Search Results

In this paper, a new computational method is provided to identify the uncertain parameters of Load Sensing Proportional Valve (LSPV) in a heavy truck brake system by using the polynomial chaos theory. The simulation model of LSPV is built in the software AMESim depending on structure of the valve, and the estimation process is implemented relying on the experimental measurements by pneumatic bench test. With the polynomial chaos expansion carried out by collocation method, the output observation function of the nonlinear pneumatic model can be transformed into a linear and time-invariant form, and the general recursive functions based on Newton method can therefore be reformulated to fit for the computer programming and calculation. To improve the estimation accuracy, the Newton method is modified with reference to Simulated Annealing algorithm by introducing the Metropolis Principle to control the fluctuation during the estimation process and escape from the local minima.

This paper deals with the robust design of the Load Sensing Proportional Valve (LSPV). To find out the parameters which have main effect on the performance of the LSPV, the DOE based on orthogonal experiment is carried out utilizing the LSPV model built in AMESim environment. In order to save computation expense, the RSM technique is used to approximate the optimal objectives and constraints. Then a robust design methodology using multi-objective evolutionary algorithm (MOEA) is performed and a set of non-dominated solutions are therefore obtained. With specified assessments, feasible solutions can therefore be selected from the entire field of the Pareto optimal solutions. The validation is made by Monte Carlo Simulation Technique in terms of the robustness of the feasible solutions.

As is known to all, there are some contradictions between the handling and ride performance during the design process of vehicles. Sometimes owing to serious collisions of each criterion in the high-dimensional solution space, the common method to deal with the contradiction is to transform into a single target according to weights of each objective, which may not obtain a desired result. A multi-criteria approach is therefore adopted to optimize both properties and the result of a multi-criteria design is not a unique one but a series of balanced solutions. This paper is focused on the robust design of a simplified vehicle model in terms of not only ride comfort but also handling and stability using a multi-objective evolutionary algorithm (MOEA) method. Using the proposed method, the conflicting performance requirements can be better traded off. One of the most important indexes to characterize the vertical ride comfort is the acceleration of the sprung mass.

Interval inverse problems can be defined as problems to estimate input through given output, where the input and output are interval numbers. Many problems in engineering can be formulated as inverse problems like vehicle suspension design. Interval metrics, instead of deterministic metrics, are used for the suspension design of a vehicle vibration model with five degrees of freedom. The vibration properties of a vehicle vibration model are described by reasonable intervals and the suspension interval parameters are to be solved. A new interval inverse analysis method, which is a combination of Chebyshev inclusion function and optimization algorithm such as multi-island genetic algorithm, is presented and used for the suspension design of a vehicle vibration model with six conflicting objective functions. The interval design of suspension using such an interval inverse analysis method is shown and validated, and some useful conclusions are reached.