Parametric sensitivity analysis for thermal runaway in semi-batch reactors: Application to cyclohexanone peroxide reactions

The semi-batch reactors (SBRs) system, which is widely used in industrial processes, possesses an intrinsic parametric sensitivity, in which infinitesimal disturbances of input parameters can result in large variations in output variables. In this work, local parametric sensitivity analysis (PSA) was used to understand parameter variations and global PSA was conducted to examine the interaction of input parameters. The effects of these parameters on the output of the system model were analyzed based on the Monte Carlo method with Latin hypercube sampling and the extended Fourier amplitude sensitivity test model. The results showed that the evolution of thermal behaviors in SBRs were observed: marginal ignition; thermal runaway; and the quick onset, fair conversion, and smooth temperature profile. The threshold point of transition from marginal ignition to thermal runaway was at the maximal value of local sensitivity, for which the slope with respect to cooling temperature equaled zero. Moreover, the sequence of the global sensitivity of six common input parameters was computed and evaluated. The reliability of the numerical models was verified by using our previous experimental results of cyclohexanone peroxide reaction. This comprehensive sensitivity analysis could provide valuable operating information to improve chemical process safety. •A theoretical method was developed to perform the local and global PSAs for thermal runaway in SBRs.•Two special thermal transition behaviors were observed during the process of thermal runaway in SBRs.•The threshold point from thermal runaway to QFS transition was found to be ats(τ∗)φi=0.•The main sequence of first-order sensitivity was the same as that obtained from global PSA:B>Wt>Da>θd>γ>θ0.•Theoretical predictions concurred well with experimental data of cyclohexanone peroxide thermal runaway reactions.