Eliminating air locking in centrifugal pumps through cold material recirculation

Thread Content

This post was last edited by The past is frozen on 2021-2-3 at 16:59 ------Eliminating air locking in centrifugal pumps through the return of cold material. Author: Wang Shibing, February 2, 2021. Today I would like to share with you a small case from the chemical industry related to solving the problem of air locking in centrifugal pumps, so as to enable us to learn and exchange ideas together. A schematic diagram is provided at the end for easier understanding. To make it easier for more people to understand, the text includes many explanations of terms and industry basics, which makes it rather verbose; please bear with us. What is the gas entrapment phenomenon? Those working in the chemical industry have surely encountered the problem of air locking in centrifugal pumps. Air locking, commonly known as \"the pump not being able to deliver fluid\", occurs when there is air present in the pump chamber at the time of startup. Since the density of air is much lower than that of the liquid being transported, the centrifugal force generated by rotation is weak; as a result, the low pressure created in the center of the impeller is not sufficient to draw liquid into the pump. Even though the centrifugal pump is started, it is unable to transport fluid – this phenomenon is what is referred to as air locking in centrifugal pumps. Put simply, it means that there is gas inside the pump chamber, or the fluid vaporizes within that chamber; the gas prevents the pump from performing its function of transporting fluid. The solution to this problem of gas entrapment is as follows: As we all know, eliminating this issue is quite simple – one just needs to exhaust the air from the chamber before starting the pump. But how can the gas entrapment phenomenon caused by fluid vaporization be resolved? Unfortunately, we ran into this problem; due to technical constraints, it is not possible to increase the back pressure of the pump (the container in question is not a pressure vessel, and there are limits on the operating pressure). Introduction to our equipment: Our production facility includes an alkali washing tower used to wash acidic oil and gas with alkaline solution. Inside this tower, cold alkaline solution (at around 30°C) flows down from the top of the tower, where it comes into contact with the hot oil and gas flowing upward. An acid-base neutralization reaction takes place between them, along with heat exchange. During this process, part of the oil and gas condenses and flows together with the alkaline solution to the bottom of the tower. By the time it reaches the bottom, the temperature of the alkaline solution has risen to about 70°C. It is then pumped to a heat exchanger using circulating water to be cooled down. After that, it enters a oil-water separation tank where the oil and water are separated. The alkaline solution at the bottom is sent back to the top of the alkali washing tower via a reflux pump for reuse ； The oil phase in the upper layer of the separation tank overflows from the top for recovery. The unit is equipped with one reboiler pump, one reflux pump, and one spare pump shared by both the reboiler and the reflux. The flow rate of the recirculation pump is called the recirculation volume, which is what I refer to in this text as the circulation volume of the alkaline solution; a flow meter can be used to monitor this value. The flow rate of the bottom of the tower pump is equal to the sum of the alkaline solution circulation volume and the content of solvent oil, and no specific figures are available for this. The problem we encountered: Due to the high temperature of the material taken out from the bottom of the caustic wash tower, as well as its presence of volatile solvent oils (mainly n-hexane, with a boiling point of around 66°C), the pump at the bottom of the tower is prone to gas locking. We have three production lines. When we were using two lines, the solvent oil content in the alkaline solution at the bottom of the tower was relatively low, so the pump at the bottom of the tower didn’t fail to deliver fluid very often; such failures were mainly caused by clogs in the pump’s inlet filter ； Over the past two years, all three production lines have been operating at full capacity. The concentration of solvent oil in the alkaline solution in the tower bottom has increased by 50% compared to before. The pump at the tower bottom often fails to deliver the required flow rate. To maintain production, we turned on two such pumps and reduced the valve settings at their outlets in order to prevent the pumps from not delivering enough fluid. There are two disadvantages to this approach: ① Running two pumps simultaneously increases electricity consumption, and ② the second pump serves as a backup for both the reactor bottom and the tower top reflux, and is not suitable for being in use for long periods of time. 　　Even when two tower bottom pumps are started, it still happens occasionally that the pumps cannot deliver sufficient flow. Methods such as cleaning the pre-pump filter, reducing the circulation volume of the alkaline solution, and cooling the pump casing by spraying water have all been tried, with little effect. This issue has troubled us for a long time; at its worst, we reduced the circulation rate of the alkaline solution to 4 t/h, and we closed the outlet valves of both column bottom pumps to a very small degree in order to barely maintain circulation (the normal circulation rate is 17 t/h). Our first experiment: While dealing with this issue, we found that when both bottom tank pumps were operating simultaneously, gas locking occurred, and in such cases both pumps stopped pumping at the same time. Since the inlet pipelines of the two bottom tank pumps shared the same outlet, we suspected that there were solid objects inside the bottom tank, blocking the outlet. In response to this speculation, we connected a water pipe at the inlet drain of the tower bottom pump and used water injection to backwash the discharge pipeline and outlet of the tower bottom; this approach worked very well at the time, as one tower bottom pump was sufficient to meet the requirements. However, this solution did not last long – it only worked for about five or six hours. After going through this process several times, it actually returned to normal. After this experience, we are firmly convinced that there are solid objects in the bottom of the tower that are blocking the discharge outlet; it will be sufficient to clean them out during the shutdown for maintenance. (Such judgments misled us for a long time.) With concerns in mind, towards the end of the year I began to worry about the judgments I had made earlier, as our device had been in operation for eight or nine months already. Plans for maintenance for the new year had been put on the agenda; if, during that maintenance, it was found that there was nothing inside the tower reactor and that the outlet wasn’t blocked, how would we address the problem of the pump not being able to deliver sufficient flow? (People who have no experience in the chemical industry may not understand our concerns. The chemical industry is a highly dangerous sector, and most renovation projects can only be carried out when the equipment is shut down. However, starting up or shutting down such equipment is extremely costly; as a result, many of our renovation plans can only be carried out during downtime for maintenance, and if we miss that opportunity, we have to wait another year.) For the second experiment, after considering these concerns, we re-evaluated the issue with the bottom-of-tank pump. The rated flow rate of this pump is 34 m/h. A few years ago, when the circulation rate of the alkaline solution was 17 t/h, one pump was sufficient to meet the production needs. Now that two bottom-of-tank pumps are in use, during the periods when gas entrapment was most severe, the circulation rate of the alkaline solution had to be reduced to 4 t/h. Why is that? There is no problem with the pump; if there is also no issue with the discharge pipeline at the bottom of the tower, then it’s a problem with the material itself. The issue with the material is also clear; as explained earlier, the material taken out from the bottom of the tower contains volatile solvent oils, which can cause the pumps to fail to deliver the required volume of fluid” ； When the circulation rate of the alkaline solution remains constant, the higher the production load, the greater the content of solvent oil. Is it simply because there are too many volatile components, which restricts the proper functioning of the pump? With such a hypothesis, it is natural to think of countermeasures, and we selected the simplest and most feasible one for testing: we connected a temporary pipeline to draw a small amount of cooled alkali solution (about 3 t/h) from the backflow line of the alkali washing tower, and sent it to the discharge line at the bottom of the tower, where it was mixed with the liquid extracted from there. This was done in order to lower the temperature and the solvent oil content of the material extracted from the tower, with the hope that this would prevent the \"air locking\" phenomenon in the pump. The experimental results were very satisfactory: only one bottom pump was needed to meet the production requirements. Moreover, even with the pump outlet valve fully open, the pump pressure remained stable, rising from 0.25 Mpa to 0.4 Mpa. The circulation rate of the alkaline solution increased to 17 t/h, and no gas entrapment occurred during the week of observation. We made a new judgment and developed a plan for improvement. After obtaining the results of the second experiment, we essentially rejected our previous judgment. The reason is simple: if the outlet had indeed been blocked, there was a high probability that it would get blocked again during the seven-day observation period. Since this did not happen, it’s likely that our previous judgment was incorrect. Based on the new experimental data, we plan to add a DN25 stainless steel pipeline about 2 meters long during the next shutdown for maintenance, along with a gate valve to control the flow rate. Through this pipeline, a small amount of cooled alkali solution (about 3 t/h) is drawn from the alkali washing tower reflux pipeline, and driven by the pressure difference provided by the reflux pump to the discharge pipeline at the bottom of the alkali washing tower, where it mixes with the liquid taken out from the bottom of the tower. Analysis of expected effects: After this pipeline is put into use, 1) the cooling and dilution effects resulting from the return of cold material can compensate for the shortcomings in the selection of the bottom pump of the tower, eliminating the \"air locking\" phenomenon in the pump and allowing it to operate at its normal capacity. 2. The standby reactor bottom pump can be stopped, putting it in a true standby mode for use when needed. 3. Shutting down the backup reactor bottom pump can save at least 15,000 kWh of electricity per year: this pump has a power rating of 5.5 KW and an efficiency of 62%. Assuming that it does not perform any useful work, with 7,200 hours of operation per year (300 days), the amount of electricity saved is: 5.5 KW * 38% * 7,200 hours = 15,048 Kw*h. (The pump that cannot be modified is a small one; the amount of electricity saved in this case is indeed minimal. It’s a pity to apply such an excellent modification idea to this small pump – it’s like using a sledgehammer to crack a nut.) Let’s leave this as a mystery; perhaps some people will ask: Since the problem of air entrapment in the tower bottom pump can be solved by using cold alkaline liquid for reflux, why is it necessary to add an additional pipeline? Couldn’t simply increasing the circulation rate of the alkaline liquid at the top of the tower suffice? The answer is as follows: I know the answer, but I won’t reveal it for now to keep it as a mystery; if you’re interested, we can discuss it privately. Conclusion: In this case, we managed to solve the significant challenges that arose in production using just a simple pipeline. Although it may seem like a coincidence, there is actually a logical explanation for it. I believe that this achievement is tied to our years of hard work, self-study, and diligent research. Therefore, I would advise everyone that having more skills is never a bad thing; learning is an endless process, and one must never be satisfied with the status quo and stop progressing. Author: Wang Shibing, 02.02.2021. The text below can be ignored. I dedicate this work to my fifteen years of youth, as well as my dedication and perseverance in this industry. Over the years, driven by personal passion, I have persisted in this industry for fifteen years. From a 19-year-old youth to a disheveled man with nothing accomplished today, I have spent countless years alone, far away from home. Changes in age and appearance are unimportant; what matters most is the sense of guilt I feel toward my parents and spouse, a guilt that I can never fully reconcile with. That’s why it took me three days to write this article, as a way to prove to them that I am not completely without achievements. The article is written in simple language so that my spouse can understand it. Actually, she is a girl with a very strong understanding ability; it’s always been the case that she already understands, yet I’m still worried that she might not. If one day I leave this industry for my family, I believe that with this article here, it will be sufficient to prove that I was once here.