Parametric Study of a gate valve

Parametric Study of a Gate Valve:

Objective:

1. To perform a steady state analysis on a gate valve assembly and do a parametric study for mass flow rate from the pipe at various valve lifts.
2. To compare the results acquired via parametric method with the results acquired by individual simulations.

Geometric Setup:

• To perform the parametric study the given gate valve assembly was imported in space-claim to perform volume extraction operation. The volume in this case Water is to pass through the gate valve to other side where the output ‘Mass flow rate’ is to be calculated as the result for the simulation.
• The parametric study for this case is varying the Valve lift linearly to analyse its effect on the mass flow rate output. As we know the gate valve acts as a obstacle between the inlet side and the outlet side and is opened or closed to vary the mass flow through the pipe.
• Using the move option, the valve disc is lifted in the +z axis by 10mm. This motion is then parameterised in a group as driving motion. This way instead of going to space-claim and manually moving the disc for every variation we can just input the driving dimension in the ‘Parameter set’ as various ‘Design Points’, apply all the simulation settings and get results. For this study we are varying the valve lift and observing the mass flow rate as the result of it.
• To provide sufficient volume for the simulation, both inlet and outlet volume faces were ‘pulled’ to mimic a pipe of continuous cross section with a 800mm in length.

Analysis Setup:

• A steady state analysis was done for the given case, to simulate the flow of fluid – Water inside a continuous pipe which is controlled by a gate valve.
• As stated previously, the volume through which the fluid will flow is extracted in space-claim with various valve lifts using the parametric set option.
• We are only simulating a single fluid, only for the mass flow rate as the required output, the energy equation or discrete phase option is not required.
• The gravity option was activated and a figure of 9.81 m/s was input in – Z axis. For boundary conditions, an ‘pressure Inlet’ option was selected with a input of 10 Pascal. Using the standard initialization the input velocity was calculated to be 0.1415 m/s at the inlet.
• For the turbulence modelling, ‘k-epsillon’ model with 2 equations was added, with realizable model and scalable wall treatment functions. A report definition for Mass flux was defined at the outlet with plot history.

Mesh setup:

• The fluid volume extracted for the particular valve lift was meshed using default parameters as there isn’t any need for complicated mesh as the results we are interested are at the outlet.
• It is essential to mesh the contracting portion of the fluid which mimics the slightly opened valve. To make sure of that the valve portion of the fluid volume around and under which the fluid is flowing is provided with a sizing option with element/cell size of 3mm. This disrupts the ‘parametric’ condition of keeping more than 1 variation across the series of simulations, hence is avoided when going for the use of ‘Parameter’ option.
• To cope with the licensing limits, the mesh covering the rest of the fluid has a slightly bigger element size of 9 mm. which makes sure all other parts are sufficiently captured.
• For the parametric study method, no local refinement of mesh where the valve disc sits inside the fluid can be provided, hence a element size of 7.5mm was given for the whole model with default mesh parameters.

Case 0: Zero lift

• To start with the simulation it is always good practice to start from 0. This case has zero valve lift which represents a closed valve. Any output flow observed in this case can be considered as leakage. The valve disc makes sure a very minute amount of fluid passes under it due to the higher inlet pressure. The majority of leakage is observed passing over the disc across the valve.
• Maximum fluid velocity recorded: 0.11m/s near the narrow gaps between the valve shoulder and the pipes which is less than the initial 0.1415m/s input.

https://drive.google.com/open?id=1wH4PCVwAw4KO0T5J3bUKSJVERHzsLpDT

Case 0 Results:

• As expected only a minute amount of mass flux of 0.10144 Kg/s is recorded as the output. The iterations v/s residuals plot starts going to a steady state post 500 iterations which can be assumed as a convergence indication.

Case 1 – Valve Lift 10 mm:

• The valve disc is lifted by 10 mm in the positive Z axis to simulate a just opened gate valve. The ‘volume extract’ option was run to capture the opened gap through which the fluid will flow. It is made sure by taking a look at the cut-section of the whole assembly that the fluid volume is extracted accurately.
• A minute increase is recorded over the 1st case for the maximum velocity recorded at 0.121m/s which is still lower than the input.

https://drive.google.com/open?id=1xI9D5wRaF_HpsIfmx1VzOi48cmQyAZ4Q

Case 1 Results:

• A 10 mm lift of the valve disc opens up the flow very slightly, as a result only a small amount of fluid can pass through. The calculated value after 600 iteration comes out as 0.1394 Kg/s. The negative sign in the picture indicates that the volume is ‘getting out’ of the defined domain. The solution can be assumed as converged as the residuals start reaching a steady state post 300 iterations.

Case 2– Valve Lift 20 mm:

• Using the parametric group feature, a valve disc lift of 20mm is applied by creating a 2nd design point and putting the value as the input parameter which is then applied to the group. Following the previous pattern, the mass flow rate as well as the velocity of fluid increases with the valve lift.

https://drive.google.com/open?id=1SyER83xBEFY823DlH80uwhr4m45Rcz0P

Case 2 Results:

• The output mass flow rate comes out to be 0.2395Kg/s, while the solution reaches convergence at 500 iterations.

Case 3 – Valve Lift 30 mm

• A valve disc lift of 30mm is input for the 3rd ‘Design point’. At this point the valve gate is open enough to let considerable amount of fluid flow through.
• Due to a converging effect caused by the valve disc bottom and the pipe’s curvature which is there to clear the valve, a maximum fluid velocity of 0.190m/s is recorded.

https://drive.google.com/open?id=1arFXmWObMhaq_FjUELviOGElbHSAd24L

Case 3 Results:

• A mass flow rate of 0.3611kg/s is recorded for this case s seen in the image. As seen in the velocity plot a lot of turbulence and disturbances are created as the fluid makes contact with the valve disc long with all its sharp edges and corners.

Case 4 – Valve Lift 40 mm:

• For the maximum valve lift in the cases considered at 40mm the maximum velocity attained was 0.235m/s at the sharp edges of the valve disc. A small example of Bernoulli\'s principal can be witnessed at the bottom of the valve disc, where the fluid is accelerated due to convergence.

https://drive.google.com/open?id=1QHAO_vpCWqWNJZqcGYotok1eDqzDRDER

Case 4 Results:

• A mass flow rate of 0.47458 kg/s is recorded as the output with the gate valve lift of 40mm. The residuals plot indicates the convergence is attained at around 250 iterations.

https://drive.google.com/open?id=1UxQ8RqPduFwJOJvNeubvrbXmAo27PpZ5

Parametric method results:

• Using the parameter set method, the following results were acquired for the given case. As discussed previously, the downside of this method is that not enough details in the mesh which might differ from one case to another or one design point to another can be added to keep the number of variations across the simulation at the least.
• It should be noted that only limited output which can be defined as a report such as fluxes, surface variables, volume variables can be determined this way. For contours, and animated distributions and more detailed results it is advisable to perform each of the case simulation individually.

https://drive.google.com/open?id=1sm8X9TbCbas0mRVGPPTf8WpcsFidUOKc

Conclusion:

• In this challenge we learned to use parametric method to study the effects of valve lift on the output mass flow rate, using the parametric option which makes things easier as we don’t have to manually apply the motion to the valve every time. This saves time and effort but more importantly automates the process for any number of cases.
• As long as the position of component which can be controlled by giving it motion is the only difference between all the cases this method can be useful. However, any differences in the parameters related to the component in question may throw an error for example having different sized mesh for the fluid in the valve disc portion wont be updated as the actual valve position does in space-claim. Hence, care must be taken.
• When compared, the results that are recorded from contours and plot distributions, through individual simulations are only a small bit more accurate than the results we get using the parametric study method for almost half the computation time. Hence, it can be concluded that for very specific results, where not much detail is required and lot of cases are to be studied, Parametric study method can be used effectively.

Drive link to the folder with all the files:

https://drive.google.com/open?id=1Zetu4pFZdu7qAAfV7ssYh1YcbIdyUKMt

Report ppt:

https://drive.google.com/open?id=1d1Zq8FdaT2V_BLLQYbdBc97SA7pIbLxk