March/April 2010

Thriving during economic downturn by building real-time enterprise

By Peter G. Martin

Part 2 of a two-part series

The previous article in this series ( discussed the role of people, information, and technology to enhance the performance of existing plant assets in today's challenging economic environment, and how companies can cope by employing real-time techniques in enterprise management using resources they may already have. This article details the real-time business approach to measurement, employee empowerment, and operations, ultimately leading to real-time profit optimization.

A real-time enterprise requires business and operations information to be available to operations personnel and management in real time, but traditional IT systems are not designed to provide information so frequently. They are optimized around monthly, weekly, or at best, daily schedules, and they typically do not contain data that reflect the real-time operation of the business. It is impossible to take monthly data and extract minute-by-minute guidance from it.

On the other hand, automation systems were designed from inception to operate in real time. They are also connected to a real-time process instrumentation that reflects everything happening in the plant second by second and can be thought of as the real-time database of the industrial operation.

Granted, this "database" is difficult to use from a business perspective, containing information such as flows, levels, temperatures, pressures, speeds, and chemical compositions, but there is a clear relationship between these basic process variables and the required real-time business variables. Using real-time business information, such as current energy costs that can be retrieved directly from utilities in electronic format, basic process data can be transformed into required business and operating data in real time, allowing plant personnel to make real-time decisions that improve the plant performance. And this typically requires no additional capital investments.

All it requires is talent that can use existing plant automation and information systems in a new and different way to apply real-time operational and business information. When the real-time business data of an industrial operation is utilized to drive business performance improvements in this manner, the resulting operation is referred to as a real-time enterprise.

Although the concept of a real-time enterprise might seem daunting, the basics of controlling and improving such an enterprise are much simpler than they may initially appear. The key is to build on the knowledge engineers have developed over the past 50 years in controlling real-time production processes.

As with any control challenge, the first component that needs to be developed is the measurement of the variables to be controlled in the time frame necessary for control. In the case of today's business variables of industrial operations, the timeframe has to be in real time.

The challenge is how to measure business variables in real time. The business variables of an operation are commonly managed through the company's ERP system, which is anything but real time. Therefore a new database is required that will underpin the creation of the real-time business variables.

Fortunately, in modern industrial plants that are under automatic process control the database exists, even if it is not initially obvious. The real-time data sourced by the hundreds of process sensors installed in most plants provides an ideal database for the development of the real-time business metrics. If the equations of the necessary operational measures (key performance indicators) and financial measures can be determined, typically an experienced engineer can develop models of those measures primarily using the plant sensors as source data. Often additional external information may be required, such as the current price of energy. These models can execute right in the controllers of the plant automation system, providing the necessary real-time business measures.


Once the real-time business measurements have been developed, the second step is to move to bring these measurements under feedback control. As with the early process control systems, the best and easiest starting point is the move to manual feedback control by using the plant operators to take control action. In early process control systems, this was accomplished by assigning an operator to control a specific process variable by turning a hand valve and empowering the operator to make the right decisions through a gauge that indicated the current value of the process variable.

For business variables, a similar approach can be taken. Plant operators can be empowered though the creation business gauges in the form of dashboard displays of the real-time business variables.

These dashboards have to be carefully developed and contextualized to the skills and experience of each individual employee involved in the control of the business. This careful process results in a set of manual feedback business control loops focused on production value, energy cost, material cost, environmental integrity, and safety.

Some business variables are beginning to fluctuate so rapidly that it is becoming very difficult for operators to control them through a manual feedback control system. In these instances, automatic feedback controllers of the business variables will need to be developed. The algorithms for these business controllers may not be as straightforward as the general purpose PID algorithm used in process control.


But careful analysis of each business control problem can often result in the effective development of a special business control algorithm. The result is an automatic business control loop.

Humans also tend to have difficulty with business variables that respond too slowly to changes they make. For example, in a refinery an operator may change a set point to drive a business result; but due to the dead time in the process, the actual result may not be realized for hours. By the time it is, the operator may be at home eating dinner and may never find out the impact of the change. In these cases, technologies such as online simulators may be deployed. When the operator makes a change, the simulator can go into fast-forward mode and immediately let the operator know what the impact of the change will be when it worked through the process.

Once each of the five key business variables of an industrial operation-production value, energy cost, material cost, environmental integrity, and safety-are brought under control, the final step is to optimize the profitability of the plant in real time. The simplified plant profit real-time optimization model shows three business objectives-maximizing production value, minimizing energy cost, and minimizing material costs-constrained by two key business constraint functions, environment and safety.

This model clearly shows optimizing plant profitability is a multiple objective optimization problem. Focusing on any single objective in deference to the others would result in a sub-optimized business. Unfortunately, effective mathematical approaches to resolve multiple objective, real-time optimization problems do not yet exist. Studies have found a human with a moderate education and the right information in the right timeframe will earn how to solve a multiple objective optimization problem through experiential learning-so manual real-time profit optimization is available today.

Every aspect involved in moving to real-time profit optimization involves utilizing the installed plant capital investments in the form of automation and information technologies more effectively to solve new kinds of business problems. In many cases, this can be accomplished without the need to acquire any new technology.


Most control and process engineers have the background and experience to make great strides toward business performance improvement by using the models presented, along with their traditional knowledge and skills as control engineers, and the installed automation and information technologies.

This is a difficult economic time for industry. But such times offer new opportunities for improvements such as described here. With the reduction of capital projects, industrial companies can mobilize their existing human assets to build real-time operations business improvement programs using their automation and information assets. As a result, not only will they ride the downturn with less disruption, they will also be in a great position to capitalize when the economic upswing comes. The time is now. The opportunity is ripe.


Peter G. Martin Ph.D. is vice president for business value generation at Invensys Operations Management. Martin has spent 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, he received the ISA Life Achievement Award in 2009 for his work in performance measurement.