Flowbench Simulation for valve flow

INTRODUCTION:

Flow of charges or the fluid is sometimes continuous and in other times, requires intermittent operation. One classic example is in the Engines where intake and exhaust valves open for the particular range of crank angle degrees to allow charge in and out of the system. Similarly, butterfly valves which have flap needs to open at particular angle when external souce excites it to allow designed amount of mass to flow through it which finds application in compressor bypass valve, wastegate etc.

In the former case, Cam used to control the lift of the valve, dwell and fall of the valve, whether its intake or exhaust, the concept is similar. The valve consists of three important parts, Valve stem, Valve lip, Valve seat. At 0mm valve lift, valve lip closes the port completely and prevents the flow of charge into the cylinder. As the lift increases, restriction to the flow in the port is reduced and fluid starts to enter the cylinder. However, there is one more aspect which influences the flow rate. The design of the ports itself, this results in actual mass flow rate to vary with the ideal flow rate due to various losses present. This is determined by the coefficient called coefficient of discharge(Cd). In order, to understand the behaviour at various lift or at angle opening in the latter curves, couple of characteristic curves could be generated like Angle vs Cd, Lift vs Mass flow rate, Angle vs Mass flow rate.

Since, this is an observed phenomena, it needs experimental investigation which is achieved by the simple experimental appartus called flow bench. In flow bench, ports or valves under investigation is placed in the control volume and the controlled flow is allowed to pass through either by controlling velocity or mass flow rate or delta pressure across the component under investigation. This results in experimental determination of curves. Similar exercise could be carried out through numerical experiments.

CFD SIMULATION

Geometry: A cylinder of 100mm diameter is taken under investigation with port of diameter of 25mm.

Meshing: Both automatic and manual setting is utilised and the grid dependency studies are carried out to determine the best possible grid for the CFD simulation. Following cases were setup for studying the mesh dependency

From this, we notice that at design point 4, values seems to converge to a point, 5th design point is highly refined and with 8 times the cells present in design point 4 with a very small improvement in the results. Hence the run time vs solution accuracy is optimised and design point 4 is taken for further study.

Final mesh chosen, Nx = 40, Ny= 100, Nz = 40, with local mesh refinement level of 3.

Solver setup: Water is chosen as the fluid under study at temperature of 20C with delta pressure across the valve face being 100000pa. Although this is an unrealistic delta pressure, it is used to better understand the results

For the paramteric study on effect of valve lift on mass flow rate and force exerted on valve face

RESULTS AND PLOTS

Plot1: Each of these plots show how the flow enters and velocity is developed in the section of cylinder, notice the formation of swirls within the chamber.

Plot2: However, the cut section doesnt give much info, hence the Mass flow rate vs the lift is plotted as shown, The results supports the theory of curve that is expected for flow through the ports for various lifts

DISCUSSION ON RESULTS

1) The maximum flow rate at the maximum lift is ~4.3kg/s and it stabilises

2) Flow curve can be divided into three secctions, Exponential rise, Linear increase and saturation

3) Valve force after certain lift remains pretty constant, initially the value is high, because of excess restriction and increased back pressure force acting on valve face and stem

4) However the above results means nothing and cannot be used for comparison between the valves of different sizes and hence the dimesnions are normalised using the parameter, L/D which is Lift/Diameter. This helps in comparing the valves of different geometry

Although the curve looks the same, the x axis is now normalised

5) Similarly, it is not ideal to compare the valve or port performance against different density and delta pressure, hence the mass flow rate is compare against the flow rate through the ideal orifice plate with no loss. Ratio of Actual flow through bench to Flow through ideal orifice plate,

Changing Y axis would lead to dimensionless quantity on both x and y axis which would be ideal to compare the performance of the valves and hence is defined as the performance curve

6) However, Flow bench has its own shortcomings, major one being that pressure is assumed constant in the flow bench, however in an actual engine the pressure difference woul be varying

7) Second most important shortcoming is that it is impossible to define the minimal or throat area as the it changes throughout the lift. This prompts use of calibrating factors in GT power models