Comparison between wire mesh sensor technology and gamma densitometry

Experimental tests were conducted recently to test both the capacitance and conductance wire mesh sensors against gamma densitometry. Similar techniques were utilised in the past at FZD to test the accuracy of the wire mesh sensor (Prasser et al 1998). It is interesting to note that comparisons of the wire mesh sensor have also been done with other measurement devices e.g. ECT (Azzopardi et al, 2009). A small closed loop test facility was utilised. This consisted of a vertical round pipe approximately 1 metre in length, and around 50mm in diameter. A 16 x 16 wire mesh sensor was used, giving a spatial resolution of 3.1mm. This sensor was placed approximately 1 metre above the injector. Temporal resolution or sampling frequency of the sensor was set at 1000 Hz throughout the experiments. Deionised water was used and hence both types of WMS could be utilised. There was a small gas injector at the inlet of the vertical pipe, this allowed compressed air to be injected, and two-phase gas-liquid mixture was generated. The gas superficial velocity was varied between 0.05m/s to 1.4m/s at two liquid velocities of 0.2 and 0.7m/s. The gamma densitometer consisted of collimated source and also a collimated detector to minimise scattering. The source was Cs 137; this generated a gamma beam of 5mm in diameter. The detector used was a NaI scintillation crystal with photomultiplier read-out. The measuring period was 120s (similar methodology was used previously by Prasser et al 1998). The gamma densitometer was placed on a moving platform approximately 1cm below the plane of wires of the sensor, in order to align it accurately using a counter mechanism, with each of the wires of the WMS, and the platform could scan the full section of the pipe. One half of the pipe i.e. 8 wires of the sensor were measured with capacitance WMS and gamma densitometer, and the other half of the pipe was measured with conductance WMS and gamma densitometer. Calibration was carried out for each position i.e. 16 points and data collected for each of the 16 wires at the stated gas and liquid flow rates. In total there were 224 runs. Different flow regimes were observed e.g. bubbly and slug. The count rate from the gamma densitometer detector was converted to void fraction (Stahl et al, 2004). For the cross-sectional void, along each wire, there was good agreement between sensors and the gamma densitometer near the centre of the pipe, however there wasn’t a good agreement near the circumference or edge of the pipe. There was consistent agreement between the capacitance and conductance wire mesh sensors throughout the experiment. This experiment also demonstrated how limited gamma densitometry actually is, i.e. it provides no visualisation. The wire mesh sensor in return gave void fractions at each crossing point i.e. a 16x 16 matrix, at a very high sampling rate, together with visualisation of flow. It also gave bubble identification and size distribution of the bubbles.