A Novel Calibration Method and Computer Simulation for Multi-hole Pitot Tubes

Greenhouse gas emissions have been regarded as a global challenge and several national metrology institutes have started research on this topic. Smokestacks are one of the main emission sources and its flow measurements draw much attention due to the unstable flow conditions and complex gas composition. Pitot tubes are widely used in the environmental analysis. However, the traditional pitot tube can only provide one-dimensional flow velocity and the measurement locations also need to be placed with care. Multi-hole pitot tubes have been claimed that it can be applied to three-dimensional swirl flow measurements in the smokestack and provides more accurate results. The main drawback for using multi-hole pitot tubes is the time-consuming and complex calibration procedures before implementation. The possible way to significantly reduce the time and costs is to establish an automatic calibration traversing system and programmable calibration method. Moreover, the latest research also revealed that flow separation and hysteresis occurred during multi-hole pitot tube calibration and resulted in discrepancies in the repeatability testing. Therefore, flow visualization, surface pressure analysis and calibration data modeling need to be further studied in order to establish appropriate measurement technology for quantifying the greenhouse gas emissions effectively. In this paper, ANFIS (Adaptive-Network-based Fuzzy Inference System) method was first applied to multi-hole pitot tube calibration modeling owing to its capability of efficient learning, easy implementation and excellent explanation through fuzzy rules. The results showed that ANFIS method can help identify the dominant parameters and construct the network of pitot tube calibration parameters among non-dimensional pressure coefficients, flow angles and flow velocity. Additionally, a commercial CFD software, ANSYS Fluent 14, was used to simulate the flow hysteresis during pitot tube calibration. The simulation was carried out by unsteady computation and the Shear Stress Transport (SST) ê-ù turbulence model was also adopted to study the adverse pressure gradients and flow separation. The simulation results showed that the location of recirculation area can be identified by the contour of negative X velocity and vorticity. It’s helpful for elucidating the laminar boundary layer separation and the behavior of flow transition on the pitot tube when flow hysteresis occurs.