Steel Cord Producer Manages Pressure

The facility produces “Steel Cord” and is a division of a large corporation. The following information was produced from a compressed air system analysis done over seven days. The compressed air system is an essential utility required for all aspects of operation. Clean dry compressed air is delivered to two production areas known as HP (half product) and FP (final product). The compressor room is situated between these two production areas. HP receives air through a single 3-inch header and FP receives air through a 1-inch diameter loop and a 1 1/2 inch diameter loop. Piping is copper throughout. Point of use piping is fed from troughs on the floor to each individual machine.

There is a 1065 gallon wet compressed air storage tank, two heated desiccant dryers in parallel and a 1500 gallon dry compressed air storage tank. There are five compressors on the supply side. They are as follows:

• #1 Atlas Copco model GA55VSD (lubricant injected) rated 270 acfm which was online
• #2 and #3 are Atlas Copco ZT45 oil free compressors each rated for 216 acfm, only #2 was online.
• #4 Atlas Copco model ZT50 VSD oil free rated for a max flow of 242 acfm and online.
• #5 is a rental which is an Atlas Copco model GA55 lubricant injected (fixed speed) rated 312 acfm and online.

During the assessment the fixed speed rental GA55 ran fully loaded along with #1 GA55VSD trimming and #4 ZT50VSD trimming. The compressor #3 ZT45 was online but never loaded into the mix. Average flow varied between 500 and 600 scfm. Wet receiver pressure (discharge pressure) averaged 111 psig with a 6 psig loss through filters and dryers. 106 psig was therefore the dry receiver pressure. The HP side fed with a 3 inch header had no pressure gradient from supply to demand. FP side because of the undersized header system lost an additional 8 psig from dry tank to end use.

Measurement Methodology

Measurements creating a baseline of the compressed air system are required to gain a basic understanding of the dynamics occurring in the plant. This compressed air analysis consisted of three days of onsite study and seven days of data collection. The running compressor power was recorded and Supply Side pressures were recorded at the common discharge of the compressor (wet receiver) and after the cleanup equipment (dry receiver). Demand side pressures were recorded in four areas: at a remote receiver on the HP side, at wastewater treatment, at a remote receiver on the FP side and on column F53 in FP.

With this data we were able to create a pressure profile and identify where the pressure drops or draw-downs were taking place. We needed to identify whether or not the main header was at fault (possibly undersized for the flow) or if the reported pressure drops were at the end users local piping or from the diaphragm pumps that run each day. Two thermal mass type flow meters were installed: one at the discharge of the dry receiver for total flow and one in HP for that area only.

The data collection lasted for seven days and sample intervals were at 3 seconds with an average of every 7 samples. This equates to a 21 second interval. A total of 28,801 readings were taken at each test point with the 45 second averaged interval.

AirMaster+ Software was used to analyze the data collected in this system. AirMaster+ is a Windows-based software tool used to analyze industrial compressed air systems. AirMaster+ is intended to enable auditors to model existing and future improved system operation, and evaluate savings from energy efficiency measures with relatively short payback periods. More information on AirMaster+ Software can be found at www.compressedairchallenge.org

Annual Electricity Costs

The data was tabulated within the AirMaster+ software tool. A weighted average of $0.04545/kWh was used. Using the seven days of data, we have extrapolated them to provide a profile of one year of energy costs. These costs do not include maintenance fees, labor or water/sewage use.

Pressure Profile

The pressure profile chart shows the pressure gradients that exist from supply to demand. The occasional pressure drops correspond to the HCL or LL pumps coming online. Later we will see that these pressure drops are due to the compressors not responding to the demand. Pressure in the FP area of the plant is consistently 7-8 psig less than the dry receiver pressure in supply. This is due to the flow traveling in undersized piping. The two loops, comprised of 1 inch and 1 1/2 inch diameter piping, are not capable of the flows required in FP.

HP which is fed with a 3 inch diameter header is quite capable of handling all the flow
requirements during all aspects of production. This would include all the diaphragm pumps that are used since they are all fed off the 3 inch from HP. Looking back at the chart you will notice how the pressure gradient from dry receiver to end of line at the HP receiver and even wastewater treatment are all identical. Pressure draw-downs are felt plant wide and not isolated to any area.

	"The audit of this facility has discovered the source of the pressure fluctuations and recommends examining alternatives to the pumps being used that cause these demand events."

The drawing and plating machines in HP are all lower pressure applications. Pictured right is a typical regulator that is installed on most HP end use equipment. It is scaled in MPa (megapascals) (.4) to (.5) MPa setting equates to 58 –72 psig. Barring a major supply side failure, HP area will never be impacted or have issues relating to pressure. Even the diaphragm pumps can operate at lower pressures.

FP is the area of production where low pressure could cause an occasional disruption. Due to the undersized pipe, the given flow to FP creates a 7-8 psig gradient (from dry receiver to end users). The chart shows an average pressure of 93 psig in FP. When a diaphragm pumps comes online such as the trash pumps, LL (liquid lubricant) or HCL pumps, we see the pressure drawdown very easily. The draw-down lasts as long as the pump is online and then the pressure recovers. A key finding here is that given a different combination of compressors, the pressure drop was not that extreme. This would lead us to believe that the pressure drops can be cured with the right mix of compressors online. Piping issues still are a threat to stabilizing pressure but the supply side is still the bigger impact.

FP’s pressure concern is in the rewinding area. We measured pressure at the furthest machine, which was FGW-10. The regulator at each rewinder is set for .5 MPA which is 68 psig. High-speed data logging proved this. This is at least 20 psig lower than the lowest pressure recorded during the 7 days of logging. Only a compressor failure could cause the pressure in FP to drop and impact this area. Additionally the safety
switch is set at 58 psig or .4 MPA and this pressure will probably never be reached.

Examining the Pressure Draw-Down

The pressure profile chart has a red circle around the first big pressure drawdown. The chart below is a close-up look at this event. It was caused by the HCL pump coming on from 10:00 to 10:50 AM. Here we can see how the compressors reacted to this event. Flow readings on the HP flow meter went from an average of 150 scfm to 300 scfm, an increase of 150. Online were compressors #1 GA55VSD, #2 ZT45, #4 ZT50VSD and the rental GA55. The rental at 85 amps is always fully loaded while #2 ZT45 at 35
amps is unloaded during the entire 7 days.

That leaves the two VSD’s to handle any increase in flow above what the GA55 can supply. The red line is the #1 GA55VSD and you can see how it responded immediately and went full rpm. The problem is the green line which is the #4 ZT50 VSD. It did not offer its full output and therefore you can see the pressure throughout the entire plant was in drawdown. This means the demand flow exceeded the supply flow. When the pump shut down, the pressure returns to its average values. Looking back at the demand side piping, we see that the tank farm and wastewater treatment are off the 3 inch line. Yet the entire plant experiences the drawdown of pressure due to the inability of the supply side to react to the new load.

Demand Side Recommendations

Demand Side - Air Leaks

A leak survey was not performed, however, we were told that during maintenance with
no production running, one ZT45 rated at 215 acfm can just about hold header pressure to 100 psig. We would consider this the gross leakage rate. This equates to about 35% leakage.

Leakage could be occurring under the troughs that feed all production machines. Pictured here (in yellow circle) is poly flow tubing that comes up to feed each machine. Push-to-connect or ferrules could be leaking. Since they are inaccessible, the leaks never get fixed. If the company eliminates only 50 scfm of air leaks, that amounts to an additional $4,700 per year savings.

Demand Side - Diaphragm Pumps

The flow-increase from our charting shows about 150 scfm increase when a diaphragm pump is running. With the minimal piping diameters and less than desirable response from the compressors, we need to either change the pumps to electrical type or repair all piping/volume and compressor control response. To reduce energy and optimize the system we need to use less compressed air. To do so means finding other methods besides compressed air to perform the functions going on now. One way is to replace the culprit diaphragm pumps with an electrical equivalent.

Centrifugal pumps are commonly used in place of diaphragm pumps when the viscosity of the fluid is not an issue. Where viscosity is an issue, there are electrically driven “PERISTALIC HOSE PUMPS”. They are capable of the same pumping characteristics as the pneumatically driven diaphragm pumps. Although this report does not endorse any particular product, I recommend that the firm investigate this type of pump along with others that can replace the air intensive diaphragm pumps presently in use.

Distribution Header Pressure Drop

The existing 1 inch and 1 1/2 inch diameter piping making up the FP loops are undersized for the flows that are required to support production. Now remember that leakage is a real flow and is taking up some valuable real estate in the pipes. If leakage was reduced by 50%, there is a good chance that the flows would be OK in the existing piping scheme. But for now let’s look at what size piping can handle the flows required to FP. If we use 600 scfm minus the 150 going to HP or 450 scfm as a maximum
peak flow we can calculate what size pipe is needed. The more flow you try to put through a pipe the greater the pressure drop will be. Pressure drop in a pipe increases with the square of the increase in flow. Which means if you double the flow, the pressure drop will increase four times what is was!

Supply Side Recommendations

The compressed air supply, utilizing sufficient storage, and proper distribution, must meet the compressed air demand. If supply, storage and distribution are not in tune or aligned, excessive pressure fluctuations will occur resulting in increased operating costs. Most compressors ability to load or unload is controlled by line pressure. Typically a drop in pressure indicates an increase in demand. This then causes a compressor to come on line or load and thus handling the increase in flow. In this system, this takes place on a regular basis as pressure rises and falls, however the compressors cannot see the true production floor pressure and are only reacting to the pressure in the wet receiver. The wet receiver as we have seen from our data could be 10-15 psig higher than actual point of use pressures.

There is no automation in place to orchestrate the compressors starting or stopping. Because it is very difficult to successfully cascade more than two compressors, there are times when the system is running with too much horsepower, all sharing the load and they simply cannot react to demand events in time. With no automation compressors maintained a higher than normal power usage regardless of production requirements. This occurs because as the pressure increases, so does the demand for air in all unregulated uses such as leaks, open blowing, and users with the regulators cranked all the way open. This phenomenon is called “artificial demand” and it prevents the compressors from being able to equalize the pressure throughout the header.

New Compressor

If we were to install a 100 hp air compressor, such as a GA75 lubricant injected fixed speed, we can accomplish what the rental plus the ZT45 is currently doing but with less power. A single 100 hp would require about 82 kW fully loaded and output 460-480 cfm. Now we only need to trim with the GA55VSD. This puts all lubricant injected rotaries online which have a greater cfm per bhp than oil free. This equates to a $14,184 savings over the existing baseline.

Electric timed drains

Electric timed condensate drains are used on each compressor and on each receiver tank. A 1/4 inch timed drain operating for 10 seconds open every 30 minutes (at $0.05kWh) can cost $700 in compressed air loss per year. The system has six operating which potentially is costing over $4,000 in compressed air. Ask you vendor for zero air-loss type drains and replace the existing drains.

Conclusion

The audit of this facility has discovered the source of the pressure fluctuations and recommends examining alternative to the pumps being used which cause these demand events. We also recommend the installation of a new 3 inch header down through FP starting at the dry receiver. This header should be tied into the existing loops.

Purchase a fixed speed compressor rated at 100 hp which will allow a reduced energy profile to operate under the current and future demand environment. Finally, perform a thorough leak check especially on the under floor trough piping and replace timed condensate drains with zero air loss type condensate drains.

For more information please contact Donald Wirth, Tencarva Machinery Co., tel: 615-742-3101, email: dwirth@tencarva.com, www.tencarva.com