• No results found

Implementing Decision-support Portals based on Data Visualization Best Practices

N/A
N/A
Protected

Academic year: 2021

Share "Implementing Decision-support Portals based on Data Visualization Best Practices"

Copied!
18
0
0

Loading.... (view fulltext now)

Full text

(1)

Implementing Decision-support Portals based on

Data Visualization Best Practices

Grenn Holden

Matrikon, Inc.

June 2007

(2)

Implementing Decision-support Portals based on

Data Visualization Best Practices

Valuable information is hidden in the vast amounts of data being collected at today’s process industry facilities. Finding that information and presenting it to the right people is a significant challenge facing industry today. Finding the right data and analyzing it with the right tools within the correct context transforms plant data into actionable information. Actionable information

enables corporations to more quickly make business and operational decisions, respond to changes in the market, and deal with upset conditions.

Getting plant data to the user community is the goal of many industrial information projects. A more effective goal is to get the right information to the right people at the right time. Having the right people spend more time with the right information enables them to find areas for

improvement and cost savings instead of wasting their time looking for relevant data.

Before a system can truly provide value to the organization, you must consider and overcome several problems. This paper will discuss the best practices I have discovered implementing user-centric decision-support systems for small, large, and global companies. I will discuss issues surrounding proper architecture, appropriate data consolidation, the use of KPIs and metrics, advantageous visualization techniques, and decision-support strategies. Figure 1 illustrates the concepts in a hierarchical diagram.

(3)

Architecture

The architecture of the solution is the foundation of the system. This can make or break any information project. The first issue to consider is ensuring a secure and reliable connection to your data facilities such as historians, maintenance databases, lab systems, and financial information. No matter how much time and effort is spent on the later steps nothing will affect the acceptance and usability of the system more than consistent and fast access to the raw data and information users are after.

A consistent issue at many facilities is the proprietary nature of control systems and the barriers between systems as seen in Figure 2. Each system requires its own analysis and management tools, leading to training and maintenance problems. Having a thick-client application installed on all users’ computers requires a license fee and annual maintenance for every install, as well as support personnel to ensure smooth handling of upgrades and troubleshooting.

Figure 2: Different DCS, different historians, proprietary clients

Introducing tools that bring the disparate systems together at the lowest possible level alleviates the need for proprietary interfaces. OPC has become the published industry standard for

integrating the control layer with the information/application layer. This approach addresses the first part of the data consolidation problem. Having your plant data consolidated allows you to begin correlating your business data such as LIMS, alarms, maintenance, and financial.

(4)

Figure 3: Different DCS, common historian, common web client

Another issue to consider at this time is ensuring a secure control network while providing easy access to the appropriate data. This can be addressed with a three-tier network strategy that isolates the control network from external disturbances and issues such as denial of service attacks or excessive network traffic generated by worms and viruses. A three-tier network, similar to the one represented in Figure 3, restricts network access to a one up/one down style of

communication. This restricts anyone from the business network from connecting directly to a DCS and having unrestricted access. DCS data is available to the business network through an historian and web visualization application.

(5)

End users require easy access to the system and to the information they need. Balancing this with the need for a secure environment is the function of the three-tier architecture. Using network security allows for easy access and seamless connectivity to the new decision-support system. Reducing the number of passwords and logins a user has to use to do their job is important in that it makes the adoption of a new system easier.

Data Consolidation

The number of data repositories in a plant is continually growing. With each new data source users must learn a new tool to access the data and analyze it. Each silo of data has its own method of data access and it is very difficult to correlate the information from these disparate sources. Collecting operational data and compiling it into production and operations reports takes significant amounts of time and effort. Often these reports are out of date by the time they make it to the key decision makers. Making business decisions on old information prevents an organization from being agile and responsive in changing market conditions.

Consider that Maintenance history + Current operating conditions + financial implications = ability to intelligently schedule maintenance. Does it make sense to run this component to failure or does it require scheduled downtime for repair? Consolidating disparate data into one tool allows this type of analysis to be accomplished quickly and effectively.

Reducing the number of tools needed and enabling users to analyze data from multiple sources allows decisions to be made quickly and more accurately. Having access to accurate and timely reports allows facilities and organizations to respond faster to external market forces. Companies that consolidate their information systems into a single environment also reduce administration costs and license fees.

Figure 5 illustrates how consolidating plant data into a web visualization application allows multiple thin-client applications from many groups to access the data in a consistent and controlled manner. Have your experts analyzing data and making decisions on timely information, not searching for data and relying on rules of thumb or fudge factors to come up with mission- critical decisions.

(6)

Using existing infrastructure reduces the cost of implementation and allows for easier integration into existing business practices.

Products like Matrikon’s Operational Insight are good examples of software solutions that help organizations overcome their data disparity and take advantage of the information hidden within their facilities.

KPI and Metrics

All organizations use metrics and key performance indicators (KPIs) to do business and remain competitive. A decision-support portal presents this key information to the right audience.

Translating business objectives into measurable statistics requires industry specialists and domain experts. Producing the results takes instrumentation, business processes and software.

When implementing a decision-support system the equations behind the numbers must be validated and made transparent to all stakeholders. Standardizing the way performance is measured removes finger-pointing and shoulder-shrugging and brings accountability and traceability to the system.

Automating data collection and analysis frees up your domain experts to act on the information instead of chasing the numbers down and verifying their validity. Ensure the system has processes in place to verify the raw data being used in the metrics. Decision makers need to have faith in the information being presented to them. A successful data visualization system provides user-specific information. Displaying the appropriate information in context to key decision makers will allow them to make timely decisions and monitor progress.

A history of metrics and KPIs is necessary to provide concrete evidence of improvement and can be used to quantify ROI on efficiency and productivity initiatives.

KPIs and metrics provide a consistent message and quantifiable goals. Having the entire

organization on board and understanding their individual impact on the numbers aids in morale and accountability.

Visualization

The problem in today’s facilities is too many single-use tools. One tool is used to trend the raw process data, one tool to review lab results, one tool to look up maintenance details, and one tool to view alarm information. The bulk of a person’s day is spent finding the relevant data and manually correlating them into static reports that are already out of date. This non-value-added time would be better spent making decisions on actionable information.

Display the right information to the right people. Visualize information in context with user-centric displays. Having the most recent data correlated, aggregated and available at any time allows for better decisions that are more relevant to the current conditions. Figure 6 is an example of how to organize appropriate information for key decision makers. KPIs give insight into the operating status of a plant while the appropriate visualization tools provide the ability to further investigate the information behind the KPIs. This functionality in one common interface speeds up the analysis of issues affecting the operations of a facility.

(7)

Figure 6: KPI Dashboard

Use advanced display techniques to highlight key issues. Sorting through record after record or displaying information in its raw form can make it difficult to isolate the largest issues at a facility. The design of the interface should bring attention to the most important assets or problems immediately. Identifying the assets that have the largest impact on profitability and weighting the data aggregation appropriately can allow users to focus on the key issues.

Figure 7 showcases the advantages of using a technology such as treemapping to visualize the Relative Performance Index and saturation of the control loops in a pulp and paper plant. Large red boxes immediately draw the attention of an engineer to a trouble area in the facility. Historically an engineer would need to analyze large amounts of operating data to find issues such as these.

(8)

Figure 7: Advanced visualization

Having various tools has historically meant many ways of interacting and displaying data. This is a costly problem, as every user requires training on how to interact with the system and requires several sources of support. Having one consistent interface allows for lower support and training costs. Allow experts to create content and collaborate on key issues by providing an easy-to-use tool that consolidates all the data and information needed to improve productivity and profitability. When designing displays, ensure there is consistent layout and functionality to all tools regardless of where the information is coming from.

Decision support

A decision-support portal facilitates getting the right information to the right people at the right time. Notifying maintenance of an imminent equipment failure can save immense amounts of money. For example, a large power company was able to identify an issue in their turbine four weeks before a bearing failure would have occurred. This one event saved the company over one million dollars in lost production, as they were able to schedule maintenance at the appropriate time. Identifying key underperforming areas in the facility can reap large returns. Focusing the right personnel on the right information at the right time is the goal of any decision-support portal.

(9)

Use roll-up metrics to focus on key areas. That is, consolidate information into a high-level

indication of the general status of an asset. Group associated rolled-up metrics together to provide top-level displays that will highlight areas in the facility or enterprise that require attention. Figure 8 is an example of a corporate-wide, top-level display. It is showing the current health of key assets across the enterprise in real time.

Figure 8: Corporate monitoring

The ability to “drill in” to the area of concern will bring up a more detailed breakdown of the information creating the metric. Weight the data used to ensure mission critical information is not glossed over by other information used in the metric. This approach allows experts to monitor large areas and ensure they are focusing on the key issues affecting performance.

A decision-support portal encourages the sharing of domain knowledge and concepts allowing experts to create content that is shared among peers. This will capture their knowledge in a public space as opposed to Excel sheets trapped away on hard drives.

To encourage adoption of a new system, ensure end users are involved early on in the design phase. The people who will use the system know what information they need to see and will have valuable input on the ways to present the information. Having “Champions” involved during the validation stages reinforces the user communities’ trust in the information. Exposing a test audience to the system allows for genuine feedback and reduces the risk of frustrating and alienating the final user group.

(10)

Conclusion

The proper implementation of a decision-support portal will reduce time spent looking for and manually manipulating data while increasing the time spent analyzing actionable information. The proper system will increase collaboration and knowledge capture while reducing total cost of ownership.

Increase the time your people are spending on higher priority issues. Get the right data to the right people, at the right time!

About the Author

Grenn Holden has five years experience designing and implementing data visualization and decision-support platforms for companies in many process industries.

About Matrikon

Matrikon Inc. is a publicly traded (TSX:MTK) software company that enables operational excellence for the process industries. The Matrikon Suite of products addresses specific operational excellence challenges in the domains of Production Management, Asset Effectiveness and Operations

Optimization. Typical applications range from connectivity to alarm management to downtime reduction to production accounting.

Matrikon has long been committed to innovation and low cost of ownership. In 1998 Matrikon was among the first to present process data on the web, and charter membership in the OPC

Foundation underscores our commitment to open standards. Today, the Matrikon Suite of products transforms operational data into actionable information using Matrikon’s foundation of universal connectivity, physical object models and web visualization.

(11)

Matrikon™ Operational Insight

Securely access all your data, regardless of location or format Seamlessly deliver information anywhere, anytime

Create and monitor enterprise KPIs for decision support Enforce consistent operational and business processes •

• • •

(12)

Challenges

The ability to make informed operational decisions is possible only if one has access to all the required operational data - not just part of it. Without access to the whole operational picture, decisions become guesses that result in decreased profitability, quality issues and poor performance. Three key problem-areas prevent improved decision support:

1. Not having timely and easy access to operational data — Avoiding production and equipment losses is possible only if personnel can respond to events in a timely fashion. When data access is impeded by data format and accessibility issues, numerous and varied thick-client applications and poor internal processes, data collaboration becomes a slow, burdensome, and in some cases impossible process. The result: data may no longer be relevant and productivity can be compromised.

Challenges

2. Varied and numerous data systems — Existing plant IT/data systems can be difficult to tie together because of age discrepancies, different vendors, multiple interfaces and applications, and differing data types. Data collaboration is resource-intensive, and this high cost of ownership means lower user adoption. 3. Too many people doing manual work — Without access to collaborative data, key people must spend time physically going to get data to create needed information via manual reports. This time-consuming, error-prone process results in personnel resources being inefficiently used and decisions made based on outdated data.

Matrikon operational insight -

Web-based plant data Visualization

Matrikon Operational Insight is a web-based data-visualization and KPI-dashboard product that enables better operational decisions by providing the right people with the data they need, when they need it and in the format they require. Using standards-based technology, Matrikon Operational Insight connects all data sources. Coupled with a robust and configurable (vs. programmable) dashboard interface, data has never been easier to trend, chart, monitor, report and display.

“Matrikon Operational Insight has allowed AmerenUE to realize productivity and

efficiency gains by allowing real-time, web-based access to information from our

plant’s various operating systems. Matrikon’s service and support have been excellent

and the delivered product has provided outstanding value to AmerenUE”.

Tim Lafser

Engineering Supervisor AmerenUE

(13)

deliVer the right inforMation to the right

people at the right tiMe and in the forMat they

require, regardless of loCation

Improve decision-making

Get more out of existing resources Increase business agility

Empower personnel Improve workflow

Matrikon Operational Insight is for plant personnel who need fast, reliable, easy access to real-time plant data. Matrikon Operational Insight is a thin-client,

out-of-the-box software solution that provides data integration, information (Key Performance Indicators) and visualization from existing, vendor-independent IT/data system investments.

Unlike vendor dependent, thick-client or point solutions, Matrikon Operational Insight uses standards-based technology to connect to all data sources. Coupled with a robust and configurable dashboard interface, data has never been easier to trend, chart, monitor, report, share and display.

Matrikon Operational Insight means informed decision-making Use one universal client tool across your entire enterprise

Access data to provide a digital dashboard

Create user-specific views to instantly display relevant information

Manipulate and display data in any format to create actionable information Integrate plant floor and business applications

Create an environment that promotes simple and fast collaboration within and across departments

Create complex KPIs and visualize the information in minutes.

Matrikon Operational Insight means getting the most out of existing and future assets

Export data to productivity tools like word processing and spreadsheet applications Enable the capture of data through web-based data entry forms

Eliminate multiple application-specific thick clients and save on purchase and maintenance costs

Access any data bi-directionally

Create platform-independent data access Reduce IT infrastructure requirements and costs • • • • • • • • • • • • • • • • • •

(14)

Powerful Trending Monitor plant condition and analyze incidents with this advanced trending tool. Intuitive interfaces make advanced functionality like multi-axis and multi-time-range fast and easy to

use. Non-technical, wizard-driven configuration allows users to login and start investigating issues in seconds, with data from any source at any time.

Plot constants Perform fill plots Trend SQL data

Use advanced functions Graphical Presentation Get actionable information from real-time plant visualization and Key Performance Indicators (KPIs). Robust calculation functionality, advanced animations, and drag–and-drop

creation means you can easily build dashboards and overviews to present the information you need in a way that is clear, fast and effective.

rich symbol library

picture-in-picture functionality contextualized overviews • • • • • • • Ad-hoc Reporting

Access your relational data with a non-technical drag-and-drop interface and wizard configuration and give your SQL experts a break. Data filtering and parameter-based queries enable interactive web-parameter-based reports. Users can easily

hyperlink reports, enabling drill-down views and providing the ability to create master-detail reporting functionality. Charting

Build comprehensive charts to visualize relational or tag-based data. Use report templates to speed the generation of multiple chart types or use the intuitive query-building interface to translate complex queries into powerful visual representations of plant data.

Consolidate, integrate, and Visualize With

Matrikon operational insight

Trending and analysis

User defined KPIs and dashboard

Users can create interactive reports without learning SQL

Get fast information from automated charting

(15)

Excel Tools

Generate fast, live data reports using Microsoft’s Excel application. Matrikon Operational Insight integrates with Excel’s menu bar,

giving users instant and seamless access to all plant data. Expedite data cleansing and root cause analysis by enabling users to get all plant data into this familiar and powerful spreadsheet tool. Calculation

Eliminate difficult, time-consuming and frustrating calculations from combined data generated from different systems. Create your own calculations graphically by accessing data from any of your configured relational databases, historians or real-time data sources. KPI visualization

Key Performance Indicator visualization is a key component of any effective management practice. Build custom or executive dashboards using Matrikon Operational

Insight’s extensive gauge and symbol library.

Event notification

Create system events based on real-time and historical data inputs with this thin-client product. Create scripted events, such as sending an email or any other task, with triggers

based on real-time or historical data. The event configuration wizard provides easy paging, email and text messaging. Create the rules to drive the events, as frequently as every 30 seconds.

Automated Report Generation and Delivery Create, send

and save reports at any user-defined interval. A wide range of report formats are supported, including Word, Excel, PDF or web page. These reports may be automatically sent

via email, printed on a network printer, saved to a file or generated upon request. Consistent report generation means accurate and timely information is delivered to relevant departments every time, on time.

Integrate Microsoft’s Excel toolset

Easily create complex calculations and display the results in a meaningful and easy-to-interpret context

Build custom dashboards and view plant Key Performance Indicators

Custom event-driven notification

(16)

“Qatar Petroleum chose Matrikon Operational Insight as our web-based decision support

system because it allowed us to retain our existing and proven business practices while

enabling them through open connectivity, accessibility and easy usability. By sharing and

distributing information between the business processes and providing this information in

role-specific views, efficiency was greatly improved.”

Sultan Al-Abdulla

Head of Maintenance Engineering, Dukhan Facility Qatar Petroleum

Matrikon operational insight

a global solution

Matrikon Operational Insight is found around the globe, in use by some of the largest companies in the world including: Ameren BASF Bayer BHP Billiton BOC BP Chevron Texaco ConocoPhillips Detroit Edison Dow Chemicals DuPont Dynegy Eastman Honeywell Huntsman Imperial Oil Kinder Morgan Pfizer Inc Rio Tinto Rohm and Haas Saudi Aramco Shell Statoil Sunoco Syncrude Tesoro Total

(17)

industry-leading resourCes

Matrikon Operational Insight Lite - Free Download

View, trend and analyze all your OPC data quickly, easily and for FREE!

Collect and trend your real-time or archived OPC data with Matrikon Operational Insight Lite’s powerful, easy-to-use tools. Use your PC to connect to multiple and unlimited OPC servers to build your own powerful trends, export to Excel and other formats, build process graphics with wizards, access the rich symbol library with over 3000 high-quality components and create a digital dashboard.

This robust data visualization tool is easy to deploy and quick to install, complete with easy-to-follow video tutorials.

Get this Free tool at www.matrikon.com

industry leading resourCes

Exclusive Interactive Webcasts

Matrikon’s interactive webcasts are designed to inform you about innovative approaches to optimizing your operations. These presentations give you access to industry experts, broadcasting live to your personal computer or boardroom.

Matrikon also offers private webcasts for your company, allowing our experts to address the unique problems you face and discuss solutions to meet your specific needs.

Register for exclusive webcasts at www.matrikon.com

Sultan Al-Abdulla

Head of Maintenance Engineering, Dukhan Facility Qatar Petroleum

Dr. Julian Pickering Senior Control Systems Consultant

BP Exploration

“Matrikon Operational Insight provides BP Exploration with

real-time enterprise-wide performance visibility to our

operations. This collaboration environment allows greater

interaction between functional groups throughout our

organization, improving business efficiencies and providing

actionable information.”

(18)

Phone: 1-780-448-1010 Fax: 1-780-448-9191 Printed in Canada 0607-P Phone: +61 (2) 4960-1000 Fax: +61 (2) 4960-1212 Phone: +44 (1224) 659-800 Fax: +44 (1224) 659-801

[email protected]

Americas

Edmonton, Alberta Calgary, Alberta Fort McMurray, Alberta Vancouver, British Columbia Toronto, Ontario

Quebec City, Quebec Cleveland, Ohio Houston, Texas St. Louis, Missouri

Asia-Pacific

Brisbane, Queensland Gladstone, Queensland Newcastle, New South Wales Melbourne, Victoria Perth, Western Australia Hamilton, New Zealand Singapore

EMEA

Aberdeen, Scotland Barcelona, Spain KÖln, Germany Matlock, United Kingdom Riffa, Al-Hidd

References

Related documents

When the zero-profit condition for new vacancies is combined with the Nash wage equation it gives an equilibrium wage rate and job creation rate that depend on the frictions

pricing scheme, the characteristics and choice behavior of providers and customers, the travel demand pattern, and the geometry of the network. > The larger the percentage

• It is not possible to compare the methods with measured input data and noon report input data due to lack of data of noon reports for – in the best case 9 noon reports to

This paper was presented at the PIONEER OF THE MILLENNIUM DEVELOPMENT GOALS: ATATÜRK 3   ATATURK’S REFORMS EMPOWERED TURKISH

The Ministry of Justice, with a view to introducing new Turkish Penal Code (Law No.5237) which came into force on 1 July 2005 held seminars on “the Introduction of New Penal

characteristics and ―extra‖ skills that industry would like to see in today's - and future- graduating engineers and scientists... “Leadership” and

Accordingly, we propose an unsupervised backward elimination feature selection algorithm BAHSIC-AD, utilizing Hilbert-Schmidt Independence Critirion (HSIC) in iden- tifying the

[r]