Simulators do not understand real world...

This is maybe a catchy title but this is what a title should be about, isn’t it. This post is the first post about the importance of rigorous calculation compare to the simpler approach from a simulator. Don’t get me wrong simulators are very good and many processes engineers are using them every day but they do have limitation and it is the job of the engineers to know what these limitations are and how to overcome them. I created a very simple case of a heat exchanger in one of the most famous simulator in the oil and gas industry HYSYS.

People familiar with the tool will recognised that the tool has calculated the heat exchanger and that everything is fine. From the simulator point of view and an ideal mass-heat balance this is correct. We have a stream coming in at 90C and leaving at 20C, on the other side we have the exact opposite. So in summary all the heat from side is transfer to the other side, and all this without any pressure drop. Now if you give this specification to a heat exchanger expert, he will struggle to design such an exchanger. Lets have a look at why.

The first aspect which I see quite often forgotten is pressure drop. When a fluid is flowing through equipment, it has to lose pressure. In the simulator you can specify the pressure drop that you want and 0 is perfectly adequate. This means that either you expect it to be very small and that you plan to add some equipment to compensate for the loss of pressure somewhere on the line. However when you start looking at a more detailed view of the equipments the pressure drop has to be taken into account. In a very simplistic view, you can say that the pressure drop is used to create heat transfer. So the more pressure drop you have, the more heat transfer you have.
The second aspect is that the world is far from ideal... I won’t go into the details of how the heat transfer works but not all the energy from one stream can go to the next. It has to cross materials, loose heat to the outside of the equipment...

But more importantly, you need a driving force. The smaller the temperature difference is the smaller the quantity of heat transfer will be, in other terms, the smaller the temperature difference is, the bigger the exchange surface needs to be to exchange the same quantity of heat. It is a bit like if you try to reach 2 when you start at 1 and add half of the previous number 1 + 0.5 + 0.25.... Mathematically you will reach 2 only after an infinite number of additions. So in theory you will need an exchanger with an infinite surface. Each different type of exchanger needs different driving force. The engineer needs to choose the correct exchanger for the correct job. In a “normal” shell & tube exchanger it could be 5C when in a specialized plate fin exchanger it can go down to 0.1C
So a more realistic exchanger will be

This exchanger has some pressure drop on both sides, and a minimum temperature approach of 5C.

Dry-out counterintuitive phenomenon

A second post because it is my grand opening :)
The other day I had a long discussion with a customer. They had an existing model working and wanted to test some different scenario. This is something that I see more and more happening in the industry but this will be another post. One test was to run the same design with high fouling. Fouling can cause a severe reduction of performance both by reducing the duty and increasing the pressure drop. You can model both effect, but most of the time it is very difficult to know the conductivity and the thickness of the fouling so only the reduction of duty is calculated.
The fouling was added and the heat exchanger was performing better. This could only mean one thing the modelling was wrong... or there was a different phenomenon occurring. I suspect that I need to say that one side of the exchanger was boiling. Boiling heat transfer coefficients are usually very high compare to single phase. We could spend a lot of posts about boiling, but in 1 sentence: the hot side is providing a heat flux through the material which is used to boil the cold side.

The type of boiling depends of a lot of parameters including the heat flux mentioned earlier. If this heat flux is too big (the limit is very creatively called “critical heat flux” – we are engineers...) a phenomenon called dry-out occurs. During dry-out, the liquid is heated so quickly that a film of vapour is created next to the material. This vapour creates a layer with a very low heat transfer coefficient. Modern tools will calculate the critical heat flux and modify the local heat transfer coefficient to take account of the high resistance layer.

Now what is happening if you add fouling to the previous case? The heat reaching the fluid is reduced because of the extra resistance of the fouling and you could fall under the critical heat flux, and get back to a region with a high boiling heat transfer coefficient.

So in this particular case, adding fouling was improving the performance of the heat exchanger because the performance was restricted by a dry out zone.

Drawbacks of the UA Model

Quite often when I talk with process engineers, they are quite happy with the modelling of heat exchangers done in the process simulators.

Most this modelling is done via an approximation: a constant UA value.

This approach is perfectly adequate if you do not change your process conditions. When you go from a single phase application to a 2-phases application the coefficient U is changing dramatically; but even in a single phase application, the coefficient can change.

The property of the fluid can change with the temperature and pressure so changing the conditions, will change the coefficient.

Even if you consider the same temperature and pressure, the flowrate is enough to change the coefficient. The coefficient is linked to the reynolds number:

this mean that if you change your process conditions, the local heat transfer changes then the overall U value alters, which means that as the UA is constant, the Area MUST change!

I don't think that there are a lot of exchangers that can provide this feature.