Research On Fault Diagnosis Method Based On Fault Information

With the growing size and complexity of integrated circuits,the possibility of circuit failure increases dramatically.In order to improve the output and quality of integrated circuits,fault diagnosis of circuits becomes an indispensable step.Fault diagnosis can effectively improve the efficiency of integrated circuit fault diagnosis and reduce the cost of fault diagnosis.Thanks to its importance,many domestic and foreign scholars are engaged in the research of integrated circuit fault diagnosis.The main types of circuits studied are digital circuit,analog circuit and digital-analog mixed circuit.There are many fault diagnosis methods designed for different circuit types in the existing research.Based on the existing research,there is still much room to improve the accuracy of fault diagnosis for the current methods.The most efficient fault diagnosis method is designed for the type of digital integrated circuit.The faults of integrated circuits usually consist of multiple single faults.Current fault logic diagnosis algorithms will produce a large number of candidate faults when they complete the fault diagnosis,which will certainly affect the next step of the fault repair work.In view of the disadvantage of current fault diagnosis algorithms,a RRC algorithm for fault diagnosis is optimized,which combines the response characteristics to reduce the fault set.Based on the fault response characteristics,the algorithm puts forward the concept of complete diagnostic accuracy,which reduces the set of candidate faults according to the complete diagnostic accuracy,thereby reducing the size of the whole fault diagnosis.The optimized method ensures that every real failure is selected in the candidate diagnostic set and greatly reduces the size of the candidate fault set.Compared with the latest fault diagnosis methods based on test scores,this method can not only shorten the diagnosis time effectively,but also ensure high accuracy and improve the diagnosis efficiency.After that,based on the newly proposed RRC method,according to the output response characteristics of single fault,no fault and actual multiple faults,this paper further presents a FIRRC method for fault diagnosis based on fault information response ratio combined with response characteristics by analyzing and collecting fault information under the set of test excitation vectors.This method avoids the situation of high diagnostic accuracy and low diagnostic resolution in fault diagnosis accuracy.Based on the influence of fault information on the accuracy of fault diagnosis in different ways and degrees,the concept of response ratio is proposed.To compute the degree of impact of failure information on the final candidate diagnostic solution,the concept of impact factor is proposed.By taking into account each of the impact factors in the fault information,we can effectively avoid ignoring the important fault information and calculate the response ratio of candidate faults.The set of candidate failures is reordered according to the response ratio.The selected failures after the final reordering are the candidate diagnostic solutions.Comparative experiments with the latest TSDFF methods show that the diagnostic resolution and diagnostic accuracy of the FIRRC method are improved,which greatly improves the accuracy of fault diagnosis and greatly improves the diagnostic efficiency overall.In addition,the method presented in this paper effectively reduces the running time of fault diagnosis,with an average reduction of 3.4%.