water brake throttling options

[viper]

I think we have decided to give water braking a shot. Per our calcs, we will need to flow around 10gpm per 100-130HP to keep from boiling the brake water. Probably the only problem we face is how to accurately and quickly throttle the water. I remember Bruce mentioning something about using a waste gate and control for this. Not sure if that worked out or not. We had considered using a VFD on the water pump to regulate the water but concerned that because the pump will be approx 15ft from the absorber, this could cause problems in quickly controlling load.

Does anyone have any other ideas here? I have seem electric ball valves but not sure if those would even be fast enough for this application. Not sure where we will end up here.

[Tony]

The wastegate idea should be practical and be very easy to computerise. Basically you use software controlled air pressure to open an external wastegate in the usual way, against it's internal spring.

Software could pulse width modulate a small air solenoid valve to control the air pressure acting on the wastegate diaphragm. This should be dead easy to make, have very high water flow capacity, be very fast acting, and have smooth
AND LINEAR control characteristics. Exhaust valve materials don't rust either. For low cost and simplicity, this would be difficult to beat.

The other way is as you suggest, a variable speed positive displacement pump. A VFD with a serial data control cable would be a neat if expensive solution.

I believe Land and Sea use a horribly expensive stepper motor operated water control valve.
That obviously works too. But you would probably have to buy the proper control valve and the electronic interface circuit board to run it.

[jsa]

Electric actuated ball valves I have seen,used in airconditioning systems, have a 90 second travel time.

I imagine they would be way to slow.

[viper]

So has the butterfly valve (wastegate) been proven yet? We were leaning more towards the stepper or servo motor setup but I think that is just because we are more familiar with them.

I am a little concerned that an air actuated valve could be affected slightly by different flow rates and pressures but maybe any hiccups can be tuned out.


Wonder if you could PWM a basic air solenoid??? I guess, never tried it.

[Tony]

A wastegate is almost always a poppet valve, not a butterfly valve.
The flow area of a butterfly valve is highly non linear with the opening angle. Any non linearity hugely complicates the design and tuning of the dyno closed loop speed/load feedback system.

Basically you do not need to know (or care) what the pressure or flow is going into the water brake, or how far the control valve is actually open, except that there is enough water flow to prevent boiling at full maximum power.

You dial up (say) 4,000 rpm and then open the engine throttle. The engine speed rapidly increases, and then the water valve opens up just far enough to hold exactly 4,000 rpm. The engine speed/load feedback system opens the water supply just far enough to hold the required set 4,000 rpm speed exactly.

The set speed may be constant, or it may slowly sweep upwards at a desired fixed rate.

Water flow is always determined by the outlet restriction.

In other words, if the drum needs to be half full to hold the engine, water flow could be one gallon per minute, or 200 gallons per minute. When the water volume going in equals the water volume forcing its way out past the outlet restriction, then the water level in the drum will remain constant.

So flow is actually set by the outlet restriction. The wastegate only opens far enough to hold the water level in the drum constant at the required level to hold the torque from the engine.
For it to do this, the water valve needs to respond quickly, smoothly, and open (and close) in a very progressive manner.

It is a bit like like driving down the road at exactly 50 mph. You never need to know how far open the throttle is to do that. You just watch your speed, and give it a bit more, or a bit less throttle. A dyno is the same. How far open the water valve is, is of no real interest. The speed/load control system just gives it a bit more water, or a bit less water, to vary the load, to hold the required exact set speed.

[viper]

Sorry Tony, I had an exhaust brake in my head for some reason. I would agree that actual position is irrelevant and and encoder could just complicate things though might help with R&D to a degree. I was concerned of surging or flutter of the valve by using air but I presume that could be easily tuned out with proper valve spring tension.


Has this idea been tested yet?

[Tony]

As far as I know this has never been tried.

It is something I put a great deal of thought into a very long time ago when I was hoping to build my own home water brake chassis dyno. Since then, I was able to buy a low cost Telma eddy current roller dyno at a price I could afford. So the whole water brake idea has not progressed any further.

You are very right to be concerned about control characteristics, speed of response, flutter, overshoot, backlash, and sticking. Wastegates have all these exact same control problems for controlling boost pressure. They work great in that application, and should work just as well controlling water flow.

Several things to think about.
The curtain area of a poppet valve opens up in direct proportion to valve lift. And so will water flow be controlled over a very wide range. The heavy return spring is also linear, the poundage increases directly with valve lift. And lastly the air diaphragm is VERY directly connected to the valve without any high inertia motor and gearbox, static friction, or backlash. The result is a very simple, robust, fast, and linear opening flow control valve.

I honestly cannot think of anything better or more suitable to use with a water brake
Controlling the wastegate actuator diaphragm air pressure with a computer is a separate problem, but one quite easily and cheaply solved.

[viper]

Have you seen positive, successful results with a variable speed pump? We are really concerned that the pump would need to be VERY close to the brake to perform correctly, otherwise there could be a delayed action to the brake. I think this method of control just seems easier and a more accurate way to go.

We have the plan to keep our closed loop water source 10s of feet from the brake some we have to consider consequences of that.

[Tony]

If you are worried about water hammer, and rapidly changing the velocity of water flow in a long pipe, then just increase pipe diameter up to the control valve or pump.
The pipe then just becomes a very low velocity extension of the water storage tank, and has minimal stored kinetic energy.

[viper]

Well, getting back to the valve idea, one of the thoughts I have is how in world the pwm control of an air cylinder is going to work without exhausting air for reduction of pressure? I mean, say to need to dial in 30psi of air and then it needs to drop back to 25, even if the solenoid modulates the correct frequency, it will remain at 30psi until something releases the air. I have obviously not thought this through so just curious if you have? Seems the exhaust or decrease of pressure could be an issue.

[Tony]

The way to do this is to place two flow restrictions in series, one fixed the other variable.

Now imagine you have two very small restrictive air holes, (like carburettor jets) placed in series, and connected to (say) a constant 50 psi air source. They will both have exactly the same mass air flow, and therefore exactly the same pressure drops. So there will be 50psi at the inlet, 25 psi between the two jets, and atmospheric pressure at the discharge of the second jet. That is with two identical air jets. The pressure will always split evenly with two identical jets.

If you replaced either one of those air jets with an adjustable needle valve, you could then adjust the mid point pressure anywhere from 50 psi to atmospheric by just turning the needle valve either way.

Replace the needle valve with a small air solenoid valve driven with a rapid on/off digital signal. By varying the on to off ratio, you can again adjust the mid point pressure anywhere from zero to the full 50 psi electrically.

You just tap off the mid point pressure, and run it to the wastegate actuator, and it will see the full range of pressure. Provided the total air flow is sufficient, and the dead air volume reasonably small, it can respond very fast to pressure changes produced by the pulse width modulation.

[viper]

I had wondered if a simple solenoid valve could work with a pwm control but kind of concerned that the pulsating effect would cause load cell issues at lower pwm frequencies. Have you looked at this angle? That would have a nice, linear response and should perform well at higher frequencies.

I am still kind of confused as to how an air system could work because once a system is pumped up, it has no way to pump down without venting to atmosphere. The orifice restriction would only induce a pressure drop if there is flow. Were you thinking of having something venting to atmosphere all the time? Guess I am a little confused on this one.


Well, I guess I can see how you could use the pwm control to vent pressure to atmosphere. That would surely work but would be bleeding air down constantly. Not much air needed but still.

[Tony]

It does vent to atmosphere all the time.

Look at this another way. One air restriction flows continuously, pumping up the dead air volume inside the wastegate. A second similar air restriction continuously bleeds off air pressure to atmosphere, Air flows continuously in and out of the wastegate, and the wastegate pressure then sits at exactly half the air supply pressure.

If either one or the other air restriction changes in flow, then the wastegate pressure changes. It will not pulse if there is enough total dead air volume. This not only works, but similar PWM pneumatic pressure control systems are used for all kinds of applications.

[viper]

Yeah, I think we are on the same page, for the most part. I am envisioning an air cylinder circuit without a normal vent to atmosphere (single acting). If that circuit has a vent to atmosphere that flow the same as what is coming from my compressed air source through a pwm control, there would be no pressure change but as soon as my flow rate from the controller exceeds the flow rate of my vent, pressure will rise on the air cylinder.

Is that how you see this? Basically just going to be venting all the time but quantity of air would be rather minimal.


It seems many are using PWM control means on solenoid valves. Have you played with this at all? This would be a direct acting valve.

[Tony]

I have in the past, experimented with these systems, and it just requires a bit of tuning and experimentation to get something working at it's best.

You could start with a compressed air source, feed that through an adjustable needle valve flow control, into a pressure gauge. Then vent the gauge through the PWM air solenoid being tested through a second adjustable needle valve.

How the air solenoid itself behaves is something else. Voltage and PWM frequency will need experimenting with too.
Try a few different small air solenoid valve types.
With a bit of fiddling around the system can usually be made to work pretty darned well.
If the pressure gauge is a bit lively, a dead air volume (surge tank) will calm it down and reduce pulsing.

To do this, you ideally need a function generator, or some other means of having a variable on/off ratio square wave 0% to 100% that you can adjust over a fairly wide frequency range.

The wastegate air pressure does not need to be controllable right from zero all the way to the full supply pressure, but it would be really nice if it did.
You will discover that a wastegate requires a fair bit of control air pressure just to lift the valve off it's seat, and not much higher pressure to drive it right to full lift.
That means the PWM solenoid only has to work really well over a fairly small part of it's full range, which should not be difficult.

It is just a case of a bit of experimenting to optimise the parts you have. Once the system is working really well, the needle valves can both be replaced with fixed small air restrictions. Software can be written to generate a PWM signal with appropriate timing and frequency over the required operating range.

This is a great do it yourself home project, because although there will be a lot of time, messing around, and experimentation required, the parts to do it are fairly readily available and not too spendy.