Driving economic performance through the utilization of automation technologies
- Improving business performance has always been an expectation of engineers working within the manufacturing industry, never more so than the last number of years.
- This case study details how a number of different automation technologies and techniques were utilized to identify performance issues and deliver cost-saving solutions on a large fermentation compressed air system.
- Simple, robust, low-cost automation solutions used to deliver improved business performance through energy reduction and operational excellence-8% reduction in annual electrical charge and 3% reduction in CO2 targets.
By David Twohig
Improving business performance comes as a pre-requisite when you work as an engineer in the manufacturing industry. While this can be achieved in numerous ways, cost reduction is the obvious primary objective for most companies. Operations are continually trying to achieve greater throughput for less input. Therefore, this has become the longstanding challenge to automation systems, as the business tries to become more competitive in expanding their market share and gaining that competitive edge. It is no longer acceptable for control systems to just manufacture product or manage utilities effectively; we must maintain those standards but achieve them more efficiently.
In practical terms, when you consider the different ways of trying to achieve these objectives, improving your energy efficiency is an obvious choice. This case study outlines the benefits of how various automation technologies were utilized to reduce costs and improve the performance of a large compressed air system.
Typically, two of three air compressors are used to supply aeration to an Active Pharmaceutical Ingredient fermentation process. The compressors draw in air from the local atmosphere, and in compressing the air, supply a common mechanical header. Individual fermenter vessels then manipulate their own aeration demand from this header through an air flow control loop. The principle of operation for the compressors originally consisted of the production team entering a common fixed pressure setpoint through the SCADA HMI. The compressors then worked to maintain this setpoint by manipulating their volumetric throughput. Presently, the system is configured to run one compressor at full capacity while the secondary compressor varies its supply as the process demand for aeration oscillates. (Note: This configuration is applied due to the age of the equipment. The older, less efficient compressor is run at its optimal point, which is approx 100%, while the secondary compressor is utilized to vary its output because it has a better efficiency curve.)
Because the supply of aeration is critical to the manufacturing process, historically these compressors were manually manipulated to ensure there was an excess of air in the system. This practice also served the purpose of ensuring the compressors were operating safely within their own performance curves, i.e., away from surge. (Note: Surge is defined as the point at which the compressor cannot add enough energy to overcome the system resistance or in this case the header pressure. This causes a rapid reversal of flow, or surge, back through the system, which can result in vibration and mechanical damage.)
This excess air was then vented to atmosphere via a blow-off valve or through a spare fermenter vessel. And although this method was inefficient in terms of energy usage, it did provide the production team with an aeration buffer. This buffer was used to absorb sudden process oscillations that might impact product quality. And so, over the 50-year life span of the system, this technique became culturally ingrained in the team when running the compressors, and as a result, the systems' overall efficiency suffered, resulting in high running costs.
Historically, this culture would not have been perceived to be a problem given the cost of manufacturing product dwarfed the associated energy costs; however, the company has since become increasingly focused on the unit cost of energy. In addition, people have become more environmentally conscious, and so the expectation is to become more energy efficient. While previous projects had attempted to optimize the performance of the compressors, it proved difficult due to the limited availability of process data from the compressor control system. The net result of all of the above was we knew the system needed improving, but quantifying and evaluating success with its current configuration was going to be difficult. However, because the cost of energy has spiralled, and the culture in managing the compressors was poor (i.e., there was obvious wastage), the system did present itself as the proverbial "low hanging fruit" with regards to reducing manufacturing costs, and so we were about to try again. After some initial investigating, it became apparent the automation technologies needed to drive improvement had evolved over the last number of years and become more reliable and affordable. Therefore, the opportunity to drive improvement became more of a realization.
To ensure the potential savings were correctly identified and quantified for the business, a Lean Six Sigma project was implemented to quantify, optimize, and validate all elements of improvement technically and culturally.
Although the fermentation process relied on the supply of air, the compressors were controlled by an independent system that did not communicate with the main Fermentation DCS. The compressors were controlled on a PLC/SCADA system that operated on a standalone network. By contrast, the fermentation process was controlled by a large DCS that was fully integrated into the site historian. And even though they operated as systems in isolation, they were heavily dependent on each other to ensure manufacturing ran smoothly (i.e., fermentation depended on the supply of aeration, and the compressors depended on fermentation to vary its demand proportionally to ensure they avoided surge conditions). As a result, the production became the common denominator in this balancing act.
Project life cycle
To ensure the project started effectively, different automation technologies were utilized to interface the compressed air control system with the site historian and Fermentation DCS. From this integration, it became possible to evaluate the performance of the compressors in terms of their energy consumption and process performance, particularly with respect to fermentations' processing activities.
By correlating this data with the data from the electrical energy meters on each compressor, the project team was able to quickly identify and quantify the performance/cost gaps. Utilizing this data further, the team was able to develop individual energy profiles for each compressor, and these profiles provided the metrics to which the project would validate its success.
Having examined the overall performance of the entire system, there were two primary issues with the supply of aeration. First, the compressors were utilized to produce an excess of air. Over certain periods, this could be as high as 15%, and in producing this air they did so at a pressure higher than was required. Referring to engineering first principles and heuristic data, the team was able to conclude the power utilization of the compressors was proportional to the volume of air being produced and the pressure at which it was produced. Therefore, the solutions appeared obvious-eliminate the waste air and optimize the pressure. Simple in theory! However, in reality, the techniques to enable these solutions needed to be identified and then implemented on a live manufacturing process.
Initially, a solution was put forward based on ISA's Instrument Engineers' Handbook, Third Edition - Section 8.9 Optimized Load Following whereby the positions of the air-flow controllers on the supply line of the fermenter vessels could be used to determine the optimal operating pressure of the header system. By using a simple selector block to determine which valve across all the fermenter vessels was in the most open position, that value was then compared to an optimal position of 90%. Through an integral-only controller, the difference in this positional relationship was used as the offset to determine the optimal pressure setpoint for the system. (Example: As the pressure setpoint reduced, the varying compressor would reduce its volumetric throughput. As this reduced the volume of air in the header, the supply valves of the fermenter vessels would open up to maintain their recipe setpoint. The setpoint of the pressure controller would reduce the air in the system until the position of most open valves would reach 90%.) This simple solution utilized the valve positions of the vessels as an indirect indication of the air requirements of the fermentation process. And so, increased or decreased, the system pressure to optimize the electrical utilization of the compressors, while still ensuring sufficient aeration, was supplied.
We can see the results from the initial trial period (see chart above), where the optimized pressure setpoint, i.e., the purple line hugs the lower pressure limit of 1.48 barg except for short infrequent periods of high demand. (Note: This data is provided from the initial six-week commissioning period in which a lower limit of 1.48 barg was applied.) However, if you examine the secondary axis (green lines), you will notice there is still a sizeable difference between the active most open valve position and the optimal setpoint of 90%. This was a clear indication the system was producing sufficient aeration at the lower pressure. And that the pressure setpoint could be reduced further in order to drive the most open valve to the optimal setpoint and capitalize on further energy savings.
In terms of dealing with the volumetric wastage, there were two challenges. First, there was an operational issue with regards to the production teams' understanding of how the compressors should run effectively. They knew surge conditions were to be avoided, and so increasing the volumetric throughput of the compressors via wastage provided a "comfort zone" in doing so. To improve the daily management of the compressors, the production team had indicated if they had a simple way of monitoring surge, they would be willing to operate closer to the edge of the proverbial surge cliff. In response, the automation group was able to capitalize on the integration of both systems and utilized existing compressor PLC code to develop some surge profile displays for the fermentation HMIs. Because the PLC control strategy utilized some simple mathematical techniques to develop its own compressor profile, these displays were able to convey the dynamic position of a compressor in relation to its surge profile. In simple terms, the operator could look at the screen and understand the compressor needed to operate with a specific envelope and as a result could pre-empt manufacturing decisions to ensure it remained within said envelope.
In the compressor charts, we can see the operating zone displayed on the operator HMI showing the dynamic position of the compressor relative to some earlier warning alarms and the actual surge line.
Second, there were legitimate situations in which some wastage would be required. One example was if the manufacturing demand for air was more than one compressor could supply but less than the minimal output required to run both compressors safely.
To address this legitimate wastage requirement (although tolerating wastage is technically not a Six Sigma policy), the automation group developed an "Air Waste Management Strategy." Implementation of this software solution enabled the production team to activate a recipe in the DCS that would continually minimize the waste air required to safely run two compressors. It did this by manipulating the aeration throughput of an empty fermenter vessel based on the difference between the real demand and the minimum throughput required to run two compressors. By creating this dummy demand, the compressors interpreted it as a real aeration requirement and so responded accordingly. This resulted in the compressors operating safely away from surge but had the added benefit of increasing or decreasing this dummy demand as production's requirements oscillated, i.e., as legitimate demand came online, the software recipe would reduce the air demand of the dummy vessel.
Because the Six Sigma concept is to categorise results in terms of benefit, i.e., Type 1 to 4 with Type 1 being balance sheet savings and Type 4 being indirect savings such as operational improvements, it therefore provided the team with a strong framework in which to quantify the success of the automated solutions.
Initially, the team utilized compressor energy profiles to determine certain savings. In this example, we can see the impact the pressure reduction strategy had with respect to the varying compressor. Aligning the "Varying Compressor" chart with the earlier trend, and it is evident the compressor is producing sufficient air volume, but in doing so at a lower pressure, it has reduced its energy consumption. (Note: This analysis represents the previous chart, and so there is a lower pressure limit in place. Because the process data indicates the limit can be reduced further, we can conclude greater energy savings are to be realized once the positional setpoint is optimized.)
In addition, the team was able to quantify a reduction in volumetric wastage utilizing the site historian. It was evident from the data that the air being produced was being utilized more efficiently compared to before the project. The direct savings have been quantified in the region of an 8% reduction of a typical annual electrical charge and 3% reduction in the sites' CO2 target emissions.
While this is considerable, there have also been other significant improvements that have contributed to improved business effectiveness. The site historian is now used by the production and support groups to track the performance of the compressors. By monitoring various trends in conjunction to the HMI displays, the production team has been able to improve the day-to-day management of the compressors, keeping costs and environmental impact minimized.
In addition, since the process data has become available, the maintenance team has been able to monitor the performance of the compressors' equipment more effectively. This monitoring has enabled the team to proactively calibrate equipment and prevent situations that historically have shown the compressors to trip out. Because of this pro-active culture, the compressors have tripped out only once in 18 months compared to the three trips in the first quarter, prior to the project being implemented. This has not only led to improved performance, but it has contributed to changing the culture of the production team with regards to managing the compressors.
While this is a specific example that may not be applicable to all compressed air systems, it does demonstrate how, with some lateral thinking, different automation concepts can be applied to reduce costs and limit environmental impact. When you consider that the most significant contribution in terms of reducing the costs of running the compressors was the pressure reduction strategy, this in automation terms is just a simple integral only controller. The case study therefore demonstrates that low-cost automation solutions can have a large impact on improving business performance, and automation has a large role to play in driving cost reduction for companies. The smarter use of our systems should be seen as the key enabler when it comes to improved business performance and optimizing existing utilities.
ABOUT THE AUTHOR
David Twohig (Twohig_William_David@Lilly.com) has been working as an automation engineer for Eli Lilly in the U.K. for a number of years. He has a degree in Applied Physics and is presently working on a research Masters with the Chemical Engineering Dept. at Newcastle University. In addition, he is a full member of the Institute of Engineering & Technology and is presently serving as an executive committee member for the Control and Automation Community. I would like to particularly acknowledge the contribution of Ashok Subramaniam (Intuitive Engineering). His technical skill set and dedication throughout the delivery phase of the project truly ensured it was always going to be successful.