Air Flow 101 - applicable to engines
 

Following the rect vs. oval port thread got me thinking that some members might like a little background on airflow in pipes, ducts and of course ports. There's a lot more going on inside an engine besides simple steady state air flow (pulsing, resonance, swirl, mixture quality etc.) but we have to start somewhere.

Air flows between two points when there is a pressure difference between the two points. This pressure difference is called Static Pressure (SP). You must expend energy to create this SP, and nothing happens without it. This pressure difference between the two points results in a force being applied to the air and the acceleration of the air mass.

Imagine the piston drawing down (energy expended) and creating a vacuum (negative static pressure) under the closed throttle plates. Now open the throttle plates to allow the difference in pressure between the atmosphere and the cylinder to act on the air in the port. Air in the port is set in motion.

The air mass or volume flowrate in CFM (Q) and the velocity of the flow in fpm (V) are related according to the equation Q = V x A, where A is the cross-sectional area of the pipe in square feet. Therefore, if Q is constant, velocity goes up as area goes down.

Now that the air is travelling at a specific velocity it exerts a Velocity Pressure (VP) in the direction of air flow. This means that some of the Static Pressure (SP) has been converted into Velocity Pressure (VP). The amount of velocity pressure is equal to the energy required to accelerate the air.

So everything is moving along nicely in our straight pipe, and if we reduce the diameter of the pipe, static pressure goes down, velocity pressure goes up, and if we increase diameter back to its original value, both static pressure and velocity pressure go back to their original value, right? WRONG! Velocity and velocity pressure go back to their original value, but we lose some of our static pressure. Why?

Because air flow in pipes encounters resistance to flow due to friction and turbulence. Friction losses are caused by the roughness of the surface and turbulence losses occur whenever the air flow changes direction or velocity. This means that some of our original pressure difference gets used up fighting the resistance to flow created by these losses. In fact these losses are expressed in the same units as static pressure (psi or "H2O).

This is important to us because a pump, fan, or other air-moving device (engine) operating at a given RPM usually provides us with a fixed, limited amount of pressure difference (or force) to apply to the air we want to move. The more static pressure we waste on these losses, the less we have to convert into air motion. Less air = less power.

Now here's where it gets interesting. It turns out that these losses are proportional to velocity pressure, as shown in the following equation:

Pressure loss = K x VP

where K is the loss coefficient for the element causing the loss.

Velocity pressure is in turn proportional to velocity squared. So if we reduce the size of the port, velocity goes up, velocity pressure goes up by the change in velocity squared and therefore resistance to air flow goes up by the change in velocity squared. Sounds like a good reason for large ports, doesn't it?

Velocity is bad for air flow, because you use up too much of your static pressure fighting the increased resistance. In steady state flow, increasing port area reduces velocity, and according to the above equation, reduces the pressure loss, leaving more of the original pressure difference for setting air in motion. More air = more power.

So why is everyone always talking about velocity in the port as a good thing? Well it's mainly because air flow in engines is not steady state, and we are continually stopping and starting the flow. Small ports keep the column of air smaller so there is less mass to accelerate, and once accelerated to a higher velocity, the air is more likely to continue at that speed. This is a good thing for filling the cylinder in an engine. Which brings us to the subject of inertia.

When the column of air in the pipe is at rest, it wants to stay at rest. Once the column of air is moving, it wants to stay moving. Starting the air mass moving from zero velocity cost us some energy. If we then slam the valve shut, this energy is wasted, isn't it? Well, not exactly.

By carefully timing the closing of the intake valve we can allow the column of air to continue to fill the cylinder even after the piston is past BDC, and no longer providing the pressure difference that started the column moving. This is sometimes called inertial supercharging (thanks JimV) and can result in greater than 100% cylinder filling or volumetric efficiency. This is why the cam timing event that has the most effect on the location of the torque peak is intake valve closing. Since this post is definitely too long already, that's enough about cam timing.

So back to pressure losses and how to reduce them. If you are injecting fuel directly into the cylinder and don't care about velocity as a way of keeping fuel droplets from the carb in suspension, then bigger IS better. If you are supercharging and don't care about "tuning" the port for resonance and improved cylinder charging at a given RPM then bigger IS better. But if you do care about these two factors then you have to reduce pressure loss another way, without increasing area and reducing velocity.

Looking at the pressure loss equation the answer is obvious; you must reduce the loss coefficient of the elements causing the loss. If we ignore friction losses, there are three elements that cause losses:

1) Changes in direction (elbows)

2) Changes in velocity (expansions and contractions)

3) Obstructions to flow (valves)

This is what porting is all about. Reducing the loss coefficient to reduce pressure loss, without resorting to larger area and slower velocity. Increasing the radius of an elbow can reduce the loss coefficient by up to 50%. This is what raised ports accomplish. Smoothing the transition when area is reduced (from carb plenum to intake runner) and when area is increased (from valve seat to combustion chamber) can also reduce the loss coefficients in those areas. Raising a valve farther off the seat reduces the loss coefficient for this obstruction.

Of course, this is all pretty simple steady state stuff. Unfortunately, there's a valve opening and closing in this flow path, and other cylinders pulling off the same manifold, and cam timing issues and so on and so on. But reducing loss coefficients is how you make a small port give you more air flow without giving up the velocity that improves inertial supercharging and combustion efficiency.

So, whadda ya think of Tiger Woods' new girlfriend?:D

If you're asking me whether she sucks or blows, I don't know, I haven't seen her.:p

what the...??? I should have studied more in high school...

Damn, Tomcat, you're putting some technical stuff out. :confused:
Those long cold winters must give you to much time to think. :D
Keep'em coming makes for good reading.

mike

Good stuff Tomcat, I like the way you think! ;) :cool: :)

Heads
 

Holey crap wud he say?

tomcat,
Do you have a show on the Discovery Channel yet?
Good info. Much better than the technical term like "stuff" that I use.
Gary

I always thought that large intake ports would work really well for supercharging too (not to mention the low factory compression ratio)--I think thats one reason why the big blocks make more power so easily. I've also wondered how the engine speed relates to the inertial mass effects of the larger air column--seems it would have less of an effect the faster you turn (the air has less of a chance to react).

I have to apologize for the length of that post. I get carried away in my desire for "completeness" and making things "easy to understand". In fact I just added more to it. But I've read so many magazine articles about heads and cams that just don't take the time to explain things properly. Rather than take short cuts, and over-simplify things, and jump to conclusions, I'd rather start at square one.

It's not that everybody needs to know exactly how and why things work. We can just learn from experience what works, and pass on our hard earned knowledge in perfectly acceptable non-technical language like "the air in large ports is lazy". After a while, we get a gut feeling for what works and what doesn't. But when you read the Rect vs. Oval port thread and see how hard it is to kill the idea that bigger is automatically better, you see that the basic principles are not well known. So why not educate those that don't have the opportunity to learn from experience?

The enemy is pressure loss. The easiest way to kill him is bigger ports, valves and cams. But you sacrifice velocity and all the good things that go with it. So reducing loss coefficients is the way to go. This applies to more than just head porting too.

Finally, the basics are just that, nothing more. They can't predict all the possible combinations and outcomes. Experience (meaning actual data) is always the final reality. Anyway, I accept the good-natured ribbing and appreciate the opportunity this forum gives me to play high school physics teacher.;)

Hey Corey - You're right on large ports for Roots or screw compressor supercharging. First of all, you are expending a little more crank power to increase the pressure difference with the supercharger, which overcomes the pressure losses in the "pipe". You can get away with the larger ports to reduce the pressure loss, because the mixing of fuel and air by the compressor rotors compensates for the lower velocity (and less mixing) in the ports. Blowthrough superchargers like Procharger or Vortech do not provide this "egg beater" effect, so I think you might not be able to use as big a port.

At higher RPM there is less time for acceleration of the air mass and inertial cylinder filling. That's why cam duration is increased for high RPM engines. The down side is that at low speed the intake valve stays open too long and the piston starts pushing air back out of the cylinder into the intake manifold. Result, rough idle and poor low RPM power. The obvious answer is variable valve timing.

OK, what about a blower motor and the rectangular ports.

BTW - "blow" is a figure of speach...

TOMCAT, Good post , but static pressure is the force being exerted outward against the "port" walls or "pipe walls", it is not why there is flow but it is a result of the flow.

David - You caught me breaking one of my own rules, over simplifying. I didn't explain total pressure (TP). The energy expended in drawing the piston down can appear as either static pressure or velocity pressure. Static pressure can be either positive or negative with respect to the atmosphere; velocity pressure is always positive.

TP = SP + VP

So you can think of total pressure as the total energy put into the air in the pipe by the fan, pump or engine, and whether this total pressure appears as static pressure or velocity pressure depends on how much flow is allowed by the pipe. Ignoring other losses, the amount of flow is controlled by whether or not the pipe is open at the inlet.

In the case where the fan, pump or engine are drawing through the pipe, both total pressure and static pressure are negative, less than atmospheric pressure. Just for illustration purposes, assume that the piston has drawn down and the throttle is closed; there is no flow and therefore no velocity pressure, so total pressure and static pressure are the same, both negative, and VP is zero.

TP - SP = VP = 0

It is this static pressure that will cause flow when we open the throttle. Now open the throttle and flow begins. This means there is now some velocity pressure, but static pressure and velocity pressure still add up to total pressure (in this case total pressure becomes less negative).

This is the actual relationship between static pressure, velocity pressure and the total pressure (difference) created by the descending piston.

It's actually easier to picture this if the pipe is under positive pressure with respect to the atmosphere. I will try to post some diagrams that illustrate this.

Ok Tom,I think its time you took a holiday down south.The cabin fever has definately got ya.Great info but man do you need a boat ride:D

Airpacker - Cabin fever is right, there's still ice on the bay and we got 8" of snow yesterday! I can't take it anymore!!!

Tom

Diagram #1 - from the Industrial Ventilation Manual - ACGIH

Diagram #2 - The fluid in the pressure measuring tubes is water. This is where the pressure units "inches of water" comes from. Question: "Is it possible to have a pipe configured such that the SP measures zero and all the total pressure is converted into velocity?"

Tomcat......now this question does relate to air pressure (kinda), and oughta be a wiz for ya.

If I want to take a vac reading and coorelate to what powervalve I might want to run, what is the difference in vacuum between taking it from a runner vs the carb base. Or better yet, where is the proper place to take the reading.

...I'll give you about 10minutes max, that includes an answer and a chart. ..................... GO


Bob

At the carb base.

David's right, the plenum averages the different and fluctuating vacuum in each runner. Chart? Do I use charts?

Okay, now that everyone has the basics under their belt, let's put this knowledge to use and get some perspective on air flow and pressure loss in cylinder heads compared to other "pipes".

For this comparison I will use actual data for three BBC cylinder heads (good, better and best) measured out of the box on a flow bench at 0.600" lift and 28" H2O, with 2.25" intake valves and ~320cc port volume. To get some perspective on these numbers we will compare the performance of these ports to that of a sharp elbow and an ideal bell mouth inlet with the same cross sectional area as the valve window (~2.3 square inches), and the same 28 "H2O of pressure difference applied.

Head........Flow....Velocity.....Ce.......VP.....L oss........K..

GM401......310.....10,526......0.50.....6.9...21.1 ......3.05
DartPro1...325.....11,035......0.52.....7.6...20.4 ......2.69
AFRCNC.....375.....12,732......0.60...10.1...17.9. .....1.77
Elbow........395.....13,412......0.63...11.2....16 .8.....1.50
Bell............612.....20,768......0.98...26.9... ..1.1......0.04

The velocity is the air speed (in fpm) through the valve window (very high). Ce is the coefficient of entry; when it's equal to 1.00, the conversion of static pressure to velocity pressure (flow) is perfect (no resistance due to turbulence). This is impossible, but a bell mouth inlet comes damn close at 0.98. You can see that VP and Loss always add up to 28 "H2O, the flow bench test pressure. See how small the resistance loss is for the bell mouth. That's why its K is so low. K is the number we are trying to reduce when we port heads.

A K factor of 1.5 is about as bad as we get when we design industrial ventilation systems. A sharp, square elbow has this K factor and therefore uses up about 60% of the available pressure difference to overcome losses (at this very high velocity). You can see that the new Air Flow Research CNC heads, as good as they are, do not flow as well as the worst element encountered in industrial ventilation.

This is not a criticism of head manufacturers or head porters. It is the nature of the beast. These AFR heads flow almost as well as a fully (hand) ported head (without increasing port volume). Further improvements may come with raised ports, different valve angles and better chambers. But as long as poppet valves are in the flow path, this is what we have to work with.

It wasn't really fair to compare the multiple losses of the port and valve to the single element of loss in a sharp elbow. The head actually has four losses. Based on my information these are the approximate losses for each element in the AFR CNC head or any fully ported head that flows 375 CFM:

Element............K.........Description

Elbow..............0.25.....Curved port
Contraction.....0.25.....Just before valve
Obstruction.....1.00.....Open valve
Expansion.......0.25.....Into chamber and cylinder

Total................1.75.....Not bad at all!!!

The high loss for the obstruction presented by the open valve suggests that the area just before the valve, the valve seat and back of the intake valve, and the chamber immediately after the valve are the most important things to work on. This agrees with what I have been told about mild porting (concentrate on the bowl and the valve seat machining). More recently I see the importance of flow in the chamber being recognized in magazine articles.

Darn Valves! :mad:

I've never seen a breakdown like that--neat. It is easy to see why the areas right around the vlave is so critical when looking at this comparison.