Research On Testability Design Of Switching Mode Power Supply Considering Uncertainty

As a device that performs power conversion and power transfer,the switching mode power supply(SMPS)is widely used in various power electronic systems.With the concept of all-electric equipment proposed and further development,the quality and reliability of the SMPS need to meet higher requirements.In the use of power,fault is inevitable.If one considers its testability in the product design process,it will lay a good foundation for fault detection and precise isolation,and it will also reduce the cost of the product’s life cycle.Due to the tolerance of various components of the SMPS and the presence of test errors,the correlation between the fault and test is uncertain.If this problem is ignored,the design of product will be flawed,and even lead to misuse and misdiagnosis.This article will take a LED drive power as the research object,and study the testability design that considers uncertainty,the main contents are as follows:First,analyze the fault mode of the switching power supply.After analyzing the working principle of the SMPS,the circuit simulation model is established based on Saber,and the accuracy of the model is tested.Based on this,the simulation model is used to perform sensitivity simulation,and determines the components that have a major influence on the main performance indicators of the power supply;combined failure rate,the key components and fault modes of the SMPS are determined.Secondly,according to the characteristics of the SMPS,a method of combining the multi-signal flow diagram model with simulation is proposed to establish testability model.Firstly,a multi-signal flow diagram model of power supply is established to realize initial selection of test nodes.Under consideration the components tolerances,Monte Carlo simulations are performed for various fault modes.The characteristic parameters of the system are analyzed,and get the test parameter sets affected by each fault;thus the fault-test correlation matrix of the SMPS is established,and the inherent testability of the SMPS is analyzed to obtain its fault detection rate,fault isolation rate,failed fuzzy groups and test redundancy groups.Then,under the influence of the components tolerances and test errors,realize the teat optimization of power supply.According to the small sample test theory,the Monte Carlo simulation data is used as the pre-test information,and the fault injection test data of power supply is taken as a sample to estimate the distribution parameters of each test.After that,the relationship between the threshold of the corresponding test under each fault condition and the detection probability of the test is established to construct a fault-test correlation matrix that takes uncertainty into consideration;on this basis,the particles that are currently widely used for optimization are established.The refined particle swarm optimization is used for test optimization,finds the threshold value and the optimal test set that satisfy the testability index,and gives the corresponding testability index.Finally,a test verification platform for SMPS is built,and the testability design of this paper are verified using this platform.By building a human-computer interaction interface,the platform can be used to inject multiple fault modes of SMPS,set up test items,process and analyze of data,and construct fault-test correlation matrix,analyze the testability,get results of the fault diagnosis and evaluate of its testability design.