EXP 1-1 Data Quality and Accuracy

Lib:Create_Ercoftac_Article_Form

Front Page

Introduction

Review of experimental studies

Description

Experimental Set Up

Measurement Quantities and Techniques

Data Quality and Accuracy

Measurement Data and Results

Data Quality and Accuracy of Measurements

The spray, in the case without the airflow, is axially symmetrical with low residual irregularities on the velocity, droplet size and concentration resulting from the imprints of the two inlet ports. Also, the airflow is symmetrical along the horizontal and vertical axes of the test section, as seen in Figure 13.
Several repeated measurements were performed to assess the repeatability of the PDA measurements. Based on that, the measurement errors were estimated (A type of uncertainty estimate). The velocity measurement error was under ±0.3 m/s, the linear average droplet diameter ($D_{10}$) error was ±1.1 µm, and Sauter mean diameter ($D_{32}$) error was ±1.8 µm.
The PDA system provides velocity and spherical validations whose results are documented in the measurement files.
The spatial resolution is 46 μm/pix × 46 μm/pix, which restricts its geometric detection limit. Optical distortions of long-range microscopis lens is negligible compared to other errors. The images were processed to estimate the spray boundary. The standard mean deviation of the spray boundary detection was ±0.55 mm.
The wind tunnel was designed and tuned to provide controlled, repeatable and stable flow with low $Tu$ and flat velocity profile. The flow conditions in the test section of the tunnel were inspected before the measurement with Laser Doppler anemometry. The velocity and Tu were measured in 20 equidistantly spaced positions 150 mm downstream from the inlet of the test section in the case without the atomiser spraying, see Figure 13. The horizontal velocity profiles at three cross-flow velocity cases in the test section are documented there. The error bar width in Figure 13 is in all cases smaller than the symbol size. The vertical velocity profiles (not shown here) are symmetrical and similar to the corresponding horizontal profiles. The free-stream $Tu$ is lower than 0.8%. It was computed as the ratio of root mean square cross-flow velocity to average cross-flow velocity. The positioning error of the 3D computer-controlled system used for positioning the wind tunnel body relative to the measurement volume of the PDA was less than 0.1 mm.

Figure 13: Measurement positions, velocity profiles and turbulence intensities for different cross-flow velocities

The uncertainties of the operation conditions of the atomiser, namely the inlet pressure and flow rate of the sprayed water are detailed in the section Atomizer under test and its supply system.

Limitations of the experiment

The experiment is a physical representation of the case of interaction of a conical liquid sheet with flowing gas. The ideal case represents a steady continuous discharge of the swirling liquid with the formation of a smooth attenuating sheet. The gas is expected to undergo steady, low-turbulent, incompressible flow with uniform velocity in the far (unaffected) field. The real experiment is provided as a case of confined flow with the boundaries formed by the section walls and corresponding gas flow characteristics as detailed in Section Data Quality and Accuracy of Measurements.
The atomiser was carefully designed and precisely fabricated by machining with the aim to produce stable internal swirling flow and steady discharge (possible effects of fabrication imprecisions are studied in ). Also, the water supply system was checked for unwanted pulsations that could propagate from the pump through the pipe to the nozzle. The atomiser showed no unwanted pulsations or fluctuations of the produced liquid sheet or spray.
The walls of the wind tunnel were designed to be sufficiently rigid and a vibration measurement was performed to confirm that no unwanted vibrations are excited by the flow or generated in other ways that could as a result affect the flow. Also, the compensator (2) was used to isolate the spreading of the fan vibrations to the tunnel construction. The flow was probed to find any harmful fluctuations e.g. at the fan blade frequency. This tunnel testing was provided without finding any negative effects and it is detailed in .
The optical techniques for measuring (PDA) and visualisation (HSC) of two-phase flow require visual access to the flow field within the object and a limited distortion of the optical path. The section walls can be considered ideally smooth, planar and forming a rectangular shape. However, the sprayed water can deposit on the side walls and the bottom of the section, which makes them slightly wavy with a height of irregularities in tenths of mm. These deposits formed on the side walls, behind the atomiser, out of the positions of the optical path of both the PDA and HSC measurements. The sprayed liquid was torn down by the airflow, so the deposits formed behind the spray and their possible effect on the boundary layer of the windows were unimportant. The influence of test section windows on the accuracy of PDA measurement was investigated. The spray produced by an air-brush nozzle was tested for the case with and without the windows present at different arrangements of the PDA. Axial velocity and diameter of droplets were found unaffected by the windows, as detailed in .
One limitation of the experiment relies on the measurement network. Even though we made it very detailed, we have not measured, for example, the droplet sizes produced by bag break-up in distant positions from the atomiser.

References

---

Contributed by: Ondrej Cejpek, Milan Maly, Jan Jedelsky — Brno University of Technology

Front Page

Introduction

Review of experimental studies

Description

Experimental Set Up

Measurement Quantities and Techniques

Data Quality and Accuracy

Measurement Data and Results

© copyright ERCOFTAC 2024