Correcting energy balance error in heat exchanger data by maximum likelihood method

•A new method was developed to minimize uncertainties in heat exchanger data.•Monte-Carlo simulations were used to verify error propagation.•Maximum likelihood errors are calculated based on the energy balance error.•Uncertainties of sample j factors were reduced by 19 (dry) and 53% (wet).•The discrepancy of UA-LMTD and ε-NTU methods due to energy balance error is fixed.