Weaving enterprise optimization from the plant floor to glass house and through the supply chain requires sorting through difficulties, weighing options, and eliminating confusion. Related solutions and user and vendor recommendations are included in AMR Research's (AMR, Boston, Mass.) Oct. '99 Report On Manufacturing titled "Variability: The Cure is Out There," written by AMR's research director of manufacturing strategies, Roddy Martin.

The AMR Report correctly identifies that most Enterprise Resource Planning (ERP) solutions grew out of transaction-oriented discrete manufacturing applications, where material lists and routings are fixed for a product and yield and component costs have fixed relationships. Discrete manufacturing does not permit assembly of 1.5 vehicles from components assembled for one vehicle, but in process manufacturing it is possible to obtain different yields from the same quantity of raw materials and resource efforts.

Sources of process yield variability include raw material potency characteristics, inefficient equipment, intermediate product shelf life, and even seasonal weather conditions. Recognition of a need to include variability tracking and analysis as an integral part of an enterprise integration solution is the emphasis of the AMR report.

When a list of what's required to meet today's customer requirements is created, quality is always near the top. Yet capturing, analyzing, and determining root causes of variability and cross-functional implications are beyond the scope of today's ERP solutions. What's available is an ERP solution that helps administer, forecast, plan, schedule, and reschedule, but doesn't help determine the root causes that initiated the need to reschedule and more importantly how to permanently eliminate those causes.

Mr. Martin says, "Total Quality Management (TQM) and Total Production Maintenance (TPM) initiatives engrained on most plant floors are not yet a part of the overall ERP solution space. ERP and Manufacturing Execution System (MES) vendors have generally remained at arm length from one another with ERP vendors downplaying the importance of MES applications. In some cases ERP vendors claim ERP solutions deliver the needed functionality, but users increasingly find this not always the case."

AMR's report cites an example where a beverage manufacturer implemented a popular ERP application but discovered the time to obtain variation reports took substantially longer than in pre-ERP days. AMR's report also explains how a brewing company has successfully provided short cycle production feedback to a discrete-oriented ERP solution to ensure responsiveness to production variabilities.

Dennis Brandl, enterprise initiative director with Sequencia Corp. (Scottsdale, Ariz.) and member of ISA's SP95 Enterprise/Control Integration standards committee says, "Process manufacturing companies are rediscovering that MES functionality must reside at the plant level but above the control and automation system. Traditional MES functions include production:

• Schedule modifications;
• Costs optimization;
• Asset and resource availability information;
• Material usage information; and
• Quality and process control statistics.

SP95 committee activities have advanced from what information to communicate to/from the control domain to how information will be communicated. Exchange protocols being considered by the committee include SQL (Structured Query Language), XML (eXtensible Markup Language), and IDL (Interactive Data Language) object definitions."

It is adoption and use of standards and de facto standards, like XML, Microsoft's BizTalk, SQL, SP95, OPC (OLE for process control), Microsoft SQL Server, DCOM (Distributed Communication Object Module), Java, ActiveX, and Visual Basic, that permit avoiding a repeat of built-from-scratch point solution shortcomings (see CE, Nov. '99, p. 25). Today's solutions start with a popular database, use a familiar graphical user interface, include pre-tested object modules, and follow structured development rules. The result is Built-For-Purpose (BFP) solutions designed to identify and remove variability yet easily integrate with other business applications.

For example, Intellution (Norwood, Mass.) offers VisiconX, a set of ActiveX controls that enable users to point and click to develop and execute unique SQL queries against disparate relational database and use the collected data to populate business reports, all without understanding SQL.

Bradley Ward Systems (Atlanta, Ga.) is focused on BFP solutions designed to meet specific plant-floor operational needs for food processing. Included in Bradley Ward's Success suite is Hazard Analysis Critical Control Point planning and execution modules designed to meet USDA-FSIS (U.S. Drug Administration-Food Safety Inspection Service) regulations.

Hilco Technologies (East City, Mo.) Real-time Production Management (rtPM), and Sequencia's Open Batch, Procedure Manager,and Production Manager are all BFP solutions for batch processes designed around the ANSI/ISA S88 batch standard models. Hilco's rtPM provides MES functionality including material receiving and tracking, recipe scheduling and execution, and quality assurance modules. Sequencia's Production Managerprovides synchronization of plant floor and business systems data (i.e., SAP R/3 software) without programming and is built around Microsoft's Distributed Network Architecture for Manufacturing.

ABB's (Wickliffe, O.) Advant Enterprise Historian and Honeywell's (Phoenix. Ariz.) Business Hiway are BFP solutions designed to integrate process industry operating domains with the PP-PI (Production Planning for Process Industries), PM (Plant Maintenance), and other application modules of SAP's R/3 inter-enterprise business solutions.

Foxboro's (Foxboro, Mass.) enterprise integration comes in the form of I/A Series Information Suitedesigned to integrate software applications and components from Foxboro, Simulation Sciences, and Wonderware, with leading ERP systems.

Improving things like raw material consistency and better use of BFP applications will improve planning and scheduling accuracy, but to become really good, to achieve "world class" capability, requires integrating process equipment health and status into the planning and scheduling equation.

Supply-chain management is all about reducing inventory and costs, but when yield variability in process manufacturing is not well managed, the solution is to maintain a "cushion" inventory of finished product—the most costly of solutions.

Jack Welch, ceo of General Electric (GE, Fairfield, Conn.), says, "An organization's ability to learn, and translate that learning into action rapidly, is the ultimate competitive advantage."

At companies like Allied Signal Inc. (Morristown, N.J.), Motorola Inc. (Schaumburg, Ill.), and GE that statement is put into practice daily with analysis software from companies like SAS (Cary, N.C.). Such software aims to gain competitive advantage for its users by removing variability (see CE, Jan. '99, p. 62, and Mar. '99, p. 87).

When first-, second-, and third-tier supply chain vendors are added to the variability complexion, many manufacturers find it necessary to educate and support supplier efforts to remove incoming variability in order to eliminate outbound cushion inventory. Thus the overwhelming need for the enterprise information network is to be able to capture, analyze, and identify root causes of plant-floor variability in real-time.

Key to determining root cause is the ability to measure critical parameters and events, and it's not always obvious things that create variability. For example, one company's continuous improvement team optimized a process only to find a few days later that variability had returned. Using MDT Software's (Alpharetta, Ga.) MASS AutoSave, root-cause analysisrevealed unauthorized and undocumented programmable logic controller software changes occurred during maintenance activities.

Determining root causes often require detective-like persistence, but tools and training such as those available from System Improvements Inc. (Knoxville, Tenn.) improve efficiency and effectiveness by prompting investigators through a consistent set of questions and categories. Responses are recorded in a searchable relational database to aid in identifying recurrences or similar root-cause events. When used proactively, root-cause analysis can identify potential production interruptions caused by equipment failures as well as reveal undiscovered production improvement opportunities (see CE, Dec. '99, p. 25, "Managing risk improves production").

It is mystifying that just before a two-week scheduled maintenance shutdown, many process plants are run their "hardest" in an effort to meet customer requirements and build inventory cushion. It is possible to conclude that either the equipment isn't really in bad enough condition to justify a maintenance shutdown, or heightened focus on reducing processing variability is indicating the processes real yield capability. Regardless, equipment health, status, and capability are key elements in achieving advanced planning and scheduling success.

Knowing the ongoing health of equipment requires establishment of a Reliability Centric Maintenance (RCM) organization (see CE, Jul. '99, p. 46). Being able to use equipment health, status, and capability in planning and scheduling scenarios requires another form of BFP in the form of Enterprise Asset Management (EAM) or Computerized Maintenance Management System (CMMS) software from companies like Indus International (San Francisco, Calif.), PSDI (Bedford, Mass), and Mincom (Norcross, Ga.). When RCM is working at it's best, and EAM/CMMS applications are integrated, supply-chain optimization can be further extended to include equipment spare parts inventory and contract maintenance services.

Identifying and attacking sources of variability to achieve world-class enterprise integration is not an overnight accomplishment, and likely not available as a single-source solution.

Help identifying, evaluating, and choosing software that's right for you is available from companies like Expert Buying Systems (Las Vegas, Nev.) with its ChooseSmartproduct and services.

If all this is not enough complexity there is yet another piece of the integration puzzle. The enabling tool for gathering and sharing data to optimize every asset is an enterprise information network, a sort of digital nervous system.

Since it is impossible to foresee enterprise information requirements several years in advance, a key requirement is for the digital nervous system to be "future proof" (see CE, Jun. '99, p. 46).

When considering a digital nervous system networking infrastructure, three main components must be evaluated:

• Physical;
• Transport mechanism; and
• Application level.

The physical medium provides means to move data among managing devices. Widely known and widely available is Ethernet cabling. This ubiquitous collection of twisted-pairs of wire can be found in virtually any business today. The magnitude of Ethernet-related connectivity devices dwarfs any other physical networking medium today. Furthermore, the Ethernet knowledge base within most organizations far surpasses that of any proprietary industrial network.

How data are transported on the wire, or wirelessly, is the second component to evaluate. This really doesn't have much to do with the data itself, but is more about how data move from one device to another over a selected medium. Clearly, Transport Control Protocol over Internet Protocol (TCP/IP) is the undisputed global leader. The beauty of TCP/IP is it's ability to bind to many types of physical media, from wire to wireless, LAN to WAN, and on the other end, to encapsulate virtually any type of data. TCP/IP then becomes the real enterprise integration enabler. Because it can adapt to the physical and application layer in virtually any way, its use is nearly limitless. This leaves the application layer, arguably the most contentious among the industrial automation community, to evaluate.

Most Information Technology (IT) organizations rely on Ethernet as the communication network of choice because it provides an interoperable method to share data from one device to another. That doesn't mean all devices speak the same application language. For example, SMTP (Simple Mail Transfer Protocol) was designed to communicate e-mail.

Does it make sense to burden SMTP with the ability to transfer files among devices? Probably not. That's what FTP (File Transfer Protocol) is designed to do. And FTP works alongside SMTP and a rather large number of other task-specific protocols. Applying the scenario to the plant floor, products such as Zoneworx's new (Temecula, Calif) Pathways and Opto 22's (Temecula, Calif.) SNAP I/O, communicate multiple protocols, such as Modbus TCP, HTTP, ActiveX, OPC, and others, over Ethernet.

To expect one gargantuan protocol to cover all possible data types across an entire enterprise is ridiculous. This is where the plant floor could benefit by borrowing a page from the IT side of the business (see CE, Dec. '99, p. 33 and p. 45).

Regardless of how it's achieved, making plant-floor assets an integral part of the enterprise's digital nervous systemis a "must do" to eliminate variability and improve responsiveness to customers. Waiting for another new standard or "promising" new technology introduces one more source of variability, and who needs that?

A best-of-breed solution contains ERP, MES, and control and automation products that make the best use of open standards to link business processes and best fit your company. Choosing just such products and technologies is the best way to use what's best today and be able to change to what's best in the future.