Latest revision as of 16:01, 11 February 2017

Induced flow in a T-junction

Application Challenge 3-02 © copyright ERCOFTAC 2004

Overview of Tests

The experiment is described precisely in ‘Deutsch, Montanari, Mallez’ (1997). Note that the experimental set-up allowed to locate the junction at least 1000 mm downstream of the nearest bend of the main pipe (no precise measurements have been made to assess the development of the turbulent flow, nevertheless a fully developed flow is assumed).

Experimental data have been obtained on this mock-up using flow visualization method and particle image velocimetry (PIV). They show an helicoidal structure (‘corkscrew’) extending along the auxiliary pipe. The length of the vortex penetration has been measured by averaging one thousand instantaneous velocity fields obtained using PIV system. Moreover, the rotation velocity has also been measured along the auxiliary line. To make comparisons easier and less dependent on the very fine details of the geometry at the T-junction, these data are scaled by the rotation velocity near to the junction (at H/D = 3) : this gives quite a clear picture of the rate of decay of the swirl in the auxiliary line.

All three experiments differ only by the bulk velocity in the auxiliary pipe. Only EXP1 will be detailed hereafter. Table 2 EXP A provides a summary of all experimental test cases.

Test Case EXP-1

Description of Experiment

Instantaneous images of the flow are obtained by illuminating the flow with a Nd:YAG pulsed laser. Light scattered perpendicular to a laser sheet is captured by a CCD camera. The flow pattern in the dead leg is determined by careful observation of the motion of tracer particles.

Instantaneous velocity fields are obtained using particle image velocimetry. The illumination and acquisition set-up relies on two frequency doubled, pulsed Nd:YAG lasers mounted side by side on an optical bread board. The beam paths are combined with a set of mirrors and a polarization beam splitter. The initial polarization of the lasers is such that the one-axis beam is transmitted by the splitter while the off axis beam is reflected. The beams pass through both spherical and cylindrical lens to form horizontal light sheets intersecting the auxiliary pipe. Laser output is on the order 200mJ per pulse and the time delay between pulses from the two lasers is 225 to 2000 microseconds. The flow is seeded independently with Silver coated hollow glass micro spheres. The particle density is close to 1 to avoid drift velocity between fluid and particles. The acquisition frequency was 10 Hz.

Boundary Data

The flow at boundaries is described in Table 1. EXP1 correspond to VA/VM = 1%. Table 2 EXP A provides a summary of all experimental test cases.

Measurement Errors

No measurement of the error is available here.

Measured Data

Particle image velocimetry of the velocity field has been obtained in different planes of the dead leg. Flow visualization has been carried out in vertical planes slicing the dead leg. These data allowed to determine a vortex penetration length which is the experimental test data provided here (Figure 4).

Table 2 EXP-A Summary Description of all Test Cases
Name	GNDPs	PDPs (problem definition parameters)		MPs (measured parameters)
	Inlet Re	Main Pipe Bulk Velocity (m/s)	Auxiliary Pipe Bulk Velocity (m/s)	DOAPs
EXP 1	895,000	9.2	0.092	Vortex Penetration
EXP 2	895,000	9.2	0.046	Vortex Penetration
EXP 3	895,000	9.2	0.023	Vortex Penetration