Executive Summary

 

The goal of this laboratory is to scientifically quantify the operating performance of air compressors and other rotating machines (pumps, turbines, motors, etc.) when utilizing synthetic lubricants. This first experiment is performed on bench-scale a ir compressors charged with commercially available lubricants. Operating conditions will be monitored using a PC-data acquisition system. Data pertaining to the operating conditions and energy consumption will be compared for each lubricant.

 

This report contains a description of the preliminary set up of the energy conversion efficiency laboratory, a brief description of the air compressors, a description of the instruments being used for data collection, the procedure for running the comp ressors and obtaining data, and a discussion of preliminary data.

Preliminary data analysis looks favorable for the advanced synthetic oils, but many more hours of data need to be collected before any results can be stated.

Introduction

As long as industry relies on fossil-fueled power plants to operate industrial equipment, the need to reduce the power consumption of this equipment becomes paramount as pollution continues to accumulate in our shrinking ecosystems. O ne method of reducing power consumption is to reduce the friction losses present in all moving machinery by the use of advanced synthetic lubricants. These lubricants have the ability to be absorbed into the surface of the moving parts providing smoother surfaces and hence, reducing the friction losses. Reducing the friction between moving parts will reduce the power consumption and extend the life of the machine. The drawback of advanced synthetic lubricants is they are substantially more expensive th an traditional petroleum-based lubricants. Unfortunately, when industrial decision-makers are given the choice of investing in the future and helping our environment or making more money, they tend to make the unwise choice of making more money. The goa l of concerned young engineers is to prove that investing in the future and helping the environment by reducing energy consumption can be profitable.

The goal of this research is to determine if energy conservation can be obtained through the use of advanced synthetic oils and to what extent these oils can have on energy cost savings. It is proposed to determine the real energy consumption o f bench-scale air compressors utilizing commercially available lubricants. Operating conditions and cost effectiveness will be determined for each lubricant. This research should provide guidance to Wisconsin industries attempting to determine whether o r not to switch to advanced synthetic lubricants.

Wenniger Compressor Co. of Milwaukee, Wisconsin supplied four Atlas Copco air compressors for use in the Energy Conversion Efficiency Laboratory. Instrumentation is set up to monitor the power consumption, oil temperature, exit air temperature and pressure, ambient air temperature and pressure, sound pressure, and ambient air relative humidity (RH). A PC-data acquisition system and data acquisition software package are being used to record, display, and store the experimental data. A thermal imaging camera and related software is also being utilized for real-time temperature monitoring and storing of any significant temperature responses that may be observed.

 

 

Lab Setup

The Energy Conversion Efficiency Laboratory is currently running four Atlas Copco air compressors. The compressors are piped with 1" PVC to a ventilation duct. The piping is terminated with noise filters supplied by Wenniger Co. The piping includes an orifice plate installed at a union 1’ from the compressor exit air hand valve. The purpose of the orifice is to simulate a load on the c ompressor. The orifice plate was sized to allow the compressor to stay running at a constant load without unloading. The compressors are being run, and the operating conditions being monitored include power consumption, oil temperature, exit air tempera ture and pressure, ambient air temperature and pressure, sound pressure, and ambient air relative humidity. The operating conditions are monitored with a data acquisition unit that is connected to a PC. The monitored conditions are analyzed and graphed using Sigma PlotÒ software.

 

Brief Description of Air Compressors

The Atlas Copco compressor model GA 11 is a stationary, single stage, oil-injected screw compressor driven by a Siemans 460V 3-phase 11 kW electric motor. The compressors are equipped with electronic regulators designed to control and pr otect the compressors and monitor components subject to service. The loading and unloading pressures can be programmed. The regulator monitors the rate of decreasing pressure when unloading. When the expected unloading period exceeds a programmed value , the regulator will stop the compressor. If the expected unloading period is below this programmed value, the regulator will keep the compressor running to prevent short cycling. The regulator unloads the compressor for 30 seconds after manually stopp ing the unit.

The electronic regulator also monitors temperature at the inlet and outlet of the compressor. If the pre-programmed limits are reached the regulator will shut down the compressor. The compressor will also be stopped in case of overload of the compres sor motor. This will be indicated on the display to inform the operator of the reason for the shut down.

 

The electronic regulator also continuously monitors critical components subject to service (oil, oil filter, air receiver, and air filter). The hours logged on each component is compared to pre-programmed limits. Exceeding these limits causes a messa ge on the display to warn the operator to replace the indicated component. Similarly, the regulator also tracks running hours, loading hours, and compressor starts and stops.

 

Description of Sensors and Data Acquisition

Watt Transducer

A watt transducer supplied by Ohio Semitronics supplies an output of 0 – 10 VDC, which corresponds to 0 – 60 kW. To achieve more accuracy from the watt transducer th e supply voltage wire was wrapped around the current transformers two times. This would give an output reading of 0 – 10 VDC that corresponds to 0 – 30 kW. The watt transducer can handle up to 100 amps and 300 VAC on each line. The accuracy is ± 0.2% of reading or ± 0.04% of full scale. These accuracies include the combined effects of voltage, current, load, and power factor. The response time to read 99% of a step change in power consumption is less than 400 ms.

 

The watt transducer, model GW5-008D, contains transformers for each leg of power. The transformers step down the in-coming current. This current is then wired to the transducer. The transducer is also wired to each leg of the in-coming 460V supply. The transducer requires 120VAC power. By monitoring the current and voltage, the transducer supplies a DC output proportional to the kW power consumption.

 

Thermocouples

Omega type T thermocouples are mounted in the oil fill plug of the air receiver and on the moisture trap at the air exit. Type T thermocouples consist of copper/copper-nickel alloy combinatio n. The temperature range and accuracy is – 76 to 212F 1.8F. The thermocouple leads are fed into the data acquisition unit. The t ime required to reach 63.2% of an instantaneous temperature change is between 4 and 5 seconds. These thermocouples can be connected directly into the data acquisition unit. The thermocouples are secured in place by compression fittings that are machined into the oil fill plug and an access cap on the moisture trap. The compression fittings make an airtight seal around the 6" ungrounded 0.125" dia. 304 stainless steel sheath.

Pressure Transducer

The pressure transducer supplied and mounted on the moisture trap is being used to monitor the exit pressure. The pressure transducer is a Kavlico Corp. model P165-5162. It has an input pressure range of 0 to 246.569 psig. Its output voltages are 0. 500 VDC at 0 psig and 4.500VDC at 246.569 psig with a maximum allowable deviation at 68F of 1.25% of full scale. The response time is 15 ms maximum to read 63 % of full-sca le pressure with a step change of pressure on the unit. The output readings are monitored from the electronic regulator and fed into the data acquisition unit

 

Ambient Condition Sensors

The ambient conditions being monitored are air temperature, pressure, and relative humidity. The ambient air sensors are supplied with DC voltage by a dual output lab bench DC power supply, model HP E3620A. The correct DC voltage sho uld be known and set on power supply before powering any sensor.

 

 

A Humitterâ 50Y, manufactured by Vaisala Inc. is used to sense relative humidity and ambient air temperature. The Humitterâ requires a power supply of 7 - 28 VDC. The humidity sensor, INTERCAPâ (part no. 15778) has an output of 0 – 1 VDC corresponding to a range of 0 – 100% RH. The accuracy is ± 2% RH at 10% RH and ± 3.5% RH at 90% RH. It takes approximately 15 seconds to respond to 90% of a 1-% change in RH. The temperature sensor used in the Humitterâ 50Y is a 1000-ohm platinum resistor (Pt 1000). It has an output of 0 – 1 VDC that corresponds to –40 - 60° C. Its accuracy is ±0.5°C at -10°C and ±0.8° C at 60° C. Its response time is dependent on air movement.

 

A SETRA, model number 276 barometric pressure sensor is used to measure ambient air pressure. This sensor requires 9 – 14 VDC input and supplies an output of 0.1 – 5.1 VDC that corresponds to 800 – 1 100 milli-bar. The accuracy is ± 75 mbar and requires less than 5 ms to respond to a 75-mbar change in pressure.

 

 

Data Acquisition Unit

The data acquisition unit is a Hewlett Packardâ HP 34970A. This unit allows for direct measurement of thermocouples, resistive temperature devices (RTDs), thermistors, dc and ac voltages, resistances, dc and ac currents, frequency, and period.

The unit has three module slots in the rear to accept the data acquisition or switching modules. A HP 34902A 16-channel reed multiplexer will be used to accept the output signals of the instrumentation used to monitor the operating conditions and ele ctrical power consumption. This multiplexer has 16 channels of 300 V- 2 W switching capability, a built in thermocouple reference junction, and a maximum input current of 50 mA. This module is not capable of monitoring ac or dc current. Each channel is fully isolated, yet the module can be configured for 4 – wire resistive measurements. The wires carrying the monitoring signals are connected to the screw terminals inside the 16- channel multiplexer.

 

Following will be a brief description of the front panel controls followed by a description of the current data acquisition setup.

 

 

Front Panel Controls

A channel is selected by turning the knob. (see Fig. 10) The front panel display will page through the channels. The channel number is a three-digit number. The first number corresponds to the slot number (100, 200, 300). T he next two digits correspond to the channel number. The navigation arrow keys on the front panel are used to skip from the slot number to the channel number.

 

The measurement parameters are then selected. Pressing the measure key will bring up measurement choices. The knob is used to scroll through the measurement choices. When measure is pressed again, the menu automatically guides one thro ugh all relevant choices to configure a measurement of the selected function. To monitor any configured channel, turn the knob until the desired channel is displayed, then hit the MON button. Only one channel can be monitored at a time, but one c an change the channel being monitored by turning the knob, even during a scan.

To set the HP 34970A to record and store data, a scan list must be set. Configuring a channel for measurement sets that channel into a scan list. A scan can be initiated by simply pushing the SCAN button to start the scan and pushing the SC AN button again to stop the scan. The instrument automatically scans the list of channels in ascending order. Any channel that has not been configured to read an input signal will automatically be skipped during a scan. Up to 50,000 readings can be stored in the unit’s non-volatile memory. Readings are stored only during a scan and all readings are time stamped. If 50,000 readings have been received and stored and the scan is still initiated, the MEM enunciator will appear on the fr ont display and the new readings will overwrite the first readings stored. Each time a new scan is started the unit erases all readings stored in memory from the previous scan. The initiation of a scan can be controlled through external signals received or by a remote software command. The unit’s internal timer can be programmed to start and stop a scan. A time delay between scans can also be programmed into the scan list.

 

The rate at which the unit scans through the channels is dependent on the signal being read by each channel. The channel can be configured to read a signal within a certain range of inputs or to automatically determine the range. It will take more ti me to read a channel if the unit is set to automatically determine the range. The time required to read a channel also depends on the accuracy desired and the amount of time needed to reject power line noise voltages.

 

The HP34970A utilizes an integrating analog-to-digital (A/D) converter that can reject spurious signals. This is achieved when the HP34970A’s internal digital multi meter (DMM) measures the average input by "integrating" it over a fixed peri od. When the integration time is set to a whole number of power line cycles (PLCs) of the spurious input, these errors (and their harmonics) will average out to approximately zero. When power is supplied to the HP34970A’s DMM, it measures the power line frequency (60Hz), and uses this measurement to determine the integration time. For better resolution and increased noise rejection, select a longer integration time.

 

The data acquisition system also utilizes the HP BenchLinkâ Data Logger software. This software is a spread sheet program that allows the data acquisition unit to be controlled through the PC. The software has real-time monitoring capabilities with full graphic response allowing the user to monitor two dependan t channels on the same graph. All front panel operations of the HP34970A can be done through the BenchLinkâ program. The units communicate via. an RS232 cable. The software is installed on the hard drive of a 486 Pentiumâ 32 MB RAM, 150 MHz P running Windows 95â . This PC is also equipped with a Coloradoâ back up, which allows data to be saved on a 3.2 GB (compressed) mini cartridge. A full Help menu and a system set-up example are available in the BenchLinkâ program.

 

To configure the HP34970A to interface with the PC via. an RS232 cable, the baud rate, parity, and flow control must be set. The baud rate is set to 57600. The parity is set for none (8 data bits), and the flow control for XON/XOFF. Th e flow control determines the method of data transfer between the acquisition unit and the PC. For further explanation please consult the HP34970A user’s guide.

 

 

 

Current Setup

Channel

Measurement

ID

Name

Function

Range

Resolution

101

Watt Transducer

DC Volts

± 10 VDC

6.5 Digits

102

Sound Pressure

DC Volts

± 1 VDC

6.5 Digits

103

Oil Temperature

Temp. (Type T)

   

104

Exit Air Temperature

Temp. (Type T)

   

105

Exit Air Pressure

DC Volts

± 1 VDC

6.5 Digits

106

Ambient Air Temperature

DC Volts

± 1 VDC

6.5 Digits

107

Ambient Air % RH

DC Volts

± 1 VDC

6.5 Digits

108

Barometric Pressure

DC Volts

± 10 VDC

6.5 Digits

 

This current setup was configured in the BenchLinkâ program then downloaded to the data acquisition unit. By accessing the channel properties, the power line cycles and channel delay can be edited. Setting the PLCs to two, and the channel delay to zero, the data acquisition unit was able to read each ch annel in approximately 65 ms. This allowed the time to scan all eight channels (scan sweep) to be less than one second.

 

There are two ways to manipulate the frequency of a scan sweep. One way is to configure a wait period from the start of one sweep to the start of another sweep. This is called a scan-to-scan interval. Currently, the scan-to-scan interval is set to 1 0s. If the scan interval is less than the time required to scan all channels in the scan list, the unit will scan continuously. The second way to manipulate the frequency of a scan sweep is to insert a channel delay between multiplexer channels. A chan nel delay can be between 0 and 60 seconds with a 1millisecond resolution. A different delay can be set for each channel. The programmed channel delay overrides the units default channel delay. To obtain the fastest scan interval, the time delay for eac h channel should switched from automatic to time, and the time set to zero.

Data is recorded from the sensors as "raw data"; i.e. the unit records the exact output of the sensors. To obtain the measurements in the correct units, scaling would be required. Scaling of the data would require the acquisition unit to ta ke more time recording individual channels. Any necessary scaling will be performed at a later time.

Thermal Imaging Camera

The imaging radiometer, model 760 IR, is manufactured by Inframetrics. This IR camera is designed for real-time analysis of static or dynamic thermal patterns. It is a completely self-contained thermal imaging, archival, and analyti cal system. The base unit contains a integral color LCD and a 3.5" floppy diskette drive, so the unit is capable of down loading and retrieving monitored images to and from a diskette. The typical minimum detectable temperature difference is 30° C across a span of 3 – 12 m m. The worst case temperature measurement accuracy is ± 2° C or ± 2% of full scale. The IR camera is interfaced with a PC running Windowsâ 3.1. This PC has Thermagramâ 95 software installed on the hard drive. This software allows the use of a full size screen display of images. The software is also capable of analyzing images in the same way that the camera analyzes images.

 

The IR camera could be used to monitor slight changes in temperature across different parameters of the air compressor unit. Pictures of the air compressor could be taken at certain times during different runs to monitor any changes that might occur. The IR camera could also be useful in detecting any faulty wiring. Taking IR pictures of the watt transducer, the compressor’s electronic regulator, or the disconnects could reveal "hot-spots" in the wiring. "Hot-spots" would indica te an increase in resistance, which in turn would indicate a bad connection and a possible hazard.

 

 

Laboratory Procedure

The following step by step procedure should be followed when running the compressors to collect data.

The following procedure should be followed when shutting down the compressor and data acquisition components.

Handling the Data

The data will be labeled with the following file format. C1_aaa_01

The first two digits correspond to the compressor number. The next three letters correspond to the run number, and the last two numbers correspond to the channel number. Each individual channel is saved as a text file and copied via. ftp onto the uni versity’s server. Once all channels are on the server, an integration program can be run to determine the average value over a predetermined time interval. The integration program is given in Appendix I. The integration being performed is based on the routine by Gill and Miller, which performs a 3rd-order finite differentiating method using uneven time intervals. The integration program is designed to return all the integrated averages, the time interval of each integration, and the average of the whole data set. The output files of this integration program can then be imported into the Sigma Plotâ spreadsheet program. Once plots are arranged to display the data in graphical form, they can be called up as a template and the new data will be automatically fit into the previous setup. Examples of these plots are shown in Appendix E.

 

Discussion of Preliminary Results

Preliminary data shows a 1.3 % decrease in power consumption for the advanced synthetic oil Royal Purpleâ . Compressor three accumulated 20 hours of data with the original HD-Rotofluidâ oil charge that showed an average power consumption of 12.63 kW. The oil and oil filter were changed and Royal Purpleâ was added and the compressor was run for 50 hours. The 50 hours of data with Royal Purpleâ showed an average power consumption of 12.47 kW. Ambient conditions for both runs did not change.

The following table summarizes the preliminary data.

 

Averaged Values for Runs on Compressor 3

 

HD-Rotofluidâ

Royal Purpleâ

% Difference

kW

12.63

12.47

-1.27

Exit Air Pressure (psi)

83.4

83.6

0.23

Barometric Pressure (psi)

14.4

14.4

0.0

Oil Temperature (° C)

83.1

82.7

-0.48

Exit Air Temperature (° C)

29.6

30.5

2.95

Ambient Air Temperature (° C)

23.1

23.2

0.43

Ambient Air % RH

42.9

42.5

-0.93

8000 hr Power Consumption

101040

99760

-1.27

Cost @ $0.0708/kWhr

$7153.63

$7063.00

Saving of $90.63/yr

 

A less dramatic change occurred between HD-Rotofluidâ and Summitâ SH-68 run in compressor four. The Summitâ SH-68 showed an improvement of only 0.85%. This less dramatic effect could be contributed to the decrease in ambient air relative humidity. The relative humidity dropped from 44.2% to 35.9% between the two runs. Since the barometric pressur e and temperature stayed the same, it could be suggested that there was more air mass in the same volume of atmospheric air. When applying work to this air mass, the pressure would increase faster if the air mass contained water vapor. If the air mass w ere void of water vapor, more work would be required to create the same pressure due to the molecules having a smaller, more compressible nature. Yet, the SH-68 shows a slight decrease in required work.

Averaged Values for Runs on Compressor 4

 

HD-Rotofluidâ

Summitâ SH-68

% Difference

KW

12.88

12.77

-0.85

Exit Air Pressure (psi)

82.4

82.4

0.0

Barometric Pressure (psi)

14.4

14.3

-0.69

Oil Temperature (° C)

83.3

83.3

0.0

Exit Air Temperature (° C)

29.4

29.4

0.0

Ambient Air Temperature (° C)

22.3

22.3

0.0

Ambient Air % RH

44.2

35.9

-18.78

8000 hr Power Consumption

103040

102160

-0.85

Cost @ $0.0708/kWhr

$7295.23

$7232.93

Savings of $62.31/yr

 

 

Although the preliminary results look favorable for the higher-grade synthetic oils, many more hours of data need to be collected before any conclusive evidence can be stated. The oils being tested have an 8000-hour life. It is suggested that at leas t 5000 of these hours be recorded and analyzed for each oil before comparisons can be made.

 

Conclusions and Recommendations

The use of advanced synthetic lubricants appears to make a difference in the power consumption of these bench scale air compressors. Many more hours of data will be required to make any real comparisons. The compressors should be run in a way that simulates industrial use. Resizing the orifice plate to allow the compressor to run continuously loaded was a step in the right direction. It also made collecting and analyzing data easier.

 

It would better simulate industrial use by loading the compressors to different pressures. It is recommended that a programmable pressure-regulating valve be purchased and installed in place of the fixed orifice plate. A programmable pressure-regula ting valve in conjunction with a solenoid valve would enable the compressor to be loaded to an initial pressure for a pre-determined time, unloaded, then loaded back to a different pressure. This would work the compressor and oil in a way that would bett er simulate "real life" usage.

 

It was observed that by slightly covering the air intake of the compressor, the oil temperature would rise. This significant change in the operating conditions gave light to the fact that more consideration should be taken in monitoring the operating conditions. In order to obtain a more complete understanding of the operating conditions, it is recommended that more sensors be utilized. It is recommended that a more defined boundary layer be constructed around the compressors. For example, the side panels should be installed when the compressors are being run. This would allow for temperature measurements on both sides of the panels. Another recommendation is to build shallow ducts around the front air intake and air exhaust on the top of the uni ts. This would serve the dual purpose of protecting the air intake from obstructions and allow for the measurement of mass flow readings across the air and oil heat exchanger. These changes would contribute to a better understanding of the energy flux a cross a boundary layer.

 

 

The work by Justin Seipel has contributed to the having the experimental runs more automated. Other steps in this direction could include an automated condensate drain system. The condensate drains from the compressors could be dumped into a condensa te drain pump. The condensate drain pump would work on a float system. Once a certain amount of condensate is dumped from the compressor moisture trap into a reservoir, a pump will push the condensate through a flow meter and directly into the oil-water separator. The flow meter will record the amount of moisture being expelled by the compressors.

 

Another important suggestion is the completion of compressor logbooks. The logbooks should be a daily summary of what is being done with each compressor. They should include, among other things, the time and date of each run, the oil charge, and the file name given to each set of data. An example page for the compressor logbooks is given in Appendix F.

It should be noted that this report does not include any mention of oil analysis. Little has been done in this area to determine the viscosity of the lubricants. A full analysis should be performed on the oils before the compressors are run, and afte r the oils have been run for 5000 hours. The change in oil viscosity will be a major part of the final analysis.