Dairy production uses ISA-95 to integrate batch information with business requirements
By Fabian Yesid Vidal Lopez, Libardo Steven Muñoz Trochez, and Oscar Amaury Rojas Alvarado
If manufacturers want to achieve an integrated automation production process, it takes a melding of the process itself, the company’s organizational structure, the equipment, and management of the supplies. Yet the specification and the definition of these aspects become more complex due to economic policies and globalization. Now more than ever it is becoming more important to take advantage of the company’s business process information flow for the most productivity and competitive gains.
Companies devoted to manufacturing food products in particular have generated a great deal of information for business systems and production processes to guarantee fulfillment of quality standards established by legislation. Their goal is also to increase productivity levels and improve process performance.
To achieve this goal, manufacturers need effective communication among different information systems that generate and manage the data in different levels of the company. In the following case study, it is particularly essential to integrate the production program information to help communicate production requirements, which the business level establishes after analyzing product demand.
We are offering a practical vision of how to work with models established by the ISA-95 standard, which has been widely accepted worldwide for information integration among business and manufacturing systems in solving enterprise-control integration problems.
With ultra-high temperature in the heated-milk production area it is not only important to fulfill quantity and presentation requirements, it is just as important for the business level to know the results of the production, details of quantities, and types of product. Being aware of the different resources used will help management conduct an analysis to determine costs and establish actions for improvement.
The ISA-95 standard has a history of solving difficulties from information exchange between the business level and manufacturing. It has also enjoyed world-wide approval. Therefore, we presented requirements in this case study with recommendations and outlines established in the ISA-95 standard.
In integration projects where we use ISA-95, we found it necessary to conduct a modeling phase followed by an implementation phase using middleware to exchange the developed Business to Manufacturing Markup Language (B2MML) documents. We focused on the information modeling phase in applying ISA-95 and B2MML documents.
To achieve this goal, during the integration process, we modeled the functional structure of the company using a data-flow model. Subsequently, object models see use in the analysis and modeling of the material resources, equipment, personnel, and process segments information. Finally, we elaborated on the production and performance program documents in B2MML, which allow the exchange of information in the implementation phase.
Functional data flow model
The functional data flow model has been the most useful tool in understanding the structure of the company. It has allowed us to clearly establish in an organized way the operation of the company, detailing in every function the responsible people, the type of decisions they make, and the way in which each interacts to develop their activities.
In developing the functional data flow model, we collected information through interviews with people involved in quality control, maintenance, inventory, procurement, and production. The questions were designed to help us learn how those within the company carry out activities defined in Part 1 of ISA-95.
The result of this modeling is a document that clearly identifies which function the different departments, people, or information systems belong to. The document specifies if activities they perform correspond to activities in Levels 3 and 4. This document has been a great help for the company, adding more clarity about its business processes and its manufacturing operations, helping identify activities affecting production performance and business processes as well as information flow between them. The document has also been a great complement to the documentation established in the quality certification ISO 9000-9001.
The functional data flow model design is general; it is easy to use and implement, regardless of whether a medium or big company develops it, since all the possible activities and information flows in these companies can be established within the functions and data flows considered in the model. The functional data flow model is not strict in its total development. It allows companies to analyze functionalities and information flows according to their own requirements.
Developing object models has allowed a more organized approach to the implementation phase. We can structure information within four categories that come from the standard: production schedule, product definition, production capability, and production performance. We obtained models from three resource categories: materials, equipment, and process segments. Although the models are not meant for designing databases, they are a great help when defining main fields in the database that will store the resource information. So they facilitate the information exchange in the way the ISA-95 standard proposes.
Using the materials model during the production process, we included information about raw materials, medium products, by-products, and finished products. The materials model required identifying material resources considered raw material, medium products, and finished products. We defined identities and properties of each of the materials, grouped together each of the defined materials in general classes, and developed material definitions and material classes under the B2MML structure. In order to get the material model in the B2MML structure, we produced a document including the B2MML required information in a summary chart.
We organized information for equipment involved in the manufacturing process of the ultra-high temperature flavored milk using the equipment model, which required the following steps:
Refer to the equipment hierarchy model previously developed in order to identify and organize the equipment that intervenes in the manufacturing of the flavored milk.
Define the equipment class and their properties, grouping them according to similar characteristics.
Define the equipment with their IDs and properties.
Develop the equipment class and equipment with their properties in the B2MML structure.
After conducting an analysis based on the equipment hierarchy model, we discovered a process cell in the flavored milk manufacturing process is composed of six units.
Process segment model
After conducting an analysis of the ISA-95 standard, we found the definition of a process segment must identify all resources used in a common way in a production step for the manufacturing of different products. This does not include resource specifications required for a particular product since this information is detailed in the product segments. The process segment model must establish a general view of the manufacturing process for the business system, letting us know activities the manufacturing system can perform, the materials, equipment, and the personnel used in each of the production steps.
The case study’s flavored milk production system included six segments within the process cell, allowing reception, heating, addition of additives, standardization, ultra-high temperature heating, and packaging activities. Each of these activities constitutes a process segment in a global segment that represents the resources in the process cell.
In developing the process segments model, we identified process segments and existing dependencies between process segments, established routing for process segments, and specified material and equipment segments for each process segment, as well as process segments identified in B2MML structure.
Production schedule model
In the flavored milk process cell, parameters include the type of product and the quantity to be produced, the quantity of raw milk, the presentation (milliliters content in the package), the ID lot, and the date of finished production.
Materials information is only differentiated by the material lot specification defined in the production schedule. We obtained specifications of the equipment requirements using the equipment ID class to be flexible when assigning equipment.
In the production schedule, Level 4 (apart from indicating in Level 3 the production requirements through information in the segment requirement field) can also request important information be returned at the end of the production execution using the information in the segment response field. But the following questions arise: Are existing MES systems intelligent enough to identify the information to be returned? Or is it necessary to use an external application to carry out the identification of this information? For the case study, the production schedule only specifies the production requirements, and the required information is reported with a defined structure at the end of the production execution, which cannot be modified in a dynamic way using the production performance document.
ABOUT THE AUTHORS
Fabian Yesid Vidal Lopez is an industrial automation engineer with Omnicon Ltda., in Cali (Valle del Cauca), Colombia (firstname.lastname@example.org). Libardo Steven Muñoz Trochez is an industrial automation engineer with Omnicon Ltda. (email@example.com). Oscar Amaury Rojas Alvarado is an electronic and telecommunications engineer at the University of Cauca in Popayán (Cauca), Colombia (firstname.lastname@example.org). This is an edited version of a paper originally presented at WBF’s 2007 North American Conference in Baltimore, Md.
Better business with integrated batch
By Sandra Vann and Mike Williams
In the food and beverage and pharmaceutical industries, customers’ expectations are changing. It is not good enough to say your product meets a certain set of criteria; you must be able to prove it was produced according to those criteria. That is the challenge of integrating information flow in manufacturing systems today, especially with the constant merger of disparate IT platforms. A successful batch process will assemble all the information surrounding it into an integrated report. Because a batch process typically produces a consistent batch size, optimization requires focus on cycle-time reduction and not just mass. This event-based information is vital to identifying where process hold-ups occur and why.
A batch plant must be able to analyze cycle time as well as batch quantity produced, tracking every minute and every pound. The data must be available in near real-time (within an hour of batch completion) to capture reasons for not meeting target as close to the event as practical. You should be able to provide a framework to compare performance across similar production facilities globally as well as the ability to report data by process cell, unit, operation, and material produced. The more ways you can slice and dice the data, the more opportunities you have to identify correlations and trends.
Your batch platform must accommodate multiple plant data sources including proprietary process information systems, commercial information systems, as well as manual spreadsheets. It is important to provide a central location to collect, aggregate, and report performance data across a business and provide secure, role-based access to information over a corporate network.
Finally, a successful system will have the corporate information systems organization maintaining it with no need for a client resident application or local support.
Take a look at how web-based tools see use in the Measure, Analyze, Improve, and Control processes.
The plant operations personnel assigns reasons for the gaps between target and actual performance—at the end of a batch for a batch process and at the end of a 24-hour reporting period for a continuous plant. Different views can accommodate different needs, yet all views have access to a common cache of information representing a list of the batches made with the production quantity and time consumed for each operation.
Next, the plant process engineer reviews the performance data and verifies its accuracy. In some cases, there may be multiple reasons for a performance gap, therefore the gap goes into different reason categories based on investigative evidence. Typically, validation occurs one to four times monthly to capture input from not only operations but supporting functions such as maintenance, process research, or supply chain. The tool facilitates rapid assignment of corporate- and business-specific reason categories and equipment locations to expedite the reconciliation and validation processes.
The first step in analyzing batch cycle time: compare and contrast each process step versus the approved target times. The tool views facilitate the selection and sorting of performance data based on operational step-by-unit operation and by material produced to identify defects and common causes of failure. This provides histograms, trends, and averages of operation step times for detailed analysis.
The improvement engineer analyzes performance across multiple plants, which includes multiple process cells composed of multiple units. We group losses into 17 major categories displayed in a Pareto chart format. To identify common areas of interest, the improvement organization can view trends of performance loss by category over the last 24 months. Further detailed investigation occurs through additional drilling capability, which provides summaries by reason and a list of individual gaps rolled up to that reason.
After installing the improvement, track the value of the project through realization to confirm the return on effort and sustain the gains achieved through project implementation.
Asset Utilization, Asset Capability, and Asset Mechanical Reliability are key performance metrics that drive manufacturing excellence for us. To present a high-level view of overall manufacturing performance a key performance indicator dashboard displays at the business level. From this high-level view, business leadership can quickly survey Current Day/Month to Date/Year to Date for up to 24 months to evaluate manufacturing performance in an overview. This cursory analysis provides a state of the union and indicators to mobilize resources to correct non-performing assets.
This view provides total production summary versus target at different levels of aggregation by equipment and by material produced and provides 24-month trending and facilitates navigation to specific facility views to allow launching ad hoc data extracts and enabling custom reporting and analysis by corporate governance bodies.
ABOUT THE AUTHORS
Sandra Vann is a Six Sigma Black Belt at Dow Chemical Co. in Midland, Mich. (Sjvann@dow.com). Mike Williams is a senior project manager at Dow (email@example.com). This is an edited version of a paper originally presented at WBF’s 2007 North American Conference in Baltimore, Md.