I often feel that I am a lone voice 'crying in the wilderness' about the abysmal state of control in most plants worldwide. It really pleases me when I can produce supporting evidence from other people that this is in fact the case.
This article, written by Dr Maurice McDowell, is about the really excellent work performed by Erich Holm and his control team at Chrome International South Africa. I first heard about it when Erich gave a presentation at a recent Crusader Systems Users Group conference. What I find so interesting about it is that it is one of the few cases I have come across where a plant is prepared to publish information about the previous state of their controls, and then actually present the economic benefits received from the improvements they have affected with optimisation.
I would also like to point out that the work done so far is only on a few of the loops where excessive problems were immediately picked up by a recently installed continuous on-line loop performance monitoring package. This type of device, which I believe is one of the most useful tools to be developed in recent times, immediately shows up severe faults. It must be noted that it does not in itself perform optimisation, or even show up all existing problems in control loops. However, once a plant has been fully optimised it really comes into its own, and will immediately give warning of degradation in the control performance in any loop.
It is also very interesting to note that the substantial financial benefits obtained so far have come about by optimising only a few loops. What further gains will the plant obtain once all controls have been optimised?
Erich Holmes is the maintenance engineer for Chrome International South Africa (CISA) and as such has overall responsibility for electrical distribution, instrumentation, control and IT systems at the company's plant in Newcastle. Erich was with CISA from before the company was registered but, as a result of his wide range of responsibility and first priority of getting the plant online, it was only in 2002 that he had the time and opportunity to look more closely at control loop optimisation and process control in the broader sense.
The plant itself has some 200 important control loops and, based on spending 5 h to get each loop under control on a manual basis, this implies six months just to go through the whole plant once. An in-house study over a period of time indicated that this was totally inadequate as, in terms of non-optimised control loops, some 40% were related to tuning, 15% were process related, 25% operator related and 10% design related (eg, inadequate sizing).
In order to address this seemingly insurmountable problem, Erich, with the support of his management, decided to take a software-based approach to monitoring and optimisation and local company Crusader Systems of Pretoria was approached to assist. Crusader Systems, established in 1997, has developed a range of software solutions that target the process industry both locally and globally and one of its products is CSense (See Sense). A specific product offering in this portfolio is CSense Process+, a control loop diagnostic and optimisation software, with local service being offered by Michael Brown Control Engineering, and international sales being handled through Outokumpu*.
CISA commenced the use of Process+ in January this year with 100 control loops being initially configured. The only additional hardware requirement was a Siemens WINcc OPC scada/server, which replaced an existing obsolescent COROS system, and increased the overall scada functionality in the control room. During April this year the evaluation system was extended to include a total of 195 control loops. At the same time, learning from the initial evaluation, additional sections were added and the loops contained within each section were revised. Additional requirements from CISA included that all control loops should operate within ±5% of limits, safety devices should be in a group of their own and that no controller should be in the off position when the plant was online.
After the installation of Process+, the out of limits values were monitored. Ignoring the results up to April, when the database was being established and the majority of effort was concentrated in the chromic acid section where the number of control loops was small, the starting point on week 14 was that some 18,2% of all control loops were out of limits. By the end of week 29 this had been reduced to just 10.8% and this value will be improved further. These percentages represent the time that the control loops are outside of an acceptable range from the set point. It probably makes more sense, however, to express it in a different way from the reporting system and to say that the control loops by week 29 were controlling within the ±5% band for 89,2% of the time.
Obviously an objective of installing Process+ was to ensure that control loops were being used in automatic mode. In the week that the software came online (week 7) the reporting system indicated that almost 60% of the control loops were switched off. This problem was highlighted and immediately addressed with the results showing a dramatic decrease in following weeks to between 20 and 30%. By week 30 a further dramatic improvement saw only 10% of the loops not operating automatically. Another indicator of the effectiveness of the system is the percentage of time that control loops are operated in manual mode (data again provided by the reporting system). During week 8 this was almost 10% but was reduced in roughly linear form to just over 3% by week 29. Obviously an increase in this percentage would highlight a problem that needs to be addressed.
Overall, the time that control loops were either off or in manual mode was reduced from 41% to 14%, a figure that represents a vastly improved utilisation of the installed capital base. Obviously, with these loops operating optimally in automatic mode, production performance will be significantly enhanced with lower manufacturing costs associated with reduced wastage.
One particular example of optimisation was the loop that controlled steam pressure to the SMC plant. Investigation showed that from start-up of the plant (four years ago) it was found that the pressure controller was not operating effectively on automatic and a decision was taken at that time to operate it manually. Not only was the loop operating 100% manually, a nightmare for the process controller, it was found that 100% of the time it was operating outside of the limits for the set points. As a result of this, steam flow was fluctuating a lot and the whole process not being in control was affecting five evaporators and other downstream plant.
The actual control loop was made up from a pressure transmitter and a steam control valve. There was also a flow transmitter that was used for indication purposes and density measurement. The steam is used to transfer heat via a heat exchanger to the sodium mono chromate (SMC) liquid going to the evaporators with the main purpose being to control the density of the SMC. The system had originally been designed to run the density in cascade with the steam controller.
The solution adopted by CISA was to use steam flow as the controller with pressure being the indicator. After confidence had been established in the functioning of the new steam controller, the next step was to get the density process variable sorted out with the density control loop being placed in true cascade with the steam flow controller. The next phase was to stabilise the flow rate in the downstream plant as this also affected the flow rate through the evaporators. This final control was carried out using two loops, pressure and flow control, and as these directly affected each other they were effectively de-tuned so that they act out of phase and increase stability.
The bottom line for any financial manager is of course the return on investment (ROI) and how quickly capital costs are written off against increased profits. Here the results obtained are phenomenal, considering the fact that to date only a small number of control loops have been optimised. On the expenditure side of the ROI calculation the cost of CSense Process+, installation, S&T, use of internal resources and the new scada system were taken into account. Erich Holmes himself estimated the cost savings to CISA based on the 10 optimal control loops. The cumulative expenditure for the first two quarters of 2003 amounted to some R312 500 while savings came in at an estimated R430 000, indicating total recovery of cost. Extrapolating these figures over a one year period shows that a direct cost saving in excess of R900 000 or an ROI in excess of 200% will be achieved.
The above information is based on the status of CISA as of July 2003 when only the few selected loops that had a high financial impact had been optimised. As of September this year, more than 90% of all control loops are running within limits 24 h a day. The next target is for CISA to reach 95% which, considering that some loops are still out of limit on manual mode, is close to what is theoretically possible.
The improvement in performance at the CISA plant is a good example of why other process plants should be looking into optimisation of their control loops.
Dr Maurice McDowell has many years' experience as a technical journalist, editor, business manager and research scientist. His third party analyses of world-class companies and processes, as well as his insight into industry and technology trends are well respected.
Michael Brown is a specialist in control loop optimisation, with many years of experience in process control instrumentation. His main activities are consulting, and teaching practical control loop analysis and optimisation. He gives training courses that can be held in clients' plants, where students can have the added benefit of practising on live loops. His work takes him to plants all over South Africa, and also to other countries. He may be contacted at Michael Brown Control Engineering, 011 486 0567, [email protected], www.controlloop.co.za
* About Outokumpu
Outokumpu is a Finnish-based metals and technology supplier for the international metals industry. Areas of involvement include construction, industrial machinery and equipment, transportation, domestic appliances and general products, electrical and electronics, and processing. Outokumpu's main markets are Europe, North and South
| Email: | [email protected] |
| www: | www.controlloop.co.za |
| Articles: | More information and articles about Michael Brown Control Engineering |
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version