Industrial Utility Efficiency

Reviewing Compressed Air Demand at a Food Processor


This article reviews portions of an audit report of a compressed air system in a food industry factory located in the U.S. Although the audit explored different supply-side options the client should consider to improve dynamic efficiency, we will focus on the demand side of the system for this article.

 

The Compressed Air System

This plant is currently served by four rotary screw air compressors providing compressed air at 105 psig average plant pressure. The four air compressors are at the end of their expected life cycle although they are still operating well. They send compressed air, at one point of entry, to a 1750-gallon “wet tank”. There is a single refrigerant dryer with a cold coalescing filter and an after-filter providing compressed air treatment. The compressed air then goes into a 5000 gallon “dry tank”.

The system normally operates with two compressors running in MASTER mode through the sequencer. One is fully loaded and the other loads and unloads to maintain the target pressure in the dry storage tank. The demand side is regulated down at 80 psig through the flow control valve.

McAuley graph 1

Measurement

During the course of the audit, compressed air flow was directly measured after the pressure regulation valve set at 80 psig as it exited the compressor room.

We also monitored compressor motor amperage on all operating compressors and common header pressure throughout the plant. This allows us to measure the performance of the system in terms of Dynamic Efficiency (DE). Dynamic efficiency is defined as Scfm per Kw. Dynamic Efficiency is a key indicator of system performance and is a readily and easily repeatable measurement. We use this audit measurement as a key benchmark to guarantee and measure improvement in system performance.

We measured the compressed air flow out to the users in twelve second intervals and continuously measured the compressor power in 12 second intervals for one week. We then compiled a frequency analysis of total system flow in 50 Scfm increments.

Below is a histogram representing the actually measured demand side flow data.

McAuley graph 2

We directly measured this flow from Monday morning until Thursday afternoon. From Thursday afternoon until Monday morning we collected amp data only on the operating compressors. This data indicated the flow was slightly less during the remainder of the week. We recorded -4% power on the lagging compressor ( loading and unloading) and +0.6% power on the leading compressor ( fully loaded). This indicates that the weekly demand profile shifted slightly to the left from the data directly observed above. The net result is that the annual cost data extrapolated maybe slightly inflated. For our purposes here I consider this to be irrelevant. We will use the direct flow profile we measure for four days to model the annual system operating cost.

In order to model the annual power required by the compressed air system we measured both the flow out to the plant and the average amperage on each compressor over the course of the data collection period. We then extrapolate this data over a year. The spreadsheet in the appendix details the actual demand BASELINE annual cost of electricity for the compressed air system based upon an entire year of production (8760 hours) and electricity cost of 8.7 cents per Kwh.

The audit result showed a poor compressed air system Dynamic Efficiency (DE) measurement of 4.36 scfm/kW and an extrapolated annual energy cost of \$162,000.

 

Allocated Unit Costs for Compressed Air

The allocated unit cost for compressed air at this plant is \$1.99 per hour of operation per 100 Scfm of usage. This number is derived from the total cost of compressed air operations, the average demand side flow rate observed in the system (930 Scfm), and the total hours per year (8760). It is a useful number when comparing costs.

The allocation of unit cost will allow plant management to estimate the financial impact of various management decisions including the operating cost associated with new equipment and the efficacy of better (more) intense maintenance practices. Caution: It is not the marginal cost of the next 100 Scfm of use nor the savings derived from the last 100 Scfm of use eliminated. This number is average for a facility of this type and size in the United States.

 

Cooler Performance Evaluation

During the audit period, air compressor cooler performance was measured and all appeared to be functioning well and not fouled.

Water-in temperature of 105F indicates the cooling tower is not functioning properly. We estimate about 45 tons of cooling is required for compressed air and about 17 tons for the new improved vacuum system.

Excessive temperature is the enemy of compressors. These types of maintenance failures can be identified early and often remedied inexpensively before an emergency breakdown. Discharge temperature should be logged every shift and provisions made to correct elevated temperature situations as soon as they arise.

During the course of the evaluation we observed the water-cooled 2000 Scfm PYRAMID dryer stop working at least twice. We never determined why it stopped working or how it began working again. It is thought that high cooling water temperature may cause these shutdowns.

 

McAuley graph 3

 

Vacuum System Operations

We were asked to collect data on the vacuum systems with an eye toward improvement. We measured common vacuum header and found the system not working nearly where it should be.

We observed the vacuum header in the 7-10 inch mercury range. Normally these systems operate at about 18-20 inches. We observed three 25 hp vacuum pumps. Two operated and one repeatedly overheated and shut down. We believe the plant needs additional vacuum capacity and as much as 100 hp more to adequately meet the needs of the plant. This includes migrating ten splicers from compressed air generated vacuum to this central vacuum system. This estimate will be confirmed through vacuum system experts and relayed to the plant.

 

“Wet Tank” Piping

The wet tank has the air going in and out near the bottom on the tank. The proper way to pipe a wet tank for moisture removal is in the bottom and out the top. This will inevitable cause moisture to drop out of the air stream and will reduce the load on the dryer and all downstream drainage equipment. The audit recommends the plant pre-plumb the required section of pipe and during the next shutdown reconfigure the wet tank for its intended purpose.

 

Potentially Inappropriate Uses of Compressed Air

Because compressed air is ubiquitous in most industrial settings, and since it is a highly adaptable energy source, it tends to be the first choice for power at remote locations even when it is clearly not the most efficient choice. We surveyed the entire plant looking for potentially inappropriate uses of compressed air.

It is worth repeating here this indisputable fact; 85% of the input purchased energy used to drive an air compressor is immediately dissipated in the heat of compression and lost forever. Only 15% of the purchased energy is actually imparted to the fluid ( the compressed air) to later go out and do useful work in the plant. If it can be done with electricity it is automatically 8 times more efficient than with compressed air.

We surveyed the entire plant with an ultrasonic detector and identified only one open blowing application that could be modified.

  • On the Rennco baggers a quarter inch open blow is used to close the open end of the bag prior to tie wrapping the end. This has been modified such that it is intermittent. I observed approximately 25% on time for this application. Assuming 6000 hours per year for each and six in operation at 20 Scfm full flow and our \$1.99 per hour per hundred Allocated Cost model, this use is costing the plant about \$3600 per year.

(20/1000) x .25 x1.99 x 6000 hours x 6 baggers = \$3582/year

   Even though a blower would use only 30% of this energy it would have to stay on 100% of the time, thus there are no net savings using a blower here.

  • The 15 bag houses on the roof are all well controlled via differential pressure and not on time. This has dramatically reduced the compressed air cost here from over 350 Scfm estimated in 1999 to less than 30 Scfm now. I observed only a few firings while surveying the equipment on the roof over a ten minute period.
  • We also looked at the CO2 mold cleaning operations. We wanted to be sure they were not dragging down the local header pressure at the mold cleaning operations. They were not, at least not at the normal mold cleaning station. We observed that the mold cleaning operation used about 200 Scfm for less than ten minutes. This is about 33 cents worth of compressed air for each cleaning operation.

200/100 x .165 hours x 1.99 \$/hr/100 Scfm = 33 cents

   The compressed air use appears efficient and effective in this application. Here is a flow tracing of a typical CO2 mold cleaning operation.

McAuley graph 3

  • Ten Thermoformers (TF) are equipped with Airvac ( Milford Ct) compressed air vacuum generators. These use 0.41 Scfm each. There are twelve on each TF. We estimate they use 65 Scfm or 15 Kw of power (65/4.36). Vacuum pumps are estimated to do this at 30% of the power of compressed air or for a net savings of \$7545 per year.

(65/100) x \$1.99 x .7 x 6000 hours per year = \$7545

 

Effective Leak Management Processes

Leaks were estimated at about 25% of flow. This is typical for a existing plant with lots of motion and heat. We noted many normal leak modes including push pull connectors, connections subjected to excessive heat and vibration (behind the safety fence) and many final connections to applications including unions, pipe clamps, and FRL.

We recommend the plant invest in an ultrasonic United Electric Ultraprobe 3000 hand held detector and recruit a Plant Champion to identify and fix leaks on a regular basis . This monthly shift task, with the Ultrasonic detector, can ultimately achieve reduction in leak loss level to 15%. In this plant, this could be responsible for up to \$16,000 in yearly savings using our allocated cost model.

930 Scfm x ( .25-.15) x 1.99 \$/100Scfm x 8760 hr/yr = \$16,212

 

We’d like to highlight, at Extruder 504, a very large long-term leak at the MAC solenoid manifold.

 

Summary

The compressed air system at this food processing plant is not now operating near peak performance. In fact, it runs at a combined dynamic efficiency (DE) of about 4.36 Scfm/Kw. We believe we can boost this dynamic efficiency to 5.69 by modifying the supply side situation and addressing the demand side issues outlined in this article.

 

For more information contact James G. McAuley P.E., tel: 832-563-6395, www.JimMcAuley.com.

To read more System Assessment articles, visit www.airbestpractices.com/system-assessments.