Here are a few words from my colleague Alicia DuBay who has just returned from the spring meeting of the Center for Operator Performance (COP).

I have not been very deeply involved with the COP for the last few years due to a new job assignment.  However, it was fantastic to get back in touch, not only with the people, but with all of the existing and new research projects that we are funding.

Since I last attended a COP meeting, there have been quite a few new projects started or continued with additional phases.  These projects include a continuation of Event Prediction, Effectiveness of various Training Methods for process operators, Data to Information Handbook, Alarm Formatting and Presentation and perhaps most exciting - an update on our Overview Displays project.

While all of these projects are valuable and the research is providing some terrific insights into expertise, how operators learn and some guidance on grouping and visualizing data, the Overview Displays project is the one I am currently obsessed with.  This project is member run and is focused on taking some of the results from our completed research, combining that with member experience and developing a set of “best practices” for creating a plant overview display for use by operators in a control room.  

Of course such displays exist today, but little exists in the way of defining the best way to aggregate and display the necessary data.  We are currently part way through creating the best practices from the results of our research and from reviewing existing graphical displays to distill common mistakes and areas for improvement.  Once these guidelines are in place, a new Overview graphic will be implemented at one of our member company’s facilities. 

Once the operators become familiar with the new display, additional research testing will be done to measure the improvement in operator performance based on criteria like how long it takes to find critical information or reaction time to a critical alarm.  This project will provide a real world example of how to apply this research and will be supported by metrics quantifying the improvement in performance.  I can hardly wait to see the results.  I hope to have additional information to share after our fall meeting.

You can find additional information on the Center for Operator Performance at www.operatorperformance.org.  Check it out.  We are accepting new members.  As always, member or not, we would love to hear your thoughts on our Overview Display project.

Awesome report from our colleagues in Corporate Research in Germany - Dr. Martin Hollender and Mr. Moncef Chioua

************************************************************************************************

ABB Corporate Research in Ladenburg has successfully tested a Mass Data Display (MDD) in a waste incineration plant. Mass Data Displays are able to ergonomically show hundreds of process signals in one single view. MDD are good candidates to be permanently shown on large screens. This allows variables to always be shown at the same spot - which supports the spatial recognition capabilities of the human brain (SDCV - spatially dedicated, continuously visible).

A key concept of Mass Data Displays is to show normalized values. Normalization is an important element of recent initiatives such as High Performance HMI. The idea is to focus (not only) on the absolute value of a given process signal but also to show its value relative to the desired operating ranges of the process. As attractive and straight-forward this idea might appear, a closer analysis reveals that several issues need to be considered:

• During the transient phases of the plant such as startup, shutdown or grade changes the desirable operating ranges of several signals is dynamically changing. A complex dynamic simulation model of the plant would therefore be required in order to accurately predict these ranges. As such models are usually not available; we have decided to focus on steady state only.
• For some of the process signals, operating ranges can be easily derived from technological considerations, e.g. a too high lubrication oil temperature might indicate a problem with the bearing but for other signals, the desired operation ranges might not be so obvious.
• Some ranges vary with respect to the operation mode of the plant, e.g. they are different depending if a power plant is running at its full load or at its half load capacity.

In the picture below, each process signal is associated with a “compass needle” on the display. If the needle is in horizontal position, it means that the process signal value is optimal for the current plant state. The closer the angle of the needle gets to 90° or even to 180°, the more “unusual” the signal behaves. By clicking on suspicious “needles”, the operator can investigate more deeply the corresponding signal behavior in a trend display.

This visualization tool enables monitoring slowly developing effects like e.g. the efficiency of a heat-exchanger due to a dirt accumulation. Good knowledge about the degree of fouling is very important for the planning of maintenance actions.

Although the concept of Mass Data Displays has been around since quite some time, the required configuration and calibration effort has limited its widespread usage in industry so far. High-fidelity process models are usually cost prohibitive. The key focus of our research project was therefore to provide tools allowing an efficient calibration of Mass Data Displays. Selected historical data episodes with validated nominal process behavior are used to define a reference for the desired signal ranges.

Calibration of MDD elements

For each plant state, typical intervals are recorded and displayed as histograms. They show how the process signal values are distributed over time. The center of gravity (4 in the example illustrated above) can be used as reference value for the process signal. The more the value goes outside the range of three standard distributions, the bigger the angle of the “needle” becomes. The sensitivity of the Mass Data Display depends on the factor of the standard deviation corresponding to an angle of 90°. Using a high factor value makes the needles mostly horizontal with the drawback that interesting effects like developing faults will be less visible.

The current state of a plant is determined with the help of a reference signal. In the case of a power plant, this is usually the generated electrical power signal. Once this signal is in steady state, we use linear approximation from previous steady states to estimate the center of gravity for all the other signals.

With such a calibration, the Mass Data Display can show how the process has behaved in similar situations of the past. It is driven by a static steady state model of the process. Deviations from horizontal will sometimes appear even in normal situations, but as the basic concept can easily be understood by operators, the Mass Data Display is acknowledged as a helpful support tool.

********************************************************************************************************
Let us know what you think of the usefullness of such a tool and if you like to see it as part of your 800xA Control System.

If you want to download this in white paper form, click the following link.

This post is the final in a series of 5 by our colleagues Marc Leroux and Simo Saynevirta on Tranforming Data into Actionable Intelligence.
******************************************************************************************************

How to get there?
Technology is just an enabler; investment in technology will not, by itself, translate into value. Further, having a multiyear improvement plan can become a liability. It is often better to have a five-year plan where one is always at year one and, as new factors are incorporated, opportunities and results are re-evaluated. This means that companies can stop focusing on sunk costs and look at the incremental value that change brings. 

A key action of this strategy is to drive business value to all levels of an organization. Every decision, from the operator level to the senior management, should be based on its value to the business. This brings one back to the collaborative environment: Seamless access to information allows operational decisions that factor in customer requirements, quality, asset condition and cost to provide an optimal recommendation. Data from multiple systems must be consolidated to provide a picture that focuses on business objectives and drives business value.  

A prerequisite for this is quality industrial software that has been developed with integration in mind. An architecture that is designed to seamlessly integrate with other systems, collect information at high resolution from underlying decisions and store it so it can be quickly retrieved and visualized is key.

The underlying automation also needs to be designed to take advantage of this integration. This is the case with ABB’s flagship automation systems, Symphony Plus and System 800xA, as well as with the underlying sensors and devices. And the solutions must extend up to the enterprise level to include operations, reliability and enterprise asset management (EAM) products.

Finally, all of this needs to be tied together by services carried out by people who understand the domain and who know how to utilize the technology and solutions to drive operational excellence and business value.

ABB has the technology, the products, and the services to implement a collaborative manufacturing system. All the components are in place. The last action is to work toward this goal together, manufacturer and supplier, in a collaborative fashion.

************************************

I hope you enjoyed this article and if you would like to download the complete document, click here.  As always, we look forward to your comments.