November/December 2012
Cover Story

The need for enterprise control

Automation engineers improve business performance

Fast Forward

  • Who needs real-time enterprise control?
  • The challenge: connecting the islands of automation and information.
  • An exciting new role for the control engineer.
 
By Peter G. Martin
 Need1

The concept of delivering enterprise control on an enterprise control system (ECS) was introduced to industry in 2006. As with any new approach in the control and management of industrial processes and businesses, it has taken some time for the value of enterprise control to be understood and embraced. The forces accelerating the pace of business are stronger than ever, however, which increases the urgency to find a workable real-time control solution that encompasses the entire enterprise, not just the plant floor.

The move toward real-time industrial business

Although many factors are driving the need for effective real-time control and automation, the one that executives seem to be struggling with the most is the transition of their critical business variables from transactional stability to real-time variability. Only a decade ago, the business variables associated with industry had been highly stable over long time frames. Electricity pricing changed every six months. Raw material costs may have changed monthly. Product costs were set for months at a time.

The business environment was highly stable for well over a month at a time. With this type of stability, the profitability of industrial operation could be effectively managed by simply controlling the efficiency of the plants in real time and managing financials monthly. Increasing operational efficiency directly translated into increased profitability. Real-time control systems could do their thing somewhat independently of business systems, and the industrial business would be successful and profitable.

Over the last decade all of this has changed. Today, the price of electricity and natural gas changes every 15 minutes in the U.S. The price of raw material may change multiple times a day. And the production value of an operation may change multiple times each minute. Just watch a metals price tracker on any financial network and you may see the market price of metals changing every 30 seconds.

Today, not only do the production environments require real-time control, but the business environments do as well. Monthly financial statements are just not frequent enough to control business variables that change multiple times a day. The solution to this daunting industry challenge is to apply control theory to the real-time business variables. This will require resurgence in the field of control engineering, expanding real-time computing domains well beyond the traditional distributed control system (DCS) and programmable logic controller (PLC) boundaries.

Automation engineering talent is often reluctant to take on the real-time business challenge. When industrial automation engineers review annual reports and other financial reports and see how complex these reports are, their first reaction may be shock. Financial reporting of industrial operations is indeed quite complex, but today not all business variables fluctuate in real time. The primary industrial business variables that may change in real time are energy costs, raw costs, and production value. These components of profitability are typically constrained by the physical process, plant safety, and environmental concerns. Since the physical process cannot be altered in real time, the real-time profitability model deals with the real-time interaction of the other five variables as displayed in the simplified real-time profitability model (Figure 1). Applying control theory to these real-time components of profitability is a good starting point for bringing industrial profitability under control.

Need2  

The solution: enterprise control

The real-time business environment presents a daunting challenge to industrial executives. The solution requires expansion in both the scope and the functionality of automation systems. The scope must expand real-time control to encompass both plant efficiency and plant profitability. The functionality must expand the real- time systems domains to encompass the entire industrial enterprise.

Real-time control has been successfully applied to control industrial processes for many decades to maximize the efficiency of the production operations. It is so pervasive that it is sometimes perceived as basic functionality, even though the technology of process control has been continually evolving and is actually quite sophisticated. The effective application of real-time control can add order to what might initially appear to be a chaotic situation. When process plants were first developed, the operation appeared chaotic because of the large number of variables that changed value in real time and had to be controlled together to produce the desired result. Control engineers throughout industry have learned how to bring order out of chaos and improve efficiency on the plant floor by applying control theory effectively.

Today, not only are plant processes themselves more challenging; the business processes also present challenges, which require real-time responses. Effective application of real-time control of both the plant and business domains is the only solution. Control engineers must expand their horizons to bring the control of business variables into their domain of expertise. The same basic control theory that has worked so successfully in real-time plant operations also applies to real-time plant profitability. The key is to build profitability control loops (Figure 2). This can be accomplished by first measuring the critical business variables in a real timeframe. The information commonly stored in an enterprise resource planning (ERP) system does not have the timeliness for a profitability control loop. Yet experience has shown that these real-time business measures can be modeled in the installed industrial automation system from the database comprised of the vast number of process sensors already installed in the plant along with some key business variables, such as current energy cost, raw material cost, and product value.

Operators close the loops

It would be nice if a profitability controller existed with the same general use characteristics as the traditional PID controller used for process control, but it does not. Part of the reason is that there are no specific natural periods for profit loops, as there are for process loops. But the first approach to closing the loop may be to emulate early process control systems and use process operators to manually close the loop. This can be accomplished by providing the real-time feedback to the operators so they can understand how the actions they normally take, such as setting a setpoint or putting a control strategy in manual, impact the profitability of the operation. This can be accomplished by supplementing them with real-time performance scorecards or dashboards. Over time, they will learn how to bring the profitability control loop into control and to maximize profits from their domain of operation. As better understanding of the dynamics and control approaches of the profit control loops is gained, mechanisms may be developed that automatically close the loops.

Need3  

It is important to realize that real-time profitability control does not replace the traditional process control. Rather, effective real-time process control must be in place for the profitability control to work. The relationship between process control and profitability control can be thought of as a cascade control strategy with profitability control as the primary loop and process control as the secondary loop (Figure 3). Profitability control must first be applied across each plant and then across the entire fleet of plants in the enterprise to realize maximum business benefit. Clearly, profitability control has broader implications and scope than process control and a broader perspective of real-time automation environments is required to make enterprise control a reality.

Need4  

Enterprise control systems enabling enterprise control

One of the critical barriers to implementing enterprise control across industrial businesses is that the installed control technology-even within a single plant-is often comprised of multiple systems and software applications that do not work well together. There are silos of automation and information throughout industrial operations (Figure 4). Implementing enterprise control requires that these silos not only work together, but that they also work together in real time, because the problem is a real-time problem that goes across traditional domains.

Need5

This is not unlike the situation at the IT and business system levels of industrial organizations that led to the development of ERP systems. There had been numerous systems and applications at the business systems level that were not designed to interoperate. Companies, such as SAP and Oracle, responded to this situation by developing enterprise service-oriented architectures to pull the plethora of systems and software together into a common business-computing domain.

Industrial automation needs to follow suit to meet the demands of enterprise control. The development of industrial service-oriented architectures that enable various real-time systems and applications to work together as a common enterprise control system is exactly what is required to meet the daunting challenges facing industrial businesses. The emerging enterprise control systems will act as the real-time counterpart to the transactional ERP systems (Figure 5). The combination of open enterprise control systems working with ERP systems will cover both the real-time and transactional needs of industrial enterprises and provide the technical basis for true enterprise control.

Need6  

One key to profitability for industrial companies is being able to respond to market and business fluctuations as they occur. Many existing process manufacturing plants were not designed to operate with agility, and providing more comprehensive business controls alone will not make them agile enough for real-time operation. Implementing a real-time enterprise control solution across an integrated system, however, will allow them to manage whole fleets of assets as a single business entity. This certainly will not necessarily improve agility of any single asset, but it will add a level of business agility at the fleet level. This may or may not be the complete answer to the requirement for more agility to match the real-time market demands, but it is a strong step in the right direction.

The industrial automation community has never been in a stronger position. Industrial automation and real-time control are the keys to driving profitable success for industrial companies. But capitalizing on the traditional control and real-time automation expertise requires a shift in perspective. The scope of where control can be effectively applied must increase to include the control of real-time business variables. And the scope of real-time automation must increase to encompass entire plants and even entire enterprises. The industrial automation industry is the key to the ongoing profitability of industrial companies. The move to enterprise control is necessary, and the expansion of automation systems to enterprise control systems is essential.

Although the technology underlying enterprise control systems was first introduced to the industrial market in 2006, few industrial companies have been in a position to take full advantage of it and develop corporate enterprise control strategies and implementations. But there have been a number of companies that have started to capitalize on enterprise control approaches in a bounded manner. The results to date have been promising, but the true potential for real-time control of industrial enterprises is only beginning to be realized.

It is a great time to be in this field. It is time to make enterprise control a reality and demonstrate the true potential of industrial automation.

ABOUT THE AUTHOR

Peter Martin, Ph.D., (peter.g.martin@invensys.com) is vice president, business value solutions, for Invensys Operations Management. He has spent more than three decades in the automation industry, culminating with the development of commercially-applied dynamic performance measurement technologies and methodologies. An established author and industry speaker, Dr. Martin received the ISA Life Achievement Award in 2009 for his work in performance measurement.