China: How does pressure swing adsorption work? 

This is what is often asked when it comes to gas separation or drying on an industrial scale. Many people immediately imagine complex installations with a lot of automation, but in reality the key principle is cyclicality, the very essencepressure swing adsorption. To put it bluntly, this is not magic, but controlled “inhalation?” and ?exhalation? adsorbent. I myself have observed how newcomers to the workshop confuse injection pressure with desorption pressure, which is why the product’s moisture content then fluctuates. Let's understand it without gloss.

The heart of the process: more than just a tower of pellets

The basic idea is to exploit the ability of certain materials, such as zeolites or activated carbon, to selectively hold molecules on their surface under pressure. But the whole trick is in the “variable”. One adsorber works onadsorption- takes the target component from the raw stream while another is regenerated by releasing pressure, often to vacuum, and purging. This is not static filtering, it is a cycle. In China, in many hydrogen production plants or natural gas treatment plants, this is the basis.

A common mistake is to think that the higher the adsorption pressure, the better. Yes, the capacity increases, but after a certain limit, energy costs for compression eat up all the benefits. We have to look for a balance point. I remember a project for ethylene drying, where the pressure was initially set at 12 bar, and eventually, after trial runs, it dropped to 9. Adsorption was a little slower, but the service life of the molecular sieves increased significantly, and energy consumption was more economical.

The key parameter that I always look at is the shape of the adsorption front in the layer. If it is too blurry or, conversely, sharp, but with a large pressure drop, something is wrong. Either the granules are caked, or the source gas contains impurities that poison the adsorbent. Visually, of course, you won’t see it, but according to online analyzers at the outlet and the pressure drop, everything becomes clear.

Desorption: where the main losses lie

The regeneration phase, the release of pressure - this is the most painful thing. place. Many people think that it is enough to simply release the pressure into the atmosphere, and the adsorbent is ready. In practice, a lot of product (like hydrogen) is lost this way, and the adsorbent itself is poorly purified. Therefore, an effective scheme is a multi-stage pressure release (equilibrium desorption) followed by purging. Often, part of the already purified product is used for purging, which creates a closed cycle.

Here is a real case: at one installation for producing nitrogen from air, they tried to save on purge gas by reducing its consumption. As a result, moisture was not completely removed from the zeolite, and after several cycles the dew point at the outlet crept up. I had to stop and carry out deep thermal regeneration - downtime and losses. This is a classic optimization mistake.pressure swing adsorption.

An interesting point with vacuum desorption (VPSA). This is often the best way to extract oxygen from the air. Reducing the pressure in the column to a high vacuum sharply increases the driving force for desorption. But there are problems here - you need high-quality vacuum pumps and absolute tightness of the system. The slightest leak and the efficiency drops before our eyes. I worked with installations that used rotary vane pumps, so they had to be maintained almost on schedule, more strictly than the main production line.

Materials matter: not all zeolite is created equal

The choice of adsorbent is 50% of success. For drying - zeolite 3A or 4A. To separate the air mixture into nitrogen and oxygen - zeolite 5A or 13X. But this is not a dogma either. For example, modified materials with improved capacity are now being actively tested to capture CO2 from biogas. Chinese adsorbent manufacturers, like those that supply raw materials to many design institutes, have made great progress in this regard.

What do I look for when evaluating material? Not only on the factory passport with data on static capacity. Mechanical abrasion resistance is important. In cyclic mode with constant pressure drops, the granules rub against each other and against the walls of the apparatus. If the strength is low, after six months there will be dust instead of granules, which will clog the tubes and valves. There was a sad experience with one seemingly inexpensive zeolite - after 4000 cycles, dusting was catastrophic.

Another nuance is the shape of the granules. Cylindrical extruded or spherical? Balls generally provide less flow resistance and more uniform distribution, but are also more expensive. For some applications where pressure drop is critical, this price difference is offset by compressor energy savings. These are details that are being worked out at the technological design stage.

Automation and valves: the nerves of the system

The adsorption-desorption cycle lasts minutes, sometimes tens of minutes. Everything is controlled by pneumatic or solenoid valves according to a given program. The reliability of these valves is the key to continuous operation. The most common failure is a sticking or slow operation of the flow switch valve. Because of this, raw and purified gas are mixed, and the quality of the product instantly drops.

Therefore, in serious installations they install not just a timer, but a system that monitors the state of the adsorption front (often using a temperature sensor or composition analysis) and can adjust the cycle duration. This is no longer a basic PID controller, but more complex logic. We implemented this at a hydrogen purification plant for an ammonia plant - we managed to increase the product yield by 3-5% due to a more accurate determination of the moment of breakthrough.

By the way, about the breakthrough. This is the moment when the adsorbent is saturated and the target impurity appears in the purified stream. Ideally, the cycle switching should occur a little earlier. But if the valve operates with a delay of even a couple of seconds, the batch of products may not comply with the specifications. It is necessary to configure the system with a reserve, which reduces the overall efficiency of using the adsorbent volume. The eternal compromise between security and economics.

From practice: from drawing to commissioning

Theory is theory, but practice decides everything. Let's take, for example, a design institute that specializes in such solutions. Let's sayChengdu Yizhi Technology Co.(their website ishttps://www.yzkjhx.ru). This is exactly the structure that grew out of a chemical technology company and is engaged in the design of industrial plants. Their job is not just to sell an adsorber, but to calculate the full technological regime for the specific task of the customer.

In their practice, as I understand from conversations with colleagues, gas separation problems for metallurgy or chemistry are often encountered. This is where all the subtleties come into playpressure swing adsorption. Let's say you need to get oxygen for the converter. Take the VPSA installation. The number of adsorbers is calculated (often 2 or 3, so that one is in reserve or in the filling/dumping stage), 13X zeolite is selected, and the valve circuit and control system are designed.

But the fun begins during commissioning. All calculations are a model. In reality, the composition of the air on site can fluctuate (humidity, CO2 content), and the ambient temperature affects the operation of the compressor and vacuum pump. Therefore, tuners spend weeks finding the optimal parameters: the duration of each stage of the cycle, adsorption pressure, degree of vacuum during desorption, purge gas flow. Sometimes they even change the originally designed valve switching sequence. This is painstaking work, the result of which is a stable 93-95% oxygen output for years.

It is in such institutes that practical experience is accumulated that you will not find in textbooks: what sealing material holds best on flanges under cyclic loads, how to properly organize condensate drainage from the dryer in front of the compressor so as not to flood the adsorber with water, how to interpret small pressure fluctuations on SCADA system graphs. This is knowledge paid for by hours of work at the control panel and analysis of unsuccessful launches.

Instead of a conclusion: thinking out loud

So, getting back to the title question...Pressure swing adsorptionis a living, breathing process. It cannot simply be copied from one plant to another and expect the same result. It is always a balance between the theory of sorption, the practical mechanics of the apparatus, the reliability of the fittings and, ultimately, economics. Sometimes it seems that by adding another pressure relief stage, you will squeeze out an extra percentage of the product, but will complicate the system so much that it will become unprofitable to maintain.

The main thing I have learned over the years of observing these cycles is that you need to feel the system. Don't just look at the numbers, but understand why today the outlet dew point is half a degree higher than yesterday with the same settings. Maybe the atmospheric pressure dropped, maybe the adsorbent began to age, or maybe the sensor just “cryed”. This is no longer pure technology, it is a craft. And in China, with its huge fleet of industrial installations, there are entire armies of such artisans who know how to make adsorbers work stably and efficiently. And companies like the mentioned design organization are precisely those nodes where this knowledge is accumulated and transformed into new working projects.