The Research On Reliability Of FPGA Software And Hardware System Based On Petri Net

The rapid development of aerospace enterprise makes the tasks of space load digital systems more and more diverse and complex.The single-event effects which caused by meteorological phenomena such as high-energy particles make some core equipment abnormal in the process of its operation,causing the problems such as data errors,core function failures,and even hardware failures.As one of the cores of the digital system,the design of the FPGA is becoming more and more complex.How to ensure the reliability of the FPGA during the design and verification phase,as well as quickly and accurately locate the potential hidden dangers in the complex FPGA system,has become a bottleneck restricting the FPGA application,and is also the key and difficult point in this field.For the past few years,researchers have proposed a series of methods for the ground verification of FPGA engineering reliability,including code correctness verification,system function test and formal verification.Compared with the formal verification method,the code verification technology cannot accurately evaluate the influence of the single-event effect.While the experimental period of the system test is long and easy to generate missed detection and misdetection.Therefore,based on the idea of formal verification and Petri net theoy,this thesis proposes a PNOF model and the analysis method of functional execution reliability,and gives an example for verification.The main work of this thesis is as follows:(1)Based on the analysis of the mechanism and influence of single-event effect and the structural characteristics of FPGA devices,the single-event upset error rate model of FPGA in orbit is presented.Then,according to the requirements of model establishment and analysis,the calculation method of single-event upset error rate of the function module in the FPGA engineering is presented.(2)According to the features of FPGA design,a topological relationship acquisition method based on functional modules is proposed.Furthermore,a classification method of connection variables in topological relations is proposed,which classified according to whether or not to perform control functions so that can describe the system more accurately.Combining FPGA topology with improved Petri Net theory,PNOF model is constructed based on the common Petri Net concept,and the practical application engineering is given to verify the correctness of the method.This model is also the foundation of the following analysis method research.(3)Based on the state reachability graph theory of Petri Net and the hierarchical reach graph method,the workflow reachable graph analysis method of PNOF model is proposed.Furthermore,the function reliability evaluation method is put forward in this thesis to analysis the PNOF model which evaluated the model in a workflow,the influence of single-event upset error on the control function of the control variable represented by the control library,and the example calculation is given to verify the feasibility of the index.(4)From the perspective of method design and protection effect,three-mode redundancy based on module partitioning and time-filter redundancy are studied.The two protection methods are evaluated for the function execution reliability index FER and protection priority.On the basis of resource and time constraint,the genetic algorithm is used to solve the optimal combination of protection methods under two constraints,and the correctness of the functional execution reliability index is proved.Compared with the existing formal methods,the PNOF model with functional modules as the granularity and control variables as the core can describe the functional characteristics of FPGA engineering in a macroscopic and accurate way.The function performance reliability index proposed in this paper is demonstrated in the example verification.Its correctness and strong practical guiding significance provide a new basis for the design and verification of space digital systems.