of the established case base and the training of the three-layer neural network model,the correlation modeling between design scenario information and carbon footprint information is realized.Furthermore,the weight and bias parameters of the training model are solved iteratively based on the differential evolution algorithm.With the minimum mean square error as the performance evaluation index of the training model,a neural network model meeting the requirements is obtained,which supports efficient and accurate prediction of the carbon footprint of the design scheme.(3)A design constraint driven low carbon design scheme optimial decision strategy is proposed.The product design constraint is defined and divided from two perspectives,unit stage and life cycle stage,and the product design constraint network is constructed.Through the integration of design space nodes and constraint network design nodes,a constraint-based multi-layer design space based on design constraints is generated,and all alternatives and associated constraints in the product life cycle are displayed effectively.The low carbon design process of a product is regarded as a constraint solving problem of the multi-layer design space based on constraints,and the key information of the space,such as design variables,range and constraints,is extracted and transformed into a constraint satisfaction problem model.In this thesis,a greedy backtracking search algorithm combining differential evolution algorithm is proposed.With the goal of minimizing the carbon footprint of the design scheme,the design constraint based multi-layer design space is heuristically traversed to efficiently generate a low carbon design scheme that meeting the constraints.(4)An uncertainty analysis method for product life cycle design scenario information is proposed.The uncertain sources of product lifecycle design scenario information are analyzed,and the probability distribution function are determined according to the characteristics of each information.The uncertainty analysis process of product life cycle design scenario information is clarified,and the Monte Carlo method is adopted to evaluate the impact of design scenario uncertainty on the carbon footprint of design schemes.Further,based on the neural network sensitivity analysis method,the influence of various inputs is quantified by deconstructing the weight of connections between neurons,and the parameters that have the greatest impact on the carbon footprint of the design scheme are identified to guide the improvement and promotion of the design.Moreover,the key parameters are improved and changed to reduce the uncertainty of the carbon footprint of the design scheme and improve the reliability of the results.(5)Based on the above research theories and methods,a life cycle low carbon design platform for mechatronic products is developed.The system development environment and mode are determined,and the software system architecture and database structure are built,which support life cycle low carbon design of product.Moreover,taking the low carbon design of wind turbine as an example,the functions and applications of the prototype system for life cycle low carbon design platform for mechatronic products were introduced from the aspects of product life cycle modeling,carbon footprint prediction and assessment,optimial decision of the low carbon design scheme,and uncertainty analysis are introduced.The research in this thesis can help designers to generate the low carbon design schemes in product life cycle under the condition of multiple and complex information and lack of inventory in the design stage.It effectively reduces carbon footprint of product and alleviate the environmental pressure brought by manufacturing,which is of great significance for the green transformation and upgrading of enterprises.