Optimization of Multiple Centrifugal Compressor System Delivers Immediate Savings

Greif | Riverville, VA

Where We Started

This project occurred at a semi-chemical medium paper mill with an annual capacity of approximately 200,000 tons. The plant’s compressed air was supplied by four water-cooled Ingersoll Rand centrifugal air compressors—two 700 HP compressors in Mill #1 and two 350 HP compressors in Mill #2. The plant operated 24 hours per day, seven days per week, and prior to this project, all four compressors were utilized to support production.

The compressed air consisted of two 700 HP compressors in Mill #1 and two 350 HP compressors in Mill #2. The two mills were separated by over 2000 feet and were operated by different production and maintenance groups. All compressors operated off their local controls and local set points, and typical air demands were between 8000-9500 SCFM, resulting in frequent periods where multiple compressors were blowing off. Prior to this project, peak demands required the use of all four compressors. Additionally, if a compressor was manually stopped, it would be restarted within a reasonable amount of time.

The Opportunity

The high-level control objective was to run the minimum number of compressors while simultaneously protecting each large motor from too frequent starts or load cycles. Ideally, all compressors would be operating in their upper throttle range, with none blowing off.  The compressed air “load” would be shared across each operating compressor and a compressor would be stopped if the available capacity in the other three machines exceeded the current air demand.

The pressure differentials between each compressor’s discharge and the production air headers in each mill also needed to be accounted for in order to accurately load-share between each compressor.

Our Solution

CAM Technologies engineered and commissioned a CAMLink™ Premier Automation System, based on the Allen Bradley Compactlogix Platform, with a main control panel in Mill #1 and a remote panel in Mill #2. The facility did not have any Ethernet connectivity between the two compressor rooms, but a shielded-twisted-pair of analog wire was present in an underground conduit. CAM Technologies utilized this existing infrastructure by developing a custom Modbus RTU slave routine operating on an Allen Bradley PLC.

Load-sharing logic was installed with a custom pressure-balancing algorithm to account for the pressure differentials in piping and clean-up. Through discussions with plant personnel and analysis of real-time air demands, the Mill #1 compressors were chosen to trim when air demand decreased.

Historical data was analyzed to develop a custom Trim Expert algorithm to determine when the air demand had sufficiently decreased to the point where the excess capacity in the other three compressors could supply the plant’s total air requirements. This algorithm analyzed real-time inlet valve and blow-off valve position along with discharge pressure to determine excess capacity. Configurable cool-down and minimum run timers protected the motor from excessive starts or load cycles.

Additionally, all system settings were optimized over multiple weeks through CAM Technologies CAMLink™ Online platform utilizing remote access to persistently monitor performance and efficiency.

 

Their Results

Improved operating efficiency.

Through coordinated control of the four compressors, an average of 650 HP was turned off resulting in annual energy savings for the facility. Annual runtime on the compressed air equipment was reduced by approximately 8,000 hours, creating an additional reduction in maintenance expenses.

Reduced total cost of ownership.

Electrical energy savings were significant, totaling approximately $284,000 in savings at a burdened electric rate of $.065/KWH. This resulted in a simple project payback of 4.7 months.

Improved system awareness and performance.

With CAMLink™ Premier system automation and the CAMLink™ Online platform installation, all current and historical data points from the system and all connected equipment are displayed and trended for real-time analysis.

By the Numbers*