Implementation of Linear Bias Corrections for Calorimeters at Mound

In the past, Mound has generally made relative bias corrections as part of the calibration of individual calorimeters. The correction made was the same over the entire operating range of the calorimeter, regardless of the magnitude of the range. Recently, an investigation was performed to check the relevancy of using linear bias corrections to calibrate the calorimeters. The bias is obtained by measuring calibrated plutonium and/or electrical heat standards over the operating range of the calorimeter. The bias correction is then calculated using a simple least squares fit (y = mx + b) of the bias in milliwatts over the operating range of the calorimeter in watts. The equation used is BI = B0 + (Bw * Wm), where B, is the bias at any given power in milliwatts, B0 is the intercept (absolute bias in milliwatts), Bw is the slope (relative bias in milliwatts per watt), and Wm is the measured power in watts. The results of the study showed a decrease in the random error of bias corrected data for most of the calorimeters which are operated over a large wattage range (greater than an order of magnitude). The linear technique for bias correction has been fully implemented at Mound and has been included in the Technical Manual, \"A Measurement Control Program for Radiometric Calorimeters at Mound\" (MD-21900)