Traceability uncertainty evaluation of electronic transformer calibrators based on DREAM-integrating Sparrow Search

Abstract Accurate traceability uncertainty evaluation guarantees the effectiveness and reliability of electronic transformer calibrators. The traceability uncertainty evaluation of electronic transformer calibrators using the guide to uncertainty in measurement (GUM) method ignores the type and width differences of the probability distribution of the inputs as well as the correlation among the inputs. To solve this problem, which impacts the accuracy and reliability of the evaluation results, a Bayesian uncertainty evaluation model based on DREAM-integrating Sparrow Search (SDREAM) is proposed. The proposed model organically combines Bayesian statistical inference with the Monte Carlo method (MCM) sampling idea, imitating the posterior probability density of multidimensional parameters with the multi-chain sampling of Markov chain Monte Carlo. Based on the multi-chain sampling of DREAM, the proposed model integrates the idea of sparrow finder search, improves the mutation strategy, and dynamically updates the crossover probability factor to balance the global and local search capabilities, which enhances the search efficiency and accuracy of the posterior probability density space of the indirectly measured multidimensional parameters. Moreover, a multidimensional posterior probability density sampling simulation example is presented based on the traceability of the digital differential method of electronic transformer calibrators. The simulation results show that SDREAM outperforms DREAM. The proposed model is also used to evaluate the traceability uncertainty of the digital differential method of electronic transformer calibrators. The obtained results are consistent with the results of the GUM and MCM methods but exhibit higher stability and good robustness.