The Plant Information (PI) System

Operating a process manufacturing plant is complex. Companies are constantly challenged to operate at an acceptable profit in the face of changing business conditions and increasing global competition. Engineering and operations management makes critical operating decisions on a daily basis that affects profitability plant efficiency and safety. These decisions include; scheduling product mix to match pricing and demand, configuring equipment for highest productivity, scheduling equipment outages, planning equipment replacements or upgrades, and conforming to environmental and safety standards. A key competitive advantage in making these decisions is to have a clear knowledge of plant operations at the fingertips of plant personnel. This enables the operations of the entire plant to be optimized relative to the equipment capabilities and chemistry of the processes.

A complete history of all plant variables is a valuable asset in achieving optimum performance in a manufacturing operation. Since it is difficult to know in advance which plant measurements will be needed for future analysis, storing everything is the best method to safeguard against not having the required information. With knowledge of the current and historical state of operations, many potential problems can be diagnosed and corrected before they have a negative effect on plant operations. People can use the information as a road map to process improvements, reduced raw materials consumption, increased production and better safety.

A plant information infrastructure is needed that serves the plant process data up to the Microsoft desktop for operations, engineering and management. This infrastructure must also serve data to software applications that require high-resolution real time and historical plant process data to analyze plant performance.

The PI System - An Information System designed for the Manufacturing Environment

The Plant Information (PI) System from OSI Software, Inc. is a set of software modules designed for plant-wide monitoring and analysis. The PI System Data Archive is the foundation of this system. It handles the collection, storage, and retrieval of numerical and string data. It also acts as a data server for Microsoft Windows-based client applications. Operators, engineers, managers, and other plant personnel use these client applications to view the plant data stored in the PI System Data Archive.

The PI data archive accommodates very large real-time and historical databases typically sized so that every recorded process point (called a "Tag") is stored on-line near its original resolution for years. The Microsoft-based PI client applications enable personnel to easily access this high-resolution data to view a plant's current condition while providing a very clear and accurate picture of past operations. Data access to and from the PI Data Archive is extremely fast. Users can retrieve the information they need within a few seconds regardless of the number of tags or the size of the archive.

PI data can be used for plant and corporate initiatives such as process improvement, total quality control, and predictive maintenance. The unified plant data repository ensures that all individuals view and analyze the same data. Software applications such as maintenance management, expert systems, LIMS, and optimization/modeling programs can use PI to gain access to real time data. The PI System can link real time control systems to production planning systems and enterprise resource planning systems to bridge the gap between business and real-time production environments.

OSI has put a premium on functionality, reliability, performance, and maintainability in the development of the PI System products. The PI System is designed to be easy to install, maintain and use. Inexpensive migration paths that preserve the existing plant historical data are provided to update legacy PI systems to the Windows NT based technology. New features are extensively tested for compatibility with older versions.

Benefits of the Plant Information (PI) System

Lower Production Costs: The PI System's data analysis and graphical tools permit problems to be resolved faster, and often prevent small conditions from becoming major upsets. Many companies report that the data stored in PI allow them to improve their preventative maintenance programs, increasing equipment life. These uses result in increased production and lower operation costs. In addition, the PI System contributes to yield improvements and less waste. When the lab system sends massages to operation via the PI System, for example, the time required to change grades decreases, increasing production and minimizing waste.

Higher Product Quality: The time to respond to a quality problem is before the product goes out of spec. Applications such as the PI-Profiles package allow inconsistencies to be noticed and fixed before they become serious problems. Operators also appreciate the PI System because they can review operation conditions from previous runs, a valued feature when producing unusual or difficult product grades.

The PI System provides an easy-to-use mechanism for obtaining and monitoring continuous process improvement. These programs frequently lead to improved quality with no increases in production costs.

Better Customer Service: With access to a PI System, customer service and quality assurance groups can quickly respond to customer complaints and concerns about product quality. This is because process data are stored at high resolution and remain online for extended periods of time. Personnel are also able to analyze past data to determine whether requests for new grades or products specs are feasible. And, because current data are also available, sales teams can review process conditions to determine when a new order cam be scheduled or delivered.

Increased Productivity and Improved Process Knowledge: Less time is wasted trying to obtain data and information relating to a problem, so more time is available for solving it. Everyone throughout the facility uses a common set of data, eliminating discussions of whose data set is "right", again contributing to a more productive use of everyone's time. Also, application developed internally with the PI System, such as performance monitoring and efficiency calculations may be shared between sites, reducing development time.

What does the PI System do?

I. Interfaces to Highly Complex Manufacturing Environments

PI provides data pipes from hundreds of different manufacturing automation devices to bring all of the plant operating data into a common data format in a time series database. Real-time and historical data is available from this database to the entire corporation via the Microsoft desktop.

Major automation vendors with standard interfaces to PI include:

ABB/Bailey, Emerson Electric (Fisher-Rosemount, Intellution), General Electric, GSE Systems Honeywell (TPS, TDC3000, Measurex), Moore Products, Rockwell Automation, Siebe Intelligent Automation (Foxboro, Wonderware), Schneider Automation (Modicon, Square D), Siemens, Toshiba, Valmet, Yokogawa and Westinghouse.

II. PI Provides Real Time and Historical Data Access

At the heart of the PI System is the PI Data Archive, a real-time database optimized for storing and retrieving time series data. This is where the plant process information, i.e. pressures, flows, temperatures, setpoints, on/offs, etc. are stored. The major design features of PI Data Archive are listed below.

Captures all data related to operations or production. By capturing all process data in a single repository, PI can create an accurate picture of current and past plant operations. All users can access the same information but with different views and perspectives. A process engineer can quickly analyze current process performance. A maintenance engineer can view historical data looking for degradation in equipment performance.

Stores data on-line long term. Several years worth of process data can be available within a few seconds of the request by the user or application. Users can pick up seasonal variations in the process, analyze equipment run times, and view cycles of production. PI utilities are available to archive off-line any old data that is no longer routinely used.

Stores data only once. Because the PI system stores data in its fundamental form, the data can be used for different purposes without any data discrepancy. Users can query and view current, historical, or statistical data with PI client software. Users and applications can request the data to be calculated and delivered in many formats including summarized or "aggregate" data. This eliminates the need to decide ahead of time how to summarize the information for reports or for data analysis.

Stores data efficiently. Traditionally companies heavily summarized data to reduce the amount of disk storage required. The PI Data Archive uses the "Swinging Door" compression algorithm to store information for thousands of points to their original time resolution, without requiring vast amounts of disk storage.

Stores data to its original resolution. The PI Data Archive is designed to store process and event information without loss of time resolution. Data is collected and stored as a function of its fundamental accuracy and time resolution. For example, if a process variable is capable of moving very quickly, data for that point is stored at a high time resolution. The "Swinging Door" compression algorithm ensures that data retrieved from the Data Archive is always represented within the accuracy specified for each point.

III. Provides Data Transport across LANs / WANs / Internet

PI provides desktop access to anyone in the corporation needing access to plant data via client/server or intranet browsing.

IV. Provides a Development Environment for Integrating ERP with the Plant Floor

Many major corporations have implemented ERP systems for integrated financial, manufacturing, sales, distribution and human resource support. To get the maximum benefit from ERP, they need to close the loop between the operating decisions made using ERP and the plant floor operations. The PI System with it's extensive library of gateways to plant floor automation is the ideal middle tier software to integrate the factory floor to major ERP vendors such as SAP, Baan, J.D Edwards and PeopleSoft.

PI as the Middle Tier between SAP and the Plant Floor: <%=company%> has developed a software product RLink as the new gateway from the PI System to PP-PI, PM and QM modules in the SAP R/3 system. The RLink gateway collects data from real-time systems on the plant floor and translates these data values into information expected by R/3 modules. RLINK supports transferring recipes, specifications, and maintenance orders back to the plant floor systems.

V. PI Provides Information Support for the Knowledge Workers

Manufacturing companies can improve performance by empowering all of the employees in the organization to become "Knowledge Workers". Employees become "Knowledge Workers" when they have the data they need to do their jobs and the tools to analyze that data at their fingertips. With timely precise information the Knowledge Workers can continuously work to improve operations and respond to planned and unplanned events. To accomplish this companies need to accumulate, integrate, share and access data both throughout the plant (desktop) and across their organizations (internet/intranet). The technologies used to provide this infrastructure must work with the company's existing manufacturing automation, databases, applications, etc.

VI. Provides Data Reconciliation / Data Scrubbing with Sigmafine

OSI has recently added the Sigmafine data reconciliation and yield accounting package to its line of products. Applications such as financial systems, planning systems, and process improvement systems require " scrubbed" or reconciled data to execute effectively. The integration of Sigmafine within the PI system allows OSI to deliver a " one number system" to process plants across a range of industries.

Material Loss Accounting. Sigmafine reduces the uncertainty in material loss calculation by guiding you to possible areas of loss with far greater accuracy than is possible with unreconciled data. For instance, for a 100,000 BPD refinery, reducing loss uncertainty from 0.5% to 0.1% reveals where oil worth US$2,500,000 per year is being lost physically or through custody measurement errors.

Yield Accounting. Sigmafine gathers, analyzes and reports on the processing and movement of feed stocks and products. Production reporting and tracking of inventory enable the plant to have consistent information about the day's or month's actual operations. Sigmafine enables a comprehensive view of the accounting process for improvement of analysis and reports as well as productivity.

Data Reconciliation. Unlike rule-based yield accounting packages that normally assign measurement errors to one or two streams, Sigmafine's data reconciliation method is based on statistical least-squares method. This method cross-checks the measurement errors on all meter and gauge readings, according to the tolerance assigned to each measurement. An accurate set of data is produced that both the yield accountants and the engineers can use.

Custody Transfer. Sigmafine can be used to verify custody transfer measurements and determine an equitable distribution of components through the system. Custody transfer meter bias can be tracked over time to provide information for possible challenges.

Environmental Compliance. By increasing confidence that the plant is in compliance with environmental regulations and permits, Sigmafine allows you to operate your plant closer to its limits, and increase processing profitability.

Unit Optimization. Many plant operators use process models to guide strategy setting. These process models require material-balanced input data. Sigmafine assists in this process by generating mass, heat, and component balances around any process unit.

Instrument Management. The solvability analysis feature of Sigmafine allows the program to analyze each stream and check to see whether it is solvable and cross-checked. The solvability or non-solvability of each stream identifies the need for additional instrumentation. The cross-checked or non-cross-checked status of each stream indicates the level of confidence the plant can have in its measurement model. Reconciliation results also include sensitivities that identify the instruments with the most impact on data accuracy.

Instrument Maintenance. The reconciliation capability of Sigmafine helps engineers to simplify instrument maintenance scheduling, by prioritizing them with a list of instruments that are ranked by performance over time. In this way, just-in-time maintenance is achieved.

VII. Provides A Microsoft Development Environment for Client Applications

PI client products are designed to be fully compliant with the Windows95/ Windows98/WindowsNT architecture. ActiveX support (server and container), and ActiveX Automation are built into the products. Microsoft's new object oriented file structure is fully supported. PI-ProcessBook has earned the "Designed for Microsoft Windows95/WindowsNT" logo by passing tests administered by an independent lab for design compliance.

ActiveX, Object Linking and Embedding. Microsoft's ActiveX provides a framework for programs and objects within programs to communicate with each other. PI-ProcessBook provides full ActiveX support. This includes a container function that allows customers to embed or link "objects" from other applications into PI-ProcessBook and a server function which allows PI-ProcessBook entities to be served to or "embedded in" other applications.

Supporting ActiveX Container functions and Object Linking and Embedding (OLE) in PI-ProcessBook means that a user can embed or link objects (spreadsheets, documents, and third party controls) from other programs into a PI-PB display. The user can create a PI-ProcessBook display containing its standard objects such as trends, bar graphs, and dynamic values while embedding a spreadsheet within the same display. Embedded objects can be edited in place by merely double clicking on the object and they can be modified and updated without having to leave PI-ProcessBook. With linked objects, any time the source of the linked object is changed, those changes are immediately reflected in the PI-ProcessBook displays.

Supporting OLE means the user can embed PI-ProcessBook objects such as trends, dynamic values, bar graphs, etc. within other ActiveX compliant applications. For example, the user could embed a PI trend display in a Microsoft Word document. The trend would actually be linked to PI and would update, in real-time, within the Word document. The user can "click" on the embedded trend, pan and zoom through the trend, all within the Word document.

ActiveX Automation
ActiveX Automation provides a framework for communication between objects and the integration of programs that were not originally designed to interact. With ActiveX Automation, a program can provide public access to the methods of its objects allowing direct manipulation of objects by other programs. In spreadsheets, objects are cells and worksheets. In other applications, objects may be text boxes, drop-down list boxes, and radio buttons. In PI-ProcessBook, objects are displays, things that have been drawn on displays (lines, trends, etc.), data sets, and PI-ProcessBooks.

Visual Basic Scripting
ActiveX Automation is complete when you add a language to the objects and methods. Microsoft's Visual Basic is a language that can manipulate objects via ActiveX Automation methods. A user can develop a Visual Basic application that can manipulate PI-ProcessBook objects. For example a script may execute that prompts a user for inputs when a PI-ProcessBook display is opened or when a value is entered into a field. This lets the user create " what if" displays or validate manual inputs.

Visual Basic scripting is useful for reporting requirements. Cells or fields in the spreadsheet report may contain PI tag names or times when data is to be presented for the report. The user simply changes a cell that contains a tag name or time and the report updates automatically. Visual Basic scripting allows for a PI-ProcessBook trend to be manipulated by the spreadsheet such that the time and PI tags it is trending will change whenever a cell in the spreadsheet changes.

PI customers have used ActiveX Automation and VBA tools in PI-ProcessBook to add value with their PI Systems. The following examples are just a few of the benefits OSI had in mind when they adapted the Microsoft client-server development environment five years ago.

Data Validation. A large pharmaceutical manufacturer who is upgrading PI from a legacy VMS version to a Windows NT version in a validated environment has developed a VBA program to compare the converted PI data with data on the Distributed Control System (DCS).

Consolidating Multiple Sources of Data. A utility trading company uses PI to leverage everything from power plant operating data, to current energy prices, to weather data to make real-time decisions on how to allocate production and deliver electricity to customers in the most profitable way.

<%=company%>, 777 Davis Street, San Leandro, CA 94577. Tel: 510-297-5800 Fax: 510-357-8136.