Optimizing Real Time Manufacturing Intelligence

Published Date: 02 Nov 2017

IT Services can now deliver Real-Time Manufacturing Intelligence ("MI"). Enterprise Resource Planning ("ERP") has been implemented for years, providing supply chain and business improvement intelligence. Real-Time MI has the opportunity for rapid ROI with enhanced yields, less downtime, and less waste.

Much research has shown that people are poor at determining correlations subjectively. Most industry experts only trust computer calculated correlations using statistical models. MI is a concept in the world of process manufacturing – unique for analytically alarming process trends to prevent out of control product problems.

Chemical, packaging, pharmaceutical and energy companies have long known that automating data collection and data analysis can lead to improving processes and yields. These companies have developed massive databases that collect detailed measurements from factory automation tools which are used later for off-line analysis.

The key to today’s MI improvement is to use new IT capabilities to leverage these existing, disparate database silos; provide real-time analysis and intelligence; and identify and correct problems in-line. This can be done without creating, duplicating or installing yet another database.

This paper looks at MI and case studies in chemical, pharmaceutical, and packaging process manufacturing and compares the major companies providing software options.

Paper for PICMET ‘13

Authors:

Jonathan Cooley, MBA, DBA Candidate, Adjunct Professor at Concordia University MBA Program & Royal Roads University, Zhuhai, China

Jim Petrusich, Vice President, North West Analytics

OPTIMIZING REAL-TIME MANUFACTURING INTELLIGENCE

In process manufacturing, time is quality or time is waste.

By Professor Jonathan Cooley, Partner, Advantage Partners USA & Asia

Jim Petrusich, Vice President, North West Analytics

We will explore and strive to define, compare and contrast core manufacturing software concepts including Material Requirements Planning (MRP), Manufacturing Resource Planning (MRPII), Manufacturing Execution Systems (MES), Enterprise Resource Planning (ERP), Business Intelligence (BI), Big Data, and Manufacturing Intelligence (EMI or MI). Although these may already be well understood, the overview leads to a more robust discussion of Manufacturing Intelligence (MI).

Much research has shown that people are poor at determining correlations subjectively. An accumulating body of research since the mid 70’s on clinical judgment, decision making, and probability estimation has documented a substantial lack of ability of both experts and non-experts to accurately draw conclusions subjectively. Evidence also shows that the same people have great confidence in their decisions and may not recognize their fallible judgment (Einhorn & Hogarth, 1978). MI focuses on manufacturing and improving subjective results by providing real-time, actionable analysis and statistically driven correlations.

The authors submit that any manufacturer that collects real-time data can implement and benefit from MI; and the IT departments of organizations can play a critical role in moving a company’s continuous improvement programs with these initiatives forward.

OVERVIEW

Manufacturing Resource Planning (MRP) & Manufacturing Execution Systems (MES)

What is Manufacturing Resource Planning and how can it help a business?

MRP began in the 1970’s. In the 1980’s Just-in-Time (JIT) and Japanese Kanban systems proved (Waldrop, P. & Scott, T. 2000 and Quezi, 2010) that manufacturing resources were best utilized when running and producing in balance. Eli Goldratt (Goldratt, 1984, 1990, 1999) developed both theory and software with complex algorithms proving that the local optimum does not equal the global optimum, meaning that maximizing production at each single unit that destroys balanced manufacturing actually is more costly than a balanced manufacturing process.

Walter Shewart (the "father of statistical quality control") and later a body of work by W. Edwards Deming (Deming, 1982) in the 1980’s provided guiding principles for gathering and analyzing data for manufacturing process and quality improvement. Deming became known as the "father of Quality Control" in Japan, dramatically improving both quality and efficiency in Japanese manufacturers. Their studies and practical applications, based on Walter Shewart’s analytical methods, proved that the tangible, intangible and hidden costs of poor quality affect not only the manufacturing process but also a company’s profitability overall.

Automated Shop Floor Controls were developed and implemented to monitor and report key factors of manufacturing processes and equipment. The goal was to monitor these processes and apply statistical quality control analyses to the resulting data. As a result, operations would be tuned. The best systems now provide this in real-time and employees can intervene in a process before significant out-of-specification product is manufactured. Off-line analyses are still a widely used alternative – albeit, usually too late for real-time interventions.

Material Requirements Planning or MRP software was designed allowing organizations to plan purchasing, monitor inventory, work-in-process, material and product levels throughout the manufacturing cycle, as well as finished goods, shipping and delivery. The goal of MRP is to help a company maximize and balance throughput with existing resources while minimizing costs. It provides a continuous work flow view at any snapshot in time regardless of whether a company’s Bill of Materials (BOM) is a simple single-level BOM or a complex multi-level BOM.

As MRP software has become more sophisticated, it evolved to Manufacturing Resource Planning or MRPII. It now suggests actions necessary such as ordering points and quantities given defined supplier and manufacturing lead times and schedules providing finished goods to meet customer requested delivery dates. Although MRP provides direction, it is a decision-making aid and not a decision maker. As with any system, the information supplied will only be as accurate as the information entered.

In a custom installations Automated Shop Floor Controls, Automated Test Equipment (ATEs) and MRP have been partially merged; however, few easily implemented options exist to facilitate this integration across the enterprise.

Manufacturing Execution Software (MES) began to be used in the late 1980’s and mid 1990’s subsequent to MRP systems. MES integrates the functionality of shop-floor data capture, packaged metrics such as overall-equipment-effectiveness, analytics and scorecards, continuous improvement capabilities, and paperless quality management. MES applications provide companies with the tracking capability to monitor manufacturing activities with greater resolution (hour/minute) and - when integrated with finite scheduling - to provide fast reaction to workflow changes. MES continually analyzes shop floor activities to be responsive to events as they occur. Manufacturers can now work with achievable shop floor workflow with the result of reduced cycle times.

Enterprise Resource Planning (ERP)

Why ERP? Enterprise Resource Planning (ERP) systems evolved to integrate business and accounting systems with MRP systems. In the early enterprise efforts, each software application was provided by different software, often from different software vendors. Without a great deal of poking, prodding and custom rewriting from an in-house or third party IT group these systems typically used disparate databases, did not share information, and did not talk to one another. For example, the accounting system did not exchange data with the MRP system or any other manufacturing software.

The idea behind ERP is that it must communicate across both manufacturing and business functions. With integrated ERP software, the financial software can create an accounts payable check as soon as the MRP portion confirms that goods have been received into inventory from a specific supplier. Similarly, accounts receivable will generate an invoice as soon as the MRP shipping clerk confirms finished goods are on a truck to the customer. All of this is done with a minimum of human intervention and paperwork.

ERP aims to replicate business processes, guide employees responsible for those processes through each area step by step, automate many business procedures, and tend to include applications that handle financials, order processing, inventory, distribution, customer relationships, etc. Various levels of automation are then implemented throughout various industries and companies within those industries.

There are a wide variety of offerings among the vendors...some, for example, include native asset management and supply chain applications, while others enable you to connect / use third-party products. In some discrete manufacturing ERP systems may handle all the needs; however, in process manufacturing such as food products, pharmaceuticals, or batch chemicals, not all ERP systems can manage this type of workflow alone. Specialized products, like MES, are typically more adept at these processes.

What are the potential ERP benefits?

Table 1

Increased competitiveness with integrated, fast, and flexible business processes;

Accelerated time to market with innovative, individualized products and services;

Simplified corporate structure, market channel, and business scenario management;

Improved corporate resource and asset utilization – and greater customer satisfaction; and

Consolidated foundation for the latest mobile, cloud, and in-memory technologies.

Source: SAP. Retrieved from http://www.sap.com/solutions/bp/enterprise-resource-planning/index.epx

The real benefits most people have been chasing have revolved around standardizing their business processes, building a clean base of data for further reporting and analysis, and removing the complexity and continuous customization problems surrounding individualized, legacy systems.

But there is a downside.

One of the essential factors that affect the efficiency of an ERP system is corporate integration of both processes and data. To ensure that the system works appropriately, business records and data must be integrated successfully. Engineering or re-engineering business processes, selecting effective software options and parameters, data migration and duplication from existing systems, and consolidation can be done but often at the cost of considerable time and expense.

The promise of a new ERP system may be great, but one needs to carefully consider the potential multi-million dollar project or the expense in terms of time, effort and human capital are worth long term gains. Implementing the software also changes what employees do and how they do their jobs. Employee resistance can be a major issue and usually requires significant change management skills (Slater, 1999). With careful planning and lots of elbow grease, ERP can work and make many an enterprise work better. 

Business Intelligence & Big Data

While data in general is ripe with the potential to be mined for insights, few software systems truly provide real-time, statistical analytic information from all the data. How does one give clear, actionable insights? Business Intelligence provides strategic insights from Big Data Analysis, but does not address critical real-time manufacturing processes. One of the major gains cited about Business Intelligence implementations are often related to the ease of report generation, not necessarily analysis of the data itself. Manufacturing Intelligence focuses on data analysis and clear presentations of the analysis - a key differentiation.

Sifting through all the data that even small to medium enterprises generate is daunting. Individual transactions can produce terabytes of detailed data in a single day, often stored in different databases which may not even communicate with each other. The recent industry buzzword for such enormous data stores is "Big Data." These data sets are so large and complex that traditional tools like relational databases, visualization applications and dashboards simply fall short of delivering useful insights (Savitz, 2012).

Compound this issue with companies operating globally, and you no longer have just site-related Big Data in information silos – one seems to have completely un-manageable sets of Big Data. Many big companies around the world still employ traditional tools manually complemented by discrete software or with custom legacy systems. In both cases the results have such latency that insights are quite stale by the time decision makers get them. One may argue that short-term decisions are strictly tactical and not as important to key decision makers while long-term decisions can be more importantly strategic. Unfortunately in today’s ever-accelerating pace of commerce, business intelligence is also quickly perishable. A "snapshot" taken only days before today may not suffice to help make tactical decisions required within a narrower window of time. Eric Savitz notes in Forbes magazine,

"Together, insightful strategic and tactical decision making creates for companies a clear competitive advantage, which can dramatically protect and accelerate their revenue and ultimately, their overall business success…It’s time for businesses to make a transformation – from utilizing Big Data trends to making both more frequent, tactical and strategic decisions (Savitz, 2012)."

The authors posit that the only way to do this in the manufacturing environment is with software that has five key features:

Table 2

______________________________________________________________________

Database Independence with the ability to access, analyze and report from disparate databases across an organization’s locations, preferably without creating yet another unique database;

Real-time, manufacturing specific statistical analysis and reporting;

Accumulated Causal Data stored in a searchable knowledgebase with statistical outlier data (CAPA or AC/CA);

Real-time Dashboard reporting for operators, managers, and executives to quickly identify key outlier events; and

Sharing and Collaboration tools across the organization to quickly react while documenting and retaining institutional memory for Best Practices.

___________________________________________________________________________

Pushing key operational knowledge and decisions directly to the appropriate level in an organization – be it the shop floor or the executive floor – enables manufacturing operations and allows transaction level insights in real-time. This is the promise of Big Data, but Business Intelligence applications do not address these five goals for manufacturing operations (Savitz, 2012).

The IT department’s role in these critical business functions is greater than ever. FedEx believes it has developed a World Class system providing real-time and historical analyses, advanced analytics and worldwide visibility across an integrated array of disparate data sources, types and platforms. The FedEx IT department - through Kim Baker in the role of Enterprise Architect - was challenged to provide package information available anywhere in the world within three seconds. Baker represents that the key is "the ability to detect events, evaluate them and optionally take action by the invocation of a service, the triggering of a business process or further information publication (Taylor, 2012)."

MANUFACTURING INTELLIGENCE

Companies such as Mercy Healthcare, The Nielsen Company, Google, Facebook, FedEx and others have all reaped competitive advantages from effective implementations of Big Data with Business Intelligence. The term "Manufacturing Intelligence" (MI) was first applied to the Lighthammer "Illuminator" product in 2001. Lighthammer was a pioneer in manufacturing intelligence and integration and was eventually acquired by SAP. MI software is frequently grouped with performance, plant, and factory management software including MRP and ERP systems; however, any company or business unit that collects process data into a database can implement Real-Time Manufacturing Intelligence and recognize significant improvements in quality and reduced process costs.

In 2011, analysts at ARC, Frost & Sullivan as well as AMR/Gartner added significantly to the body of literature on evolving Manufacturing Intelligence (MI) software - also known as Enterprise Manufacturing Intelligence (EMI).

Manufacturing Intelligence draws from the known concepts of Business Intelligence (BI). While BI focuses on acquiring data and keeping it available for management, MI connects manufacturing plant floor measurements and testing located in various data sources to automatically pull the requisite information and statistically alarm in real-time for unusual process variation.

Fulfilling the promise of effective Big Data, these systems utilize real-time advanced analytics to deliver information about manufacturing processes. This helps manufacturers optimize the performance of manufacturing processes, anticipate and prevent problems,

and improve manufacturing yields.

A common goal for using MI is to turn existing plant data – often captured automatically or manually - into useful knowledge that drives more informed business decisions.  The core components of MI include:

Data Capture both automated and manual;

Data Collection into their respective databases;

Data Integration across these data systems; and

Contextualization, a data map or production model, analytics and visualization, reporting, and simplified dashboards. 

Chart 1

As shown in Chart 1, the MI interface pipeline connects the disparate databases with real-time automated manufacturing data capture then provides real-time advanced analytics and dashboards. Manufacturing analytics within MI are statistically based and used to identify significant exceptions and events, providing rule-based visualization (Aberdeen,2008). MI software analyzes this data and provides key performance indicator (KPI) alerts, process verification, and process optimization.

When connected with Collaboration Tools (not just traditional email or reports), facilities and functional groups that were previously manufacturing data silos can begin to share problem-solving solutions and best practices. Most collaboration tools provide either quick one dimensional communication, or scheduled rich media multi-dimensional communication. The tradeoff is timeliness for a deep understanding of complex material. However the toughest manufacturing issues that consume our top engineering and scientific resources are usually both complex and time sensitive. As shown in Chart 2, most existing collaboration tools, fall short for providing problem solving solutions for these issues.

Chart 2

The software also enables the improvement of manufacturing processes, identification of best practices, and the ability to respond to exceptions and events. MI is especially effective in process manufacturing where re-work of discrete manufacturing units is not possible.

MI aids in the prediction of potential process and manufacturing problems before they impact critical factors such as cost, quality, or yield.

Why Does Our Business Need MI?

Analysts contend that a trio of demographic trends and manufacturing globalization emerged in manufacturing plants and control rooms create a "perfect storm" for MI solutions:

___________________________________________________________________________

The rising "gray wave" of post-war baby boomers in the U.S. and the results of the one-child policy in China1 is about to break on the shores of manufacturing. As the educated workers of the West leave the workforce for retirement, a significant volume of practical experience and domain expertise is leaving, creating a knowledge vacuum. Without a method to capture this knowledge and develop institutional memory, these skills will be lost.

A new, younger workforce is replacing the baby boomers. This younger, western generation is accustomed to acquiring and processing information and knowledge visually, virtually, and at high rates.

U.S. graduation rates are falling in the technical fields.

Sources: Compiled from Frost & Sullivan (2011), AMR/Gartner (2011), Poston (2012) and Cooley (2012).

1 China –the current world’s manufacturing workforce – is aging. Due to its population control campaigns beginning in 1971 and its one- child-policies, China’s elderly (over 60) is projected to grow to 31% by 2050. In 2010 the elderly represented about 12% of the population, and falling fertility rates cannot re-supply the workforce (Poston, 2012). Improvements in China’s infrastructure mean the 150 million experienced migrant workers now traveling cross country to work in Eastern China factories will begin working closer to home. Although India has a young population, it does not appear to have the infrastructure or workforce ability to fill this gap (Cooley, 2012).

The convergence of these factors is driving the heightened focus on optimizing MI solutions as it becomes increasingly clear that the way forward in manufacturing requires that the next generation of manufacturing decision makers be given access to a different set of applications than their baby-boomer predecessors.

One of the conclusions of the Aberdeen Groups’ 2008 benchmark report "Event Driven Manufacturing Intelligence 2008" was that just randomly collecting data in enterprise planning and production (Business Intelligence) was more likely to hinder effective cooperation between areas rather than encourage it.

Manufacturing Intelligence solutions are helping to bridge the gap between production and business environments. The continuing presence of inflexible legacy systems is facilitating the use of intelligent analytics to reduce complexity and simplify and accelerate decision-making.

The MI Market

Frost & Sullivan expects the world Manufacturing Intelligence market to attain revenues of $2.58 billion by 2014. Their research shows that the tactical need to integrate real-time manufacturing data with diverse business systems is driving the growth in MI and MI application approaches.

The MI market is now divided into two approaches:

1. Some on single, homogeneous databases such as SAP. These tend to be large, complex installations where significant custom programming is involved. The need for integrated systems on unified platform is driving consolidation; however, these tend to be large installations where existing legacy systems no longer are viable. These single-platform solutions have all the benefits – and the drawbacks – of large ERP and BI installations.

2. Others are dedicated MI applications providing database independence incorporating the multiple, existing databases found in and between manufacturing processes, lines and facilities. These applications leave the data in their existing databases and perform the analytics function while providing visibility. As a result, their cost and implementation time are much less.

Manufacturing Intelligence solutions help bridge the gap between production and business environments. The continuing persistence of inflexible and customized legacy systems is facilitating the use of flexible, intelligent analytics to reduce complexity and simplify decision-making. Additional needs in regulatory compliance, cost control, quality and the global pace and competition of business is driving real-time data acquisition, analysis and reporting.

So, who are the leading MI providers and what features do they provide? Chart 3 identifies key providers. Some such as SAP or Aegis can provide a single platform approach, others such as North West Analytics focus on MI as a value-added module.

Chart 3

Others such as MiniTab and JMP focus on designed experiment analytics and ignore real-time monitoring. According to Frost & Sullivan, MI vendors must provide value-added services to set their brand apart from other commercial, off-the-shelf products.

Moving to MI

What are the major considerations for Management, Quality and IT departments in moving their companies to MI? AMR Research, now part of Gartner Group, identified five core functions that every MI application should possess:

Table 3 FIVE CORE MI FUNCTIONS

___________________________________________________________________________

1. Aggregation: provides availability to data from multiple sources, most often databases.

2. Contextualization: supplies a structure or model for the data that will help users find what they need (e.g., a folder tree utilizing a hierarchy such as the ISA-95 standard).

3. Analysis: enables the user to analyze data across sources (and manufacturing sites), providing a foundation for genuine ad hoc reporting.

4. Visualization: provides tools to create visual summaries of data (e.g., dashboards) to alert decision-makers and focus their attention on the most important information at the moment.

5. Propagation: automates the transfer of data from the plant floor to enterprise-level systems such as ERP.

_______________________________________________

Source: Compiled from AMR Research/Gartner Group, Aberdeen, and North West Analytics

To this list, Aberdeen’s Best of Class manufacturing study (Aberdeen, 2008) would add the ability for these organizations to then "Innovate" efficiently. North West Analytics would add "Collaboration" innovations with a fully integrated knowledge base to capture institutional wisdom and provide enterprise-wide collaboration tied to plant-floor data to accelerate issue discovery, solutions.

Frost & Sullivan (2011) information, compiled with the authors’ suggestions, create seven guidelines to manufacturers and IT departments that are entertaining the implementation of Enterprise Manufacturing Intelligence solutions:

Table 4 GUIDELINES FOR IMPLEMENTING MI

_____________________________________________________________________

Implement MI to integrate plant data from disparate sources across all database platforms and data/operational silos with business content to deliver operational intelligence without duplicating existing databases or creating a new MI database;

Drive MI initiatives under the sponsorship of plant-floor management and IT responsible for operational performance and continuous improvement;

Flow data and analyses through analytic filters to deliver only relevant results;

Implement role-based dashboards throughout the corporation from the plant floor to the executive level;

Adopt remote access capabilities to provide visibility of critical data and to send alerts to key decision-makers;

Provide a mechanism for inter-company collaboration between facilities, management and locations; and

Consider the time value of information, implementation costs and manpower. Large, fully integrated systems have their advantages but may take man-years and significant capital investment with legacy system changes to implement while near-term strategic advantages with a rapid return on investment (ROI) may be available through MI focused applications that meet the criteria outlined in Table 4.

MI Case Studies

Aberdeen Group’s (Aberdeen, 2008) survey and analysis of Best Practices within manufacturers reflected that 49% of Best of Class organizations are more likely to align operational and compliance metrics as well as employ business process models, real-time visibility, and adjust these models based on the data.

Aberdeen’s analysis indicates that Best of Class manufacturers tend to implement all forms of automated data gathering, real-time monitoring and feedback including MRP, MES, MI, Big Data, Automated Quality Systems and more. Furthermore from their study, Aberdeen surmises that this ability to generate, store, analyze, present, and manage large amounts of data – and take timely action – provide these companies a competitive advantage. With a cultural shift to value-added decisions based on facts, they are more capable to respond to and innovate with market changes.

North West Analytics – this paper’s co-author – has implemented MI across new manufacturing systems interfacing as well to legacy systems and databases. Below, we provide a few sample Case Studies. They reflect the power of Manufacturing Intelligence concepts which can be implemented as pilot trials and expanded over time, or enterprise wide at one time.

Pharmaceutical Industry MI Case Study:

The situation was created by an expansion of a manufacturing facility in this multi-national company’s Puerto Rico operation.  The FDA requires that all manufacturing facilities whose products are bound for the U.S. market use analytic monitoring to ensure dosage quantities are correct within tight dosage guidelines.  The FDA drugs for these lines are in liquid form filling vials and syringes for intravenous injections.  Automated scales send fill weight data to a local database on the line and alert the operators when statistical variation of the fill weight exceeds control limits.  Sampling is also done from each batch by the laboratory to ensure product quality.  For overall Quality Assurance, results are stored in the Laboratory Information Management System (LIMS). Lab samples are also taken from products that come out as finished goods through the production line.  In addition, the Lab performs Stability Testing to determine and confirm product shelf-life data.  Although Stability Testing is performed during the R&D stages of product development, these tests continue to be performed throughout the life of the product in the production environment. 

Chart 4

A real-time alarming system monitors both the manufacturing and the laboratory databases on a wide range of variables.  If statistically significant changes occur in the process, the manufacturing system will alarm and notify operators, supervisors, process engineers, quality personnel, and management.  These alarms are based on analytics and allow employees to address issues before they continue; however, they do not prompt operators to react to measurements that may appear unusual but are statistically likely to be just part of random variation in the process.  Since implementing MI as shown in Chart 5, many people now have visibility to the manufacturing process and management is now well aware of how the line is functioning in real-time.  The system provides corporate visibility to the production line identifying batch quality

Chart 5

and fill issues early and thereby halting and correcting the process sooner. In the next MI phase the incoming raw materials inspection system will be implemented and added to the databases monitored by the real-time server. The company expects the benefits to accelerate and increase as other plants in the enterprise implement this approach and each operation begins to benefit from the experience, knowledge, and learning of the other facilities.

Chemical Processing MI Case Study:

In this processing plant of a major chemical producer, a wide variety of data was already being collected feeding a number of separate databases.  The goal of the MI system was to reduce down the time to find and resolve problems as they occur in the production process.  Although data is constantly being collected through automated on-line systems, there are occurrences when a problem has been raised that some of the important variables have not been captured over a long period of time.  This has caused large delays in resolving issues as the team assembled to resolve the issue finds they do not have access to all the information they need. 

  The new real-time analytic system triggers operators when such issues begin to appear.  Operators no longer are allowed to rely on tribal knowledge, shift expertise (companies that run multiple shifts usually have different procedures being followed on each shift, whether they are aware of this or not), or obsolete company practices that are no longer endorsed.  When operators respond to the alarm notifications, the data are automatically stored in an Assignable Cause/Corrective Action (AC/CA or CAPA) database.  As this database grows, the knowledge for resolving issues grows – not just in a single plant, but across the entire enterprise. Best Practices can be analyzed. Ineffective ones are identified and eliminated.  Although correlations are frequently identified for problems, determining causation is much more difficult, and state-of–the-art MI analytic tools like these help streamline the manufacturing process and cut waste. Collaboration tools allow issues to be solved across shifts and plants, and these Best Practices to be shared across the entire enterprise.  

When the system was first being installed, a meeting was called to show managers the new dashboarding capability. In the meeting one of the managers asked about a certain metric that was showing an alarm and asked if this was live or just simulated data. The data was live from the factory floor. He commented that this metric is important but sometimes ignored by operators. When the group had a follow up meeting a few days later, the dashboard history showed that the metric went back into control immediately after the previous meeting as action had been taken based on the real-time analytics and alerts. Everyone looked at the manager and he just smiled. "That is why we need a system like this," he said.

 

Packaging Company MI Case Study:

A flexible packaging maker implemented an MI system to streamline the start-up and shut-down processes of their production lines.  There are a series of operational check lists in the start-up process as each process manufacturing line begins to run.  While this is happening, the entire product running through the line is waste.  Certain variables are captured by an in-line vision system that feed a high speed database known as a "historian" database.  These vision systems capture far more information than common relational databases. Testing equipment such as these camera based systems are capable of producing massive amounts of data quickly.  Other quality and product tests are periodically performed during the process, and once the operator is convinced the line is producing quality product, the signal is given to run the production process.  This continues as long as consistent product quality is maintained. 

  The real-time MI software analyzes the real-time data and now allows the operator to know exactly when the line is running consistently, in control, and producing quality product without the additional periodic manual checks.  This has reduced the start-up time to stabilization that operators previously required to become convinced the process, product, and quality is stable.  In addition, the MI system quickly notifies the operators when the process is starting to trend out of control or becoming less stable.  Operators can now monitor the run-time process as well and stop production before additional product is produced out of specifications.  Since this is tightly controlled, the new MI system helps ensure quality and that shipments of bad product never leave the building.

THE FUTURE

While providing visibility to data is essential in today’s complex manufacturing environments, the real value comes in placing data into an operational and business context. Using that intelligence to drive effective, real-time actions is what creates competitive advantage.

Best-in-class manufacturers are implementing MI systems within individual manufacturing lines, within individual sites and across the entire enterprise and supply chain network. The momentum is building as new, low cost, easily implemented systems can survive and thrive along with legacy automation systems.

And what about the IT department?

Historically, the Information Technology (IT) arm of companies were led by the Director of Data Processing. Computing was initially seen as a back office tool for automating accounting and other financial matters. Later, when computers became more sophisticated and took on additional tasks throughout businesses, IT was relegated to assisting with the technical review of candidate software, assisting in physical hardware and software implementations, and monitoring and maintaining the resulting installation. The job title changed to Director of Management Information Systems. The breakthrough came in the mid-1980s, when chief executives increasingly invited their IT leaders to sit at the strategic business table. With the advent of the Chief Information Officer (CIO) position, IT moved into the executive business suite and IT professionals also became responsible for directing strategic initiatives.

Among the first people with the CIO title were Al Zipf of Bank of America and Max Hopper of both Bank of America and American Airlines. "Management’s Newest Star: Meet the Chief Information Officer," declared BusinessWeek magazine in a headline in 1986. "This reflected the recognition that what started as electronic data processing was beginning to become business strategy," said Harvey Koeppel, a longtime CIO who is now executive director of the Center for CIO Leadership (Icons, 2011).

"To stay relevant, the CIO role must evolve beyond the operational, shared service mentality. Droning on about uptime and upgrades is not going to cut it... In this world, the CIO becomes a mix of process officer, information broker and skunk works-type researcher… In this role, the infrastructure is far less important than the strategic direction of the company and a detailed understanding of the company’s markets, processes and relationships. Essentially the ‘Information’ portion of IT becomes far more relevant than the technical aspect (Gray, 2009)."

IBM CEO, Sam Palmisano, and others argued that it’s essential for technology executives to step up and help their companies deal with the opportunities and stresses of globalization and the emergence of disruptive new technologies, including cloud computing, social networking, advanced analytics and mobile communications (Icons, 2011).

The future of efficient, responsive and quality manufacturing lies with not only senior management, but with plant managers, quality managers and the IT department. IT can help identify significant trends and applications available for real-time MI as effective cost-benefit ratios and attractive ROIs. Manufacturing Intelligence systems provide IT an opportunity to play a critical strategic role in the manufacturing environment and across the enterprise.