Ever had to determine the pressure drops to assign to the control valve while doing hydraulics? This post will help you specify the control valve pressure drop on a rigorous and sound basis.
For this post, I’m going to assume you can work out the various flowrates and pressure drops the system will be subjected to. You know how to read the “pump curve” or “compressor curve” if said equipment is involved. (Good articles: Part 1, Part 2) Once you know the range of flowrates and pressure drops required of the valve, you can select the correct Cv and correct valve or get the help of an instrumentation engineer to do this.
This post will be focused on the initial hydraulic design of a system, when you first need to select a pump and also figure out the pressure drop you will allocate to the control valve. Let’s examine a simple hydraulic circuit problem.
Liquid flow system: Drum 1----Centrifugal Pump----Control Valve----Drum 2.
Drum 1 and Drum 2 are flash drums with their own independent pressure controls, and the control valve is installed to vary the flowrate.
Your overall hydraulic equation is: at the normal flowrate: Pressure at Drum 1 + Pump’s Pressure Addition – Pipe/Valve/Equipment Frictional Losses at selected pipe size and route + Static Head Pressure Effects – Control Valve Pressure Drop = Pressure at Drum 2
Written in variables: At Qnorm: P1 + P + SH – f – CV = P2
Known: P1, P2, Qnorm (These are usually known by the heat and material balance)
Calculated for a given pipe size & equipment configuration: SH, f
Note that SH may be positive or negative. We discussed the calculation of "f" in Determine Pressure Drops in Straight Pipe. If you have equipment that can cause friction losses like gate valves, heat exchangers, orifices, etc. their losses are included in the "f" term.
To be determined: P, CV
How to proceed after calculating SH and f
The next step is to assign a value to CV. Once CV is specified, we will know the P value we need, allowing us to select a pump.
But first, let’s take a step back and remember why we have a control valve in the first place. If the pump is a centrifugal pump, then for any given model of pump, the flowrate through the pump depends on pressure provided by the pump (depends on the value of P). We know this from the pump curve. Once we install a control valve, the control system will act to vary pressure drop CV. In operation, when CV changes due to control action, P will change to balance the overall pressure equation, thereby varying the flow.
For example, CV increases from 15 psi to 20 psi while P1 and P2 are kept constant. Therefore P will increase to overcome the extra pressure drop while still forcing the liquid into Drum 2. This higher value of P will reduce flowrate Q. Alternatively, if CV decreased from 15 to 10 psi, P would be able to decrease and Q would increase.
Therefore we conclude that the pressure drop you assign to CV is the value that your controller will vary to change the flowrate. The higher the CV value you design for, the greater your ability to vary CV and therefore control the flow. But high CV values also waste pumping energy across the valve. A tiny CV value means little energy is wasted, but there is little scope to change the pump’s flowrate.
The size of the term CV is the margin of control allowed to the control valve.
A typical “rule of thumb” when designing control valves is the “one third rule.” The rule says: CV = 1/3*f. Then calculate P. 1/3rd was found to often be a good compromise between having a good control margin while not needlessly wasting energy. Additionally, if CV < 5-15 psig, you might consider increasing CV to at least 5-15 psig because practically speaking, it is hard to get a control valve with less pressure drop than that.
Problems with the one third rule
Unfortunately, the one third rule is not always reliable because:
- You may have a system with almost no flowrate and pressure fluctuations, or very high fluctuations, in which case the 1/3rd rule may leave you too much or too little leeway in the control valve
- If P1 or P2 are >> f, small fluctuations in drum pressure could eat up the margin of control you have left in CV
- A vague rule of thumb is hard to defend. If you are pushed to reduce the control valve pressure drop allowance to save money, you might cut so deep that your system cannot control itself
A better procedure to follow is:
- Work out the different flow scenarios you may be subjected to, including the maximum design flowrate and the minimum turndown. If you have no other basis, a typical refining rule of thumb is Qmax = Qnorm * 1.1. If you expect large fluctuations (like in tower pump-arounds, or when controlling level on small drums) then Qmax = Qnorm *1.2. For turndowns, rule of thumb is Qturn = Qnorn * 0.6. Utility systems may well have a larger range of flowrates and need much larger design margins.
- Presumably P1, P2, and SH are the same at all flowrates. f, the piping frictional losses, can be recalculated at the new flowrates. If in doubt, use the calculation fmax = fnorm * (Qmax/Qnorm)^2. This will roughly let you scale frictional pressure losses to new flowrates. (There are tables out there that will suggest exponents slightly different than 2 for different kinds of equipment. For example, a certain heat exchanger type might use (Qmax/Qnorm)^1.8. But using ^2 is a decent general approximation).
- At the maximum flowrate, add a small pressure allowance for the pressure drop inherent in the control valve itself, like ~4 psig for a typical globe valve. Then add a few psi extra so that the control valve has some small measure of control even at the maximum flow. So, depending on the valve type, you may want CV = 5-10 psi at maximum flowrate.
- Ensure that CV chosen at step #3 is > 5% of P1 and P2. If not, consider increasing CV to 5% of P1 so that it can cope with fluctuations of P1 and P2. Some would prefer to add the 5% regardless, just to ensure that the control valve can handle fluctuations in upstream pressure. (So you could set CV at max flow = a few psi + 5% of P1).
- Having assigned a CV pressure loss at rated flow, you can determine the P you require at rated flow. Then find a pump able to provide this
- Time to consider normal flow and turndown flow. P will vary based on the pump curve. If you have no pump curve yet, you might assume P is constant at all flowrates as a first approximation
- Calculate the new value of CV value at your normal and min flowrates, now that you know P at these flowrates
So now you should have normal, max, and min flowrates as well as normal, max, and min control valve pressure drops. This gives you a better picture of the range of control you need and lets you specify the valve properly. Ensure that the control valve is not at extreme stoke, like <10% or >80% open, so that it is still controllable even at the extreme cases.
You can extend the basic principles of this example into many other types of hydraulic circuits. Here are three special cases to watch out for:
- If P is not a centrifugal device, the whole control strategy changes. For positive displacement devices you cannot use suction or discharge throttling effectively, so instead you will control fluid recycle or or control the speed of the positive displacement device itself
- If there is no pump or compressor, P may be 0 in all cases. (The system may still change, like the control valve may need to handle fluctuations in pressure or liquid levels(static head))
- If the fluid is vapor or 2-phase, you may have to do iterative calculations because so many properties depend on pressure. (Fluid properties vary by pressure, meaning f varies by pressure. So you must first guess at an overall pressure profile to work out f, and then refine your estimates of P and CV)
2010-09-19 - Rewrite suggested steps for better clarity.
2010-12-15 - Better explanation of determining Q normal and Q turndown, and positive displacement devices
2011-04-09 - Check out this article, if you have a Chemical Engineering Magazine subscription: Realistic Control Valve Pressure Drops