Research And Application Of NUIT "Burnup Step Size-nuclide Density" Transformation Relationship Covering All Nuclides

Software verification and validation is an important means to ensure the correctness and reliability of complex numerical simulation software.At present,the domestic nuclear power software verification mainly adopts direct comparison method,because the solved mathematical physical equations are partial differential equations,it is usually difficult to construct analytical solutions or exact solutions,and the benchmark problem,experimental data development cost is high,long period,advanced reactor software and new generation software do not have reference procedures,which makes the autonomous software verification face the problem of testing expected value(Oracle),making the verification inadequate,thus affect the safety and economy of engineering analysis and design.Metamorphic testing is an indirect comparison technique without constructing numerical solutions or verification cases,and it is one of the mainstream techniques to alleviate the Oracle problem.Because the determination mechanism comes from the physical model and calculation method,the technique is not limited by the specific implementation technology of the code,and it has been initially applied to the verification of a variety of autonomous nuclear power design software with good results,which has a broad application prospect in the nuclear power field.Metamorphic relations are the core of metamorphic testing,and metamorphic relations are composed of input and output modes.Metamorphic relations can usually be obtained by analyzing the nature of physical models,numerical algorithms,or by mining in operational data.Since the fuel consumption calculation program NUIT involves more than 3800 nuclides,it is challenging to identify the metastable relations for each nuclide.In this paper,we investigate the metastable relationship between"burnup step-nuclide density",establish a likelihood metastable classification tree,and successfully apply it to the validation of NUIT,a source term calculation program for high-temperature gas-cooled reactor cores,with good results.(1)A new method for identifying the likelihood metastable relationship of nuclide density for the fuel consumption calculation program is proposed.First,a set of major nuclides is selected,the input pattern is determined by parameter analysis,the test input pair is generated,the tested program is executed and the nuclide density is obtained,the output pattern is mined to form the metamorphic relationship and the metamorphic relationship classification is established;for other nuclides,their relationship with the major nuclides is analyzed first,and then the corresponding metamorphic relationship is selected and verified.If the requirements are met,the nuclides are put into the classification tree,and if not,new metamorphic relationships are obtained by performing the mining task.(2)For the NUIT burnout program of high temperature gas-cooled reactors,the method proposed in this paper successfully mines the nuclide density change patterns of all nuclides in the burnout program at the input pattern r1:x_j=x_i+1,i.e.,the output patterns of the likelihood metastable relations,and classifies the nuclides with the same change patterns.1194 nuclides have an output pattern in the form of a cubic functional relationship,266 nuclides have an output pattern in the form of a linear functional relationship,331 nuclides have an output pattern in the form of a power functional relationship,225 nuclides have an output pattern in the form of a quadratic functional relationship,2 nuclides have an output pattern in the form of an inverse functional relationship(for fuel nuclides),and 9 nuclides have an output pattern at The current step does not conform to the previously predetermined ten output pattern forms,and the plausibility of their ten predetermined output pattern forms are all less than 0.9.(3)The above likelihood metamorphosis relation is used for NUIT correctness verification.Forty test cases were designed based on the above likelihood metamorphosis relationship,and eight defects were detected from NUIT.The above study shows that the proposed likelihood metamorphism identification method for NUIT is effective,avoiding the dynamic mining of output patterns with a large number of test cases for each kernel,greatly reducing the workload and data redundancy,saving a lot of labor and time costs,and reducing the difficulty of metamorphic relationship discovery.The likelihood metamorphic relationship obtained from it can be directly used for the validation of the fuel consumption calculation procedure;it can be used as a reference and reference for other professions such as thermal engineering and hydropower,and provides important support for the application of metamorphic testing technology.