The Fortune 500 HVAC company maps performance for its compressors across a wide range of mass flow rates, rotational speeds, and inlet guide vane angles. This data set is used to create predictive models of water chiller performance. These models must meet strict industry accuracy standards, meaning months of painstaking and costly testing for each new design.
SmartUQ utilized intelligent data sampling to choose which points should be sampled and advanced emulation technologies (aka, predictive modeling) to build the performance maps. This reduced the total number of tests needed; minimizing cost and time.
A new sampling method and emulator-based performance mapping process were successfully demonstrated. Using these techniques would allow the condition and total number of tests to be more specifically tailored to the accuracy requirements of the performance map, thus reducing the testing requirements and costs.
