The measured data in a utility plant have little consistency in both mass and energy balances due to random and gross erros, Data Reconciliation tries to resolve this inconsistency in measured data. Gross Erros that include leak, miscalibration of measurement instrumentation, and sensor malfunction should be detected and eliminated from the measured data. Severak methods for gross error detection such as Chi-square test, Measurement test, Constraint test, Maximum power test, and Principal component test are cinoared abd analyzed on several cases in a utility plant. This article summaruzed the distinguishing features in every tests and proposed optimal processes for each tests.