A Functional Representation And Decomposition Method For The System Design Of Multi-Domain Complex Mechatronics

The multi-domain complex mechatronics is involved in cross disciplines such as mechanical, electronic, control, and so on. It is formed with the interaction of energy, material and signal through multiple components. With the increasing of complexity, it is a great challenge to implement the automated design for multi-domain complex mechatronics. Due to users usually give the overall functional requirement of the product, function decomposition, which generates the functional model hierarchically through breaking the coarse-grained overall function to some fine-grained sub-function until they are solvable, is the first step and plays a crucial role in the system design of mechatronics. There have been some researches about system design, but there is still no effective and practical method or tool to support system design. How to get the function from requirement and get the structure from function is still an open problem. Even the understanding and definition for the concept of function is still ambiguous. This paper will start from the representation methodology of functional knowledge, and propose a solution for the automated functional decomposition of system design.To support automated functional decomposition, the representation methodology of functional knowledge is firstly needed. The methodology includes the representation for function and representation for working principle. Function is defined as the relation between input flow and output flow as a whole. Its action is to exert some effects to the input flow, change the type or properties of the flow, and then obtain the aimed output flow. These effects are called functional effects. Working principle is the combination of the physical effect with the geometric and material characteristics. It describes how the function of product works. The key concept of working principle is physical effect, which determines the innovation degree of function to a large extent.Three function decomposition strategies are proposed based on the representation methodology of functional knowledge. They are effect decomposition, principle decomposition and backward search decomposition. Effect decomposition breaks the overall function which has functional effects into corresponding sub-function. Principle decomposition breaks function in different ways according to different working principles. Backward search decomposition is used when effect decomposition and principle decomposition are not suitable, which will obtain sub-function chains that satisfy the input flow and output flow of the original function. A lot of principle solutions will be generated through the decomposition process. Designers can find suitable schemes of conceptual design that contain innovative principles.Based on the above research work, a SysML-based function modeling and decomposition tool is developed. The decomposition process is shown with the example of the automobile engine.