The erroneous sensors have a very negative effect on the performances of control, diagnosis, and optimization of building energy systems. A novel virtual in-situ sensor calibration (VIC) has been developed with various calibration strategies to calibrate working sensor errors in operational energy systems or buildings, based on the Bayesian inference and Markov chain Monte Carlo method (MCMC). A reviving calibration with sensitivity coefficients is the best strategy overcoming various negative effects on the VIC accuracy, and the sensitivity coefficients are very important for the reviving calibration performance. Unfortunately, however, they have been defined manually and sometimes the manual definition resulted in the low calibration accuracy. It causes various practical issues in operational systems and hinder the VIC automation, due to the human interference and low reliability. Therefore, this study identified the existing limitations and then suggested a sensitivity coefficient optimization (SCO) method to advance the reviving calibration strategy. It defines the best sensitivity coefficients automatically using an optimization process and this automation is intended to solve the previous practical challenges of VIC in the operation stage. In a cases study for an absorption refrigeration system, the optimized sensitivity coefficients decreased the systematic and random errors of all sensors significantly, compared to the existing method. All the measurements approached to their true values after the SCO-driven reviving calibration and the maximum error decreased from 76.4% to 5.3%. (C) 2019 Elsevier B.V. All rights reserved.