Circumvent A Compressor Calamity

This Month’s Puzzler

Simultaneous failures of our ammonia compressors shut down our refrigeration unit. I’ve been asked to specify a couple of rentals so we can cope until we get spare parts from Germany. However, I face a difficulty because the file only contains the original design; the ammonia refrigeration system was expanded about 30%. (In addition, our load center is maxed-out, so current draw is a problem.)

A corporate engineer compounded the problem by attempting a material and energy balance based on the condenser data sheet. This undated exercise seems to confuse the total ammonia flow through the screw compressors with the individual flow through the condenser. It shows a mass flow rate much higher than compressor horsepower would allow. Also, the flows through the condenser and evaporators don’t add up. Fortunately, I found a useful website for a pressure/enthalpy chart.

We need two compressors to operate the plant (with an additional unit on standby) and keep two spiral heat exchangers in service at all times (see figure).

These compressors have run so smoothly that their simultaneous failure caught us by surprise. It took two days of frantic searching to find suitable oil filters.

Operations is wondering if we can continue to run the plant without refrigeration because the daytime high temperature now is about 30°F.

How do I resolve this puzzle? Can we safely operate the plant temporarily without ammonia refrigeration? Should we believe the ammonia flow calculation based on the condenser?

Look Into Oil Slugging

Here goes. (Refer to this figure.)

1. Given the introductory statement that the ammonia compressors simultaneously failed, could it be that a slug of liquid (either ammonia or oil) damaged the compressors? Compressors do not tolerate excessive amounts of relatively incompressible fluids. Since spare parts needed to be ordered, something mechanical failed in both compressors at the same time. The cheapest control for expansion feed (liquid ammonia) to the spiral evaporators would be to use a temperature probe (liquid bulb operating a mechanical operator) controlling an expansion valve feed for each exchanger. The control bulb is, of course, in the vapor line at the outlet of each evaporator. Vendors use these controls because they are cheap and help keep bids low. If the valve failed open, excess liquid could “slug” into the compressors and damage them.

I went through something of this nature with a propane refrigeration system design and specified a “superheat” meter on the outlet of the exchanger. This simple device consisted of a temperature transmitter and pressure transmitter that fed data to a distributed control system which used a manipulated Antoine equation along with the pressure to compute the superheat. An alarm and shutdown were provided to protect the screw compressor. Measurement of superheat was a cheap way to ensure the absence of liquid propane being returned to the compressor screw. However, such a superheat meter would not protect from oil slugs in this system.

For small amounts of liquid ammonia returned to the compressors, I would expect some disruption for lubrication but not simultaneous failure of two compressors.

2. Related to this comment, lube oil could be the liquid slugs that damaged the compressors. There is an oil pot shown under V-700 in the system diagram. A separate oil-rich phase must be removed from the oil pot and returned to the compressor area of the system. No detail is provided but this should be investigated.

3. V-700 appears to be a subcooler for the liquid ammonia. No vapor return to the compressor system is shown. Could there be a liquid return system problem in a vapor line from V-700 that could explain the compressor damage?

4. No function is shown for V-703. Is this an accumulator? What happens to any oil at V-703?

5. Is there a knockout pot ahead of the compressors to prevent liquid from slugging into the compressors? If there is such a pot, how is the liquid returned to the system? Is there level instrumentation with alarm and shutdown?

6. Some amount of oil is dissolved in the liquid ammonia. Is there a way for oil to accumulate in the vapor return to the compressors and then slug into them to cause mechanical damage? Is that oil intended to be returned as a mist in quantities that do not disrupt the compressors? One might assume from the Puzzler system sketch that the lines shown are short and compact but is this the case? More detail is needed.

7. The design curve for the cooling water profile and ammonia profile shows ammonia heat removal as linear, i.e., no phase condensation heat load is shown at constant temperature in the design curve. This is a minor detail and may not relate to the compressor problem.

8. Another minor detail is the cooling water connection as shown on the system sketch. The temperatures in the design curve cross (outlet water warmer than the liquid ammonia at the outlet), so it would be expected that the cooling water feed would be “adjacent” to the liquid draw. There is probably no failure explanation to be found here because the system mostly operated with great reliability.

9. As for the Puzzler question about continuing plant operation due to cold weather, there is no information as to the source of heat the refrigeration system is removing. My guess is the plant might have to be shut down but with no information about the heat source and refrigeration use, I can’t address that question.

So, the bottom-line suggestion is to look hard for ways liquid (oil or ammonia) could be slugged into the compressors. If the problem in fact stems from liquid flows into the compressor suction, some changes to correct this will be required in addition to mechanical repair of the compressors. Should liquid slugging into the compressors be determined to have caused the compressor failure, a system redesign might be considered to eliminate this problem — but it won’t be cheap. There could be other explanations and an open mind should be maintained until a convincing problem is found and corrected before the repaired compressors are restarted.
Scott DeYoung, consultant
Ridgefield, Conn.

Estimate Essential Data

At the strategic level, key issues are heat transfer area requirements in absence of refrigeration, power requirements (e.g., for air-cooled heat exchanger) relative to space requirements for heat exchanger, control system for air cooler, and heat and material balance of the existing refrigeration system.

1. Ammonia entering the vaporizer is at 38°F. (Per its Mollier chart, ammonia entering the vaporizer (V-700) at 56 psig and 38°F is a 10°F superheated vapor.) Air temperature at 30°F seems a reasonable replacement of cooling medium as far as the driving force for cooling is concerned. Beyond that, you will need to investigate the overall heat transfer coefficient of the cooling equipment (e.g., air-cooled heat exchanger) in comparison with the heat transfer coefficient of the existing vaporizer.

For estimating ammonia flow rate, you may consider one of two approaches:

a. For 10°F temperature rise of cooling water across the condenser (H-700), you can estimate heat absorbed if water flow rate is known. This heat is given off by condensation of ammonia vapor. Using the enthalpy data from the ammonia Mollier, estimate flow rate of ammonia. (To approximate flow rate, if no flow meters are on the water line, you may consider clamp-on flow meters.)

b. Alternatively, if horsepower for each compressor is known, you can use horsepower along with efficiency to calculate energy input. This, in turn, is the enthalpy change (from Mollier) of ammonia vapor as it is compressed by a compressor.

Note that it is unlikely the above two methods will give you the exact same estimate of flow rate. Physical property data as well as accuracy and precision of instruments (thermocouples, flow meters, pressure transmitters) account for differences in estimates. Use reasonable conservatism for estimates.

2. After getting an estimate of air-cooler heat transfer area, contact vendor(s) to get an idea about space and installation requirements. You also need to check delivery times and see if vendors have rental units.

3. Although information about your cooling application (i.e., the fluid being cooled) is not shown in the problem statement, it would be appropriate to provide the proper control scheme to achieve stable outlet temperature of your stream. Control schemes may include flow bypass or louvers or a variable speed motor or a combination of these or other relevant arrangements.

4. Compare power requirements (e.g., for air cooler) with available power in your motor control center. Because you will not be using refrigeration compressors, you may consider using their breakers.

5. When you revert to the refrigeration system, first try to determine the cause(s) of abrupt failure of the compressors. Their simultaneous failure suggests a common-cause origin — such as liquid entrainment in the suction or something else.

6. As you make changes in operations and equipment setup, keep operations, construction and maintenance folks in the loop and, before startup, put these changes through a management of change process.
GC Shah, consultant
Houston

Pray For Cold Weather

The fact that you can’t find a spare part indicates two things: 1) you’ve got to improve your inventory system; and 2) it’s harder to find things you don’t think you’ll need precisely because something has changed. Take a hard look at why you suddenly need an item not required for routine maintenance.

Actually, you might get away with running without the compressors if the air temperature is low. Assuming you’re running your ammonia evaporators at about 30–40°F, the extra-cold cooling water in the condenser might compensate, at least partly, for the low ammonia flow rate. The cooling tower water will probably be in the 40–50°F range if the ambient temperature is below freezing — even better if it’s colder outside.

The trouble is that you’re at the mercy of weather. A “heat wave” above 40°F will shut you down.

The safest thing to do is to have a rental contract with a compressor company for a skid-mounted unit. Unfortunately, because this system ran so smoothly, you may find there aren’t any tie-points to connect the rental. All of this discussion is in the future, although adding tie-points might be a good idea. Right now, you need to check corporate’s calculations.

First, take the field data and develop the temperature profile first. Then, match it against a pressure/enthalpy (P/H) curve for ammonia. You will need to establish the ammonia temperatures at the evaporator and the condenser; this also will help you verify if you can operate the refrigeration system using cold ambient temperatures. Once you’ve got the two horizontal lines established, do the balance across the system to estimate the flow of ammonia; this will tell you whether the corporate calculations are for one or two compressors.

Never rely on the design load of a heat exchanger to estimate a product flow through it. Remember that exchangers generally are over-sized; often exchangers are bought used and re-tasked to a process. It’s like walking into a shoe store, asking for a size 9½ and getting a size 12 — sure you can wear the shoes but they’re not a perfect fit.
Dirk Willard, consultant
Wooster, Ohio

June’s Puzzler

We installed a new automatic sampler in our catalyst bed reactor (see Figure 1). It replaced an old corroded manual station that was fouling. The manual station worked but some of our new operators warned it posed a burn risk; veteran operators had developed a knack for avoiding that risk. So, the safety department mandated putting in an automatic sampler.