China: How does pressure swing adsorption removal work? 

This is a topic around which there is so much talk, but in essence it is often reduced to simple pumping and pumping. In fact, if you dig,pressure swing adsorption removalis a whole philosophy of balance between speed, purity and cost. Many people think that the main thing is the adsorbent itself, they say, they bought expensive zeolite and the process began. But the key is often something else - in the cycle itself, in that very variable pressure, and how to implement it in a real, not paper, installation.

Not a single adsorbent: where the devil lies

I sometimes look at the diagrams sent by colleagues from new projects. The columns, arrows, and valves are beautifully drawn. And if you ask about the cycle time, especially the time of desorption and purging - silence. But this is where efficiency sits.Pressure swing adsorptiongood for gas separation where differences in vapor pressure or adsorption kinetics are small. For example, the separation of hydrogen from associated gases or the dehydration of natural gas. But if you do not correctly calculate the pressure reduction (blowdown) and purge phases, you will end up with not a pure product, but something in between, and even with huge losses for regeneration.

I had experience, about five years ago, with one installation for obtaining nitrogen from the air. The customer complained about low purity and high purge gas consumption. We came and looked. It turned out that the designers had set too short a time to equalize the pressure between the columns. The adsorber did not have time to unload normally before the desorption phase; some of the heavy components remained, and in the next cycle they went into the product line. We had to reprogram the controller, literally increased the pauses between valve switchings by seconds - and the parameters returned to the passport values. It's a small thing, but it solves everything.

Or here’s another thing with the purge gas itself. Often it is taken from a product line, this is logical. But if you do not ensure its sufficient purity and stable pressure at the inlet to the regenerated adsorber, desorption will proceed sluggishly and incompletely. The adsorbent will coke over time and the capacity will fall. Then you have to stop the line and carry out thermal regeneration, which is expensive and time-consuming. So the system for preparing and supplying the purge flow is not a secondary piping, but part of the heart of the installation.

Equipment: Not all valves are created equal

Speaking about implementation, one cannot fail to mention the fittings. Quick-acting valves are, one might say, the joints of the entire system. Their wear, response speed, and tightness when closed directly affect the economics of the process. If you install cheap valves, leaks begin between pressure stages, mixing of flows, and a decrease in separation efficiency. This is especially critical for processes producing high-purity products, where even trace impurities are defective.

I remember at one of the polysilicon production plants they used an installationhydrogen drying and purificationjust using PSA (pressure swing adsorption) technology. The problem was micro-leakage through the spool of one of the key valves. Hydrogen is like that, it will find a gap. As a result, the dew point in the dried stream periodically jumped. We searched for the cause for a long time, the sin was on the adsorbent, on the cycle program. But it turned out that there was mechanical wear of the seal in the valve, which was located on the pressure equalization line. We replaced it with a valve with tighter tolerances for tightness - the problem went away.

Hence the conclusion: when designing a system, you need to install not just a valve, but a specific type, with specific characteristics in terms of the number of cycles and tightness class. And, what is important, provide for the possibility of diagnosis and easy replacement. Because they will fail - it is a matter of time and operating conditions.

Adsorbent: selection and degradation

And, of course, the core is the sorbent itself. Zeolites, activated carbons, molecular sieves. The choice depends on what we are removing: water, CO2, sulfur compounds, hydrocarbons. A common mistake is to try to solve all problems with one adsorbent. For example, zeolites like 4A or 13X are excellent for deep gas drying. But if the gas contains polar molecules heavier than water (say, mercaptans), they can irreversibly clog the zeolite pores. You need either a preliminary layer of another sorbent, or a completely different scheme.

I observed an interesting incident at a gas processing plant. There was a standard two-column gas dehydration unit before low-temperature separation. Suddenly the cycle between regenerations shortened sharply, and the dew point in the water crept up. We figured it out. It turned out that the composition of the raw material had changed - traces of glycol-based hydrate formation inhibitors appeared in the gas. They, being heavier and having a high affinity for the zeolite surface, displaced water during the adsorption process, and were not completely removed during desorption. Gradually, the sorbent lost its water capacity. The solution was not to urgently replace the adsorbent, but to install upstream a simpler carbon scrubber filter to capture these heavy impurities.

Degradation of the adsorbent is an inevitable but manageable process. In addition to chemical poisoning, there is a purely mechanical one - abrasion due to sudden changes in pressure and vibration. Especially in large adsorbers, where the sorbent layer is high. It is necessary to provide systems for uniform gas distribution at the inlet and high-quality supporting gratings to avoid channeling and fluidization of the layer. Sometimes you see how, when opening the adsorber after a couple of years of operation, the top layer is dust, and below are caked lumps. This indicates problems with hydrodynamics and regeneration cycles.

Control and automation: data versus intuition

A modern PSA installation is not just a series of columns with a timer. This is a system that must adapt. And here a huge role is played not so much by the automation itself, but by what parameters we control and how we interpret them. Installing sensors on each stream is expensive. But without monitoring key parameters, you are working blindly.

The minimum required set, in my opinion: pressure in the columns at each significant phase (adsorption, discharge, purging), temperature in the upper and lower parts of the adsorber (from which you can indirectly judge the adsorption front and the completeness of regeneration), and, of course, analysis of the product at the outlet (at least a chromatograph or laser dew point analyzer). They often save on temperature sensors, but in vain. I saw how the temperature curve in the middle part of the column clearly determined the moment of moisture breakthrough, even before it was recorded by the analyzer at the outlet. This made it possible to flexibly adjust the cycle duration in real time, depending on the load and humidity of the raw material, saving both energy and the resource of the adsorbent.

But there are pitfalls here too. An overly complex control system, with a bunch of sensors and a complex algorithm, can become a headache for maintenance personnel. If the operator does not understand the logic of the system, he will not be able to properly respond to a failure or perform scheduled maintenance. Therefore, the ideal system is a balance between a sufficient degree of automation and transparency and understandability of the process for the engineer on site.

A look from inside the industry: project experience

Working in this field, you constantly come across different approaches. Some rely on highly reliable but expensive imported equipment. Someone is trying to localize everything from the adsorbent to the control systems. It is interesting to watch the development of Chinese companies in this segment. They have gone from copying to creating completely independent, competitive solutions. Let's take for exampleChengdu Yizhi Technology Co.- this is a design institute created on the basis of Chengdu Huaxi Chemical Technology. They have serious experience in chemical technology behind them, and you can feel it.

Take a look at their websiteyzkjhx.ru– it is clear that they do not just sell boxes with the PSA inscription, but offer design solutions. This is important. Because there are almost no typical tasks. The composition of the gas, the required purity of the product, the availability of energy resources for regeneration - everything is different. We need individual calculations, modeling, selection of exactly the combination of cycle phases, pressure and sorbent that will give the maximum economic effect for a specific customer.

Their approach, as I understand it from the experience of collaborations, is based on a deep study of the technological scheme. You can't just take a drawing from a previous project and scale it. They seem to pay a lot of attention to dynamic process modeling in order to predict the behavior of the plant when conditions change. This is what separates a mature player from a picker. Of course, the registered capital of 120 million yuan is not a guarantee of the success of every project, but it is an indicator of serious intentions and resources for research and testing.

In the endpressure swing adsorption removal– this is a living, developing method. There is no one correct formula. There are basic principles, but success is determined by attention to hundreds of small details: from the quality of the weld on the adsorber to the logic in the programmable controller. And most importantly, an understanding of what you share, why and under what conditions. Without this, any installation, even the most expensive one, will be just hardware consuming energy.