11 replies to this topic

[Bodhisatya]

Hello everyone,
Am currently working in an Air Separtion Unit. We are facing a strange problem here. We have two Direct Contact coolers in our Plant. The hot compressed air(135 DegC) after Air Compressor enters into the 1st Cooler having a direct heat exchange with Cooling water suppiled via Pump. This cooled air enters into the second cooler for further cooling, exchanging heat with Chilled water coming from a Water Chiller. The 2nd vessel is having a demister Pad to prevent entraiment .

But the Problem is that we are observing physcial carryover of water along with air ex 2nd Cooler. Intially, we suspected that the demister failed, so it was replaced. But still the problem persists. On adjusting water flow rate to both the Vessels,we found that the Carryover has reduced. But to our utter surprise the carry over has increased significantly since yesterday, even with this reduced flowrate.

Some have suggested that the carry over might be due to the Poor quality of water suppiled. We indeed have observed huge foaming in suppiled water.Is this the reason or there is something other, Please suggest.

The Air is then supposed to enter Dryers for Purification. Now water has also entered in these Vessels. I am fearing that it might have damaged the dessicants(Mol Sieve,Alumina).

Please help me out.

Regards
Bodhisatya.

Edited by Art Montemayor, 23 April 2012 - 10:07 AM.

[Bodhisatya]

Respected Delegates...waiting for Some sort of reply....

[pkannan]

Dear friend,

It is problem of velocity of fluids. The water cannot carry over to the overhead if your fluid velocity well below superfacial velocity. Please estimate the same and solve.

Regards,
Kannan

[Dipankarc84]

Bodhisatya,

Along with this data can you please tell us what is the design (rated) water flow and air flow for which the ASU has been designed? As pointed out in the previous post, you might very well be operating at very high vapor velocities (significantly higher than the design flow rates) due to which the demister performance is hampered.

Regards
Dipankar

[Bodhisatya]

Respected Delegates:

Thanks for replying. I fear that the air velocity is not the case. Here, we have a provision for running with 2 Compressors as well as One Compressor. The rated flowrate for single compressor is 5,500 Nm³/hr which cumalitively stands at 10,000 Nm/hr for both Compressors. Both these Values are well within the design rate, which the demister is supposed to withstand. But that's not happening.

The water flow rate to be supplied to Cooler1 is 33 m³/hr as per the design(Currently we are running it at 11 m³/hr) and for vessel 2 it is 5 m³/hr (3 m³/hr currently).

Regards
Bodhisatya

Edited by Art Montemayor, 23 April 2012 - 10:04 AM.

[Dipankarc84]

Bodhisatya,

I suggest that you calculate the K value of your demister based on the actual flow rates and compare them to the design values as well. Demisters function on the philosophy of impingement, coalescence and settling ,i.e, the liquid droplets must impinge on the pad, change direction and coalesce with another droplet. The resulting droplet, owing to it's larger size will settle down, thereby de-entraining the gas.

If the gas velocities are very high, the liquid droplets will be re-entrained and hence you will find liquid carryover at the top of the vessel. On the other hand, if the gas velocity is too low, the droplets might not have sufficient velocity to impinge on the pad and thereby, they simply move through the demister as mist and you notice carryover at the top of the vessel.

In general, if the K-value is less than 0.03 m/s you might experience the latter phenomenon while at K-values higher than 0.12 m/s you might experience re-entrainment.

A lot of times, owing to the low velocities in retrofit columns, demisters need to blanked off from the sides to aid in sufficient gas velocities that will effectively impinge the droplets on the pad. So, it is important to check whether your actual gas rates are "within bounds of the optimum velocity" and not just "within the rated flow rates".

Secondly, I gather that you have inspected the demister and found no damage in it. Is it a Knit mesh kind of demister? It is important to inspect the demister in its "installed condition". Even though you might find the demister to be perfectly fine once you remove it and bring it to the ground, a lot of times there are gaps in between the separate sections of the demister providing enough area for the liquid droplets to escape.

Third, what is the system of draining the entrapped water from the demister? In case, the draining provision is choked you might find that your demister is "wetting and weeping".

Regards
Dipankar

[Art Montemayor]

Bodhisatya:

By the term “Direct Contact coolers”, I understand that you are using what is otherwise called “humidifiers” instead of shell & tube heat exchangers. In other words, instead of convection and conduction heat exchange, you are employing simultaneous heat and mass transfer to cool down the high pressure air compressor discharge prior to its being sent to the molecular sieve adsorption unit for water and CO₂ removal. Am I correct in this interpretation of your description? The reason for my asking for a clarification is because I started my career in the compressed gas industry and spent a lot of years installing, operating, and modifying air separation units and I never saw, read about, or heard of anyone using a direct injection of water to cool down a compressed air stream in an air separation unit. Yours is the first application I have ever hear of using that unit operation. I personally consider that as a very bad process application and would never recommend it to anyone or use it myself. But if that is what you have to work with, so be it.

My reason for clarifying the issue is straight-forward: the reason hardly anyone uses that method is because of the potential for process upsets and the tradeoffs that it represents. You are probably going through one of the major upsets in such a system. Yours is a simultaneous heat and mass transfer operation that depends on subsequently and successfully separating the liquid and resultant saturated vapor phase efficiently. It is as simple as that – but also as troublesome as anything can get. The major and first issue to resolve is: WHO – specifically – DESIGNED AND FABRICATED the vessels and internals involved in the direct contact of both phases? Secondly, who has verified and approved the design?

Additionally, you say that you have observed a decrease and an increase in “carryover” of cooling water from the contact vessels. But how can you “observe” and confirm that? The system is under pressure you only would have a liquid level gage as a monitoring device on any post separator prior to the adsorption unit. And the liquid level gage has no capability of monitoring or measuring any carryover into the adsorption unit. How can you – or anyone else – substantiate that? Please bear in mind that I am not being critical of your installation or your operation. As an experienced engineer in this specific field, I have firsthand experience on this type of operation and how it is designed and operated. I know of its deficiencies and shortcomings in trying to operate it with quantitative results, so I know how to get to the bottom of what is happening.

If, indeed, the direct-contact coolers are operating deficiently then the adsorption unit should fail to dry the compressed, saturated air to the required -100 ^oF dewpoint level and the cryogenic cold box would result in water ice congestion and blockage – causing a total shutdown and defrosting operation immediately. But you haven’t told us that.

Nevertheless, if there is carryover of water then the issue lies in the design or the condition of the internal of coolers --- and here, I am assuming that these coolers operated satisfactorily in the past at the same capacity conditions or higher. Therefore the solution to the problem is rather simple: find the out-of-design condition presently within the coolers. This could be the spraying internals, the packing (if it exists), the demister pads, or the amount of water being accurately measured (if the measurements are accurate) - or even the water quality. But you have not told us if past operations have been satisfactory. What is the unit’s production history look like?

[Bodhisatya]

Dear Art Montemayor...

Nothing to tell u about the plant's production history. it is still under Commissioning stage as because the warm up section couldn't be established because of this water Spillage. Taking air into the Cold Section needs to be ruled out. So Cold Box is safe uptil now.

Sir, without doubting ur engineering skills, the ASU plants that I have seen till date be it Linde's or Air Liquide's. All use these Direct Coolers to cool Compressed Air. Another way out, is using Air Chiller unit to cool air which we generally use in smaller plants, but those freon units have their own inherent problems. Lets not get into that.

Coming to your point of how I observed Water Carry over. At the downstream of Direct Cooler 2, air enters the adsorbers. At the common inlet point, as well as at the respective beds, we have LPDS (Low Point Drains), on opening them before taking air into Mol Sieve beds, I have seen physcial carryover of water with air instead of saturated air coming. And Yes, ur understanding is very correct when u use the term Humidifier. When it comes to design, there are specified vendor and our design department who designed it and selected the internals, so I am not authorised to say anything. But what I can say from an operations point of view is that the performance of packing is appriciable as we got the desired outlet temp.ex direct cooler 2 - even with lower water flow rate. Expecting a solution from ur end.

Regards
Bodhisatya.

[Bodhisatya]

Dipankar Sir....

A lot of times, owing to the low velocities in retrofit columns, demisters need to blanked off from the sides to aid in sufficient gas velocities that will effectively impinge the droplets on the pad. So, it is important to check whether your actual gas rates are "within bounds of the optimum velocity" and not just "within the rated flow rates"...Please Elaborate...couldn't fathom it properly.

Coming to point no 2..yes it's a mesh type demister.I have personally seen the pad both during installation as well as during maintanance inspection..it is perfectly in order.No such deformation or gap seen.

Third Point the droplets coalesecs and settle down at the bottom,we have Level controllers for these Cooilng vessels..which maintains the level properly..by opening the control Valve.So it's malfunctioning is probably not happening as the Valves opens up on level build up..and Vice-Versa...

Regards
Bodhisatya.

[S.AHMAD]

Bodhisatya
1. Firstly, check the velocity as per previous post.
2. Secondly, have you check the water level in the vessel? another possibility is water carry over due to restricted water outlet. During commissioning, normally we do water flushing to remove any debris, cloth and worst still there are sometimes bolts and nuts left out during construction. Similarly if level control valve undersized or stuck partially closed or malfunction.

Edited by S.AHMAD, 23 April 2012 - 07:54 PM.

[Bodhisatya]

Ahmed Sir..

As I Said in my previous post, even when we ran with a single compressor with a rated flow of 5500 Nm3/hr(8.5 bar(g)) we observed water carryover.I have run the machines as per the specification provided which lies well within my boundary Limits.Now when both compressors run With a Combined Water Flow rate (10000 Nm3/hr) matter aggravates.Please suggest what is the line of action that I need to adopt from operational point of view.

Coming to ur Point 2..I have personally mointored the Flushing activities and can assure you that there is no left over in the water lines.Also the Level Controller is working fine, to verify my statement I reduced the water Level to 28%(270 mmWC) still carry over was there..

[Dipankarc84]

Bodhisatya

As explained, at lower velocities the water droplets will not have enough energy to impinge on the demister and change direction. Consequently, it gets carried along with the air. To avoid this, in a lot of retrofit columns (this is an example I used) where you already have a large diameter vessel, the periphery of the demister is blanked off. This effectively reduces the diameter of the demister and results in a higher superficial velocity of the gas. This was just an example I used to say that even "low gas velocities" do hamper demister performance.