Reliability Analysis Methods Of Circuit Tolerance Using Belief Rule Base

In the design phase of the digital analog circuit,the circuit reliability analysis is an important part of the circuit design and development process.It is very important to improve yield and reliability of product.At present,there are two main types of methods to analyze the reliability of circuit,one are based on the statistical analysis methods,the other are based on geometric approximation methods.These methods usually have some disadvantages,such as large amount of computation and special requirements for circuit performance function,which will limit the application in practical circuit design.In fact,in the process of circuit design and circuit simulation,circuit designers or users can accumulate some subjective and uncertain experience knowledge.If the subjective knowledge and circuit simulation objective data are combined to carry out tolerance analysis,not only the efficiency of circuit design can be promoted but also the product development cycle can be shortened.In order to achieve this goal,this paper presents a new method of reliability tolerance analysis and model optimization based on the combination of subjective experience and circuit simulation data.The main work includes:(1)A circuit tolerance design method based on belief rule base(BRB).A BRB model is established to describe the nonlinear mapping relationship between the parameters of circuit elements and the performance indicators of the circuit;In the complex situation of design domain is non connected and non convex,a method based on BRB to solve the initial value of the circuit tolerance and optimize the tolerance region is proposed;Take the performance index function of Rosenbrock-type circuit and the performance of the actual track circuit tuning unit as an example,it is verified that the proposed method can accurately determine the elements tolerance and center value from the design domain with a relative small computational cost.(2)A method of circuit center value design based on belief rule base.When the problem(1)is fixed,the circuit tolerance analysis is to find the optimal central value(nominal value)of the element.A circuit reliability estimation model based on BRB is established,which is used to describe the nonlinear mapping relationship between the central value of the circuit element parameters and the reliability of circuit;Then,a circuit central value design method based on BRB is proposed.Compared with other nonlinear optimization methods,linear programming is more easily and more accurate,and the method can be used to optimize the parameters from any point in the feasible solution space;In the end,with the example of designing tuning element parameters of the track circuit,this paper verifies the advantages of this method in training BRB model parameters and the optimization efficiency.(3)A BRB model parameter optimization method based on sequential linear programming.Through the first order Taylor expansion of the objective function and constraint conditions of the optimization model of the BRB parameters in the first two problems,the nonlinear programming problem is transformed into a simple linear programming problem;Then given the parameters to find the optimal range,determine the optimal range,linear search,the search results are used as the basis of the new basis points,the BRB model is to be found to meet the requirements of the stop.Compared with other nonlinear optimization methods,the method of linear programming is more easily to be solves and has higher accuracy.And this method can be used to train the parameters from any point in the feasible solution space;In the end,with the example of designing the center value of tuning element parameters of the track circuit,this paper verifies the advantages of this method in the training BRB model parameters and the optimization efficiency.