While there have been significant benefits over the past 10 years with sustained value tools, they have not been able to solve the whole problem in an integrated, systematic way, says Rob Howard.
The refining and petrochemical industry is reliant on improving return on assets and lowering operating costs to be competitive. In these challenging times, doing more with less is essential to driving better results.
To deliver sustained economic benefits, refiners have adopted advanced process control (APC) technologies to improve product yield, reduce energy consumption, increase capacity, improve product quality, deliver consistent process safety and reduce environmental emissions. Essentially, APC allows companies to do more, save more and make more.
In the past decade, model-based predictive control techniques have become a pragmatic choice in the industrial world and especially the oil refining and petrochemical industries. Many companies have recently reported that advanced control delivers benefits ranging from 2 per cent to 6 per cent of increased profit by reducing process variability and allowing plants to be operated closer to their true constraints. APC has always played a key role in helping refineries, in particular, achieve commercial success. However, as the physical plant changes over time, APC models need to be maintained and updated or adapted or they no longer represent true plant behaviour. Whilst there have been significant benefits over the last ten years with sustained value tools they have not been able to solve the whole problem in an integrated, systematic way.
Adaptive Process Control developed by AspenTech is a break-through in technology for maintaining controllers, such as Aspen DMCplus, and it allows engineers to perform maintenance on the controller while it is still on-line and continuing to optimise the plant. There is a significant difference between the traditional approach to controller maintenance (sustained value) and Adaptive Process Control. With sustained value, revamping the controller was done as a project. Under Adaptive Process Control, however, the controller is modified over time in more of a continuous process. This update of the model occurs without the need to take the controller off-line and enables a company to reap the benefits of control and optimisation while the model is under maintenance.
Model quality analysis, which continually runs and assesses the accuracy of the model, can detect when degradation of performance occurs. This same model quality analysis can pinpoint a specific part of a controller and compare it with the original model, thereby determining the underlying cause of the degradation in performance.
First generation APC tools delivered a lot of benefits, but they did not solve the biggest cost issues. With aggressive step-testing the cycles are much shorter, but they are more disruptive. Of course, when the controllers are off-line this results in lost capacity and benefit related to quality that normally accompanies an APC implementation. This type of testing requires constant supervision on the part of plant operators and engineers. The fundamental problem in the past was that the controller needed to be turned off to collect open-loop data suitable for the model identification technologies. Revamping a controller often required up to 80% of the original effort and cost. There is also the latent cost of the loss of capacity and quality. The lack of precision is also an issue in identifying problem areas of the models. Co-linearity detection and repair were not integrated with the other modelling workflows, preparing for model identification was manually intensive and generating candidate models required a lot of activity by the control engineer.
Another core issue that companies face is that APC maintenance was performed infrequently and the methodology used mirrored that of the initial project. That would produce some undesirable side-effects. For example, maintenance was commonly deferred until turnarounds and the interim controller performance was allowed to degrade over time.
Sometimes, operators would lose faith in the application and turn off the controller. The economic impact with APC maintenance is also a significant problem. The controller downtime, including revamp and turnaround in addition to normal downtime, can reduce total benefit by 35 per cent across a five year cycle. So, for a typical unit, annual benefits of $2.5m (low end) can be achieved, amounting to a loss of $875,000 per APC application.
When this is multiplied across an average of 9 major units under APC control per refinery, this totals almost $8 million over the five year period that is potentially lost in benefits from reducing the downtime of the controller. So, 60 per cent of this downtime is spent in revamp, 30% is the underutilisation period of the controller when it is not operating at peak performance and, therefore, not generating maximum benefits and the final 10 per cent is unavoidable where the unit is in turnaround. By reducing some of these downtime periods and decreasing the amount of underutilisation that is sub-optimal performance, a company can substantially increase the benefits of APC.
Making the difference
Today, with Adaptive Process Control engineers can do everything required to update the models without the need to turn off the controller. They can also ensure that the controller exhibits robust behaviour during the periods between model updates, make maintenance a built-in and continuous part of the process and eliminate the need to wait for turnarounds to revamp controllers. Crucially, the software permits the engineer to always be in the loop in terms of making decisions when new models would replace existing ones in on-line applications.
For more than 20 years, Aspen DMCplus has been the industry standard for advanced process control software. The application has been successfully applied on large and small units and is the most scalable commercial technology in the industry. Today, with a completely re-designed modelling environment and the introduction of Adaptive Process Control, companies can identify poorly performing aspects of the APC model and pinpoint the areas of the model needing attention.
The additional benefits of Adaptive Process Control are to squeeze out costs by making improvements to existing solutions and adopting new technologies:
* Reduce controller maintenance labour costs
* Minimise erosion of benefits
* Return controllers to service faster after turnarounds
* Reduce Controller Maintenance Costs by 25 per cent
* Increase benefits by continually improving performance
* 10 per cent increase in total APC benefits over 4 years
One of the most significant differences with Adaptive Process Control is that companies now have the ability to collect new model identification data using small perturbation background testing as opposed to the aggressive step-testing methods traditionally used. As the new data are collected, they are monitored in real-time and using AspenTech’s new automated slicing technology bad data are automatically removed producing a clean data set for model identification. Adaptive modelling creates candidate models and those are automatically presented to the engineer for review. This entire process is monitored by an automated test agent and the engineer is notified in real-time of any problems that occur within the workflow.
Jack Adair, Advanced Process Control Engineer, Valero Corporation recently commented on Adaptive Process Control: "The technology works. We achieved the objective of obtaining quality data without unduly disturbing the process or sacrificing economic objectives. The operators hardly noticed a step test was going on.”
APC increases stability and allows the process to operate much more closely to its maximum performance and economic constraints. The technology is designed to operate above regulatory controls and adjust the set points of the basic PID loops to achieve better control of the higher level variables. Across the plant, APC technologies improve plant stability by reducing the effects of disturbances on plant. The plant, therefore, runs more smoothly than with regulatory controls alone.
Adaptive Process Control from AspenTech eliminates the need to approach APC maintenance as a project. It creates a continuous process of collecting process data, assessing model quality and generating new models as the behaviour of the plant changes over time. With greater control, refineries and petrochemical plants can increase overall operational efficiency while keeping the process between safe limits of reliable operation and ultimately deliver greater profit.
Rob Howard, director of business consulting, AspenTech, Reading, UK. UK.www.aspentech.com