Your Smart Manufacturing IT Clients Capitalize On Manufacturing Data
By Peter Guilfoyle, Vice President of Marketing, Northwest Analytics
A major challenge faced by manufacturers is the state of data useable to glean actionable intelligence. Manufacturing, compared to industries that are using more sophisticated data analytics tools — like advertising and finance — is often scrambling to find better ways to use data to improve quality and efficiency.
It is true today, in an age of Big Data analytics, that manufacturers on the whole are lacking in “tools to collect, condition, and analyze the information. What can be done to overcome the suggestion that manufacturing suffers from “DRIP” syndrome; data rich, information poor?
Complex Technologies In Isolated Systems
Today’s manufacturing environments are made up of complex, purpose-focused technologies that gather vast amounts of data in often-isolated systems that are not easily integrated. The data collected and captured in these systems is often referred to as “data islands” or “data silos,” reflecting the fact that the complexity and specificity of individual processes result in data isolation. Data doesn’t become useful information until analytics are applied to surface the significant process signals from each data system and woven into an integrated fabric of manufacturing intelligence.
The stakes are high in manufacturing quality. The performance of key components is often a serious safety issue, as in the manufacture of automobile or aircraft parts. Critical tolerances must be monitored and tested to ensure quality and safety. Except in prototyping of newly developed products, there are no production “betas” that can be operated on the manufacturing floor. Product runs have to be right the first time. Production stoppages or product recalls can be enormously expensive, running into the millions or billions of dollars.
Many of the systems in use today were designed with several-decades-old information technology hardware presumptions reflective of the storage and data transmission limitations at the time. Do companies have to concern themselves with new systems modeled on “smart machines” and the as yet nascent-defined impact of the Internet of Things in manufacturing, or is there value to be realized from the data, technologies and processes already in place?
A closer look shows there are smart manufacturers proving that intelligence surfaced from current production data can be a powerful tool for improvement of product quality and process efficiency.
Dow Chemical Was Drowning In Data
Dow Chemical, headquartered in Midland, MI, is the second-largest chemical company by revenue in the world. More than 5,000 plastic, chemical and agricultural products are manufactured at 188 sites in 36 countries.
For Dow, a single product manufactured in different plants will likely be produced with varying equipment and systems, each set up with their own production lines, processes and data storage, many of which are not tied to a database. Across their facilities, volumes of manufacturing data were being gathered. The challenge was to turn that data into meaningful intelligence.
Lots Of Data, Not Enough Information
At Dow, as with many other manufacturing companies, much of the data gathered was never looked at after it was gathered in the initial factory process. Since the data was stored in silos segregated by purpose-built systems, integration for the purposes of analysis confronted enormous obstacles in terms of cost and inconvenience. Even the suggestion of replacing manufacturing process technology could rapidly become an investment of millions of dollars and months of migration and integration of process data.
Much more desirable would be a method of analyzing the process data in place where it currently resides, without constructing new systems or configuring a redundant, parallel data repository. What was needed was a neutral or agnostic system with the capability to analyze the data from all data silos, islands or disparate databases and provide intelligent views of manufacturing processes in real time.
Dow turned to long-time partner Northwest Analytics to implement a solution based on enterprise manufacturing intelligence; a system of real-time process analytics, alarm notifications, and role-based dashboards from existing manufacturing data sources. Called NWA Focus EMI, the new system quickly accessed data directly from Dow's many original sources, passed it through a global analytics layer, and then provided visualizations of the information and trends in real time.
The global analytics layer turned the data into the type of information and metrics anyone could use. Everyone — from the vice president of manufacturing processes to the plant manager to the operator on the plant floor — now see the same production data, can access it directly, use it to do their jobs and take action if needed. And it can be delivered in accordance with the desired sample rate.
For Dow, the return on investment for implementing a manufacturing intelligence solution was immediate. The company gained improved visibility, efficiency and quality by being able to aggregate various data islands and databases into one, integrated information system. And they gained quantifiable value by avoiding expensive mistakes in the manufacturing process, saving millions of dollars.
Recasting A High-End Foundry Into A Data-Driven Manufacturer
Palmer Foundry in Massachusetts designs and manufactures high-end, precision aluminum castings for numerous industries, including semiconductor, automotive, energy, aerospace, robotics and many more. Each project requires demanding specifications and diligent control of raw materials and processes.
The foundry's success and customer satisfaction were unquestioned. However, Palmer's existing customers, driven by their own corporate requirements, began pushing Palmer to meet qualification requirements in order to continue as a vendor. Prospective customers, even when impressed by the foundry's capabilities, balked at bringing new business to them because of a lack of demonstrated process control. Palmer realized that delivering superior products and service was no longer sufficient to secure business in a data-driven, process-controlled manufacturing world. They needed to adopt quality control processes to meet both the requirements of their customers and manufacturing industry standards.
At the outset of the project, process data at Palmer was scattered across laptops in isolated spreadsheets that could be described as “data tombs” – data was stored but rarely seen or used again. There was also a lack of staff with statistical process control skills and a small, overburdened IT department. Implementing a completely new system of software and hardware would mean taking employees away from manufacturing activities for training, which would be difficult for the company to afford or justify.
Palmer understood that in order to meet their customers’ requirements, they needed to improve internal processes and gain control over supply chain issues that could affect product quality. They also realized the need for the full analytical functionality provided by a manufacturing intelligence software solution to support business management and strategic planning. But circumstances dictated that the solution would have to be installed, configured, implemented, learned and adopted in a short amount of time with limited resources, and by leveraging their existing technology investments.
Palmer selected a full-featured manufacturing intelligence solution. Installation was conducted over the phone (with limited IT involvement) and — in less than three weeks — Palmer was operating with direct data-source connectivity, global analytics, real-time role-based visualization, and alarm notification services. Using links to existing databases without duplication or redundancy, data is accessed directly at the original sources, passed through a global analytics layer, and delivered as actionable information through the intelligence viewer.
Improved Supply Chain Validation, Greater Process Control
For Palmer, this meant accessible, real-time information at their fingertips, a visual onscreen display of relevant information, and a color-coded warning system of deviations from project or raw materials parameters. Palmer staff can now verify raw materials before they enter manufacturing, accepting or rejecting shipments or adjusting manufacturing processes to accommodate material variables. They also enjoy greater real-time control over manufacturing processes, warning operators of potential problems, required adjustments or corrections, and identification of process inefficiencies, such as high scrap and rework rates.
Palmer's new manufacturing intelligence capabilities added a data-driven showcase for their highly skilled foundry to reassure their current business partners and win new contracts. As a result, Palmer has repeatedly received vendor approval status across its customer base. Going forward, Palmer plans to continue capturing data to drive process improvement and establish long-term best practices, speeding identification and resolution of quality issues, improving supplier performance, and customer confidence.
Real-Time Manufacturing Intelligence From Existing Process Data
Manufacturers generate plenty of data with existing systems. The challenge is converting that data into the intelligence needed to make real-time actionable decisions. Forward thinking companies as large and diverse as Dow Chemical, or as specialized and focused as Palmer Foundry are demonstrating that solutions based on analytics that access existing data to produce real-time actionable manufacturing intelligence are powerful, flexible and affordable. They have implemented systems that capitalize on existing data collection technologies to analyze and visualize information in a way that improves products and empowers operators, supervisors and managers without breaking the bank or disrupting production.