with analytic applications
There are many possible missteps throughout the manufacturing process. Because each process depends on those that come before it, small issues quickly turn into large problems. Even with increased digitization, today’s manufacturing leaders face many of the same challenges as their predecessors.
The success of a manufacturing company depends on the quality of the end product. Each time a product that falls short of expectations or specifications makes it to market, the quality deviation creates negative impacts that reverberate throughout the company. Often, plant managers don’t realize a deviation has occurred until after production, leaving leaders to focus on reactive resolution.
Additionally, each time a process deviates from the norm, referred to as an excursion, waste occurs — in time, materials, or energy. Waste from excursions causes a wide range of issues, including higher costs, production delays, increased labor costs, and even reputational damage. At many plants, a single issue, such as one machine performing slightly slower than usual, can lead to a significant amount of waste that affects the company in multiple ways.
On top of this, plant managers must constantly watch their production efficiency. Seemingly minor delays can have major consequences. Alongside increased costs and lower quality, employee morale can fall in the face of consistent efficiency issues. Manufacturers looking to maintain a competitive advantage and high customer satisfaction, therefore, must ensure that each product is produced and delivered as promised.
Limitations of traditional factory management
Many plant managers use traditional reactive root cause analysis to solve these age-old issues: repairing a machine after it breaks or remaking a defective product. However, the negative impact is still felt because the issues have already happened. In many cases, the solution only solves that occurrence, and the same problem arises again.
Organizations are under fierce pressure from both customers and competitors, particularly in the high-tech manufacturing space. Equipment costs more – both to purchase and maintain. In an age where compute power and technological ability continue to grow exponentially, not least in the semiconductor space, customers are demanding more in terms of functionality and service. The “chip wars” - the competition, primarily between China and the US, to control the market in semiconductors, also known as chips - are also adding pressures given their importance to the economy. To stay competitive in the market, manufacturers must shift their approach – trying to do more simply doesn’t work.
By moving from a reactive approach to a proactive approach, plant leaders can prevent potential issues, such as quality deviations, waste, and low production. Factory data almost always provides clues that can be used to predict an issue, and plant leaders turn to technology to manage and analyze the large amount of data the factory produces. However, many turn to traditional technological tools that no longer work effectively with today’s smart factories.
Manufacturers often use multiple different types of statistical tools to manage factory data, such as control charts, failure mode and effects analysis (FMEA), and linear regression. Because the team reviews after the fact, these tools force manufacturers to operate in a reactive model. However, the biggest issue with statistical tools is that their complexity requires the user to be an expert, often in statistics or mathematics.
Alongside statistical tools manufacturers also rely on specialized tools. “Big acronym” software such as MES, ERP, and SCM may have built-in analytical capabilities but typically they connect only to the data they house, have proprietary but not necessarily best-in-class algorithms, and support only pre-defined workflows. While the predefined workflows for many of these tools make it easy to get started, customizing templates is often time consuming, and sometimes it’s still not possible to get the desired results. Manufacturing companies need tools that are both simple to use and easy to customize.
Statistical and specialized tools
Innovative manufacturing companies are taking a new approach that allows their organizations to lead the industry instead of following it. By using visual data science, manufacturers blend the best of statistical and specialized tools to gain advanced abilities, such as analyzing large quantities and configuring the software specifically for manufacturing use cases. At the same time, this approach democratizes data in a way previously not possible, allowing every engineer on the floor timely access to heterogeneous data to support a wide range of use cases.
The three key capabilities for putting this approach into action are:
Data mashup and exploration The first step is connecting and combining data from multiple sources, such as factory operations, raw materials, quality and reliability, maintenance history and product specifications. Next, engineering teams can wrangle the data (prepare, compute, filter, transform) and then explore and visualize from any angle to ask and answer cross-domain questions.
Tailored analytics Because every challenge will be different, engineers want to choose exactly the right analytical methods to suit the problem. They would like to choose from a variety of built-in methods or seek help from statistical experts to provide cutting-edge analytics that can drop into the environment and be applied in context to the problem.
Enterprise scale There are multiple interpretations of the word “scale”, and the best visual data science platform will accommodate all of them. For example, plant production mix and volumes often change quickly based on demand, driving the need to find insights quickly in technical data of all types and often in massive volumes. Because manufacturers must deliver a tool to thousands of users, they need to effectively manage the access and authorization when scaling. Multiple key engineering disciplines must also be supported, including process, equipment, and product engineers, each with their own requirements and needs to collaborate across teams. Finally, as finely tuned analytical workflows are developed, they must be deployed widely so that as many in the organization can benefit from that IP as possible, up to and including continuous, automated analysis.
How visual data science improves manufacturing operations
By using the Visual Data Science approach, Spotfire combines the best of statistical and specialized tools, giving its users the power to analyze data, collaborate and share insights and take action. With the power of automated, proactive root-cause analysis and the ability to capture expert knowledge for reuse, manufacturing leaders can generate the insights needed to make smart business decisions. Manufacturers of all sizes and industries find that Spotfire’s intelligence improves collaboration across the organization between departments.
Spotfire uses artificial intelligence as the foundation for its platform. With the AI recommendation engine, Spotfire automatically recommends visuals to best identify relationships, outliers, and use the data to solve problems. Also, AI allows manufacturers to set up automations that do not simply complete tasks but make decisions based on pre-defined criteria.
Manufacturing companies can apply Spotfire technology to these, and many other, use cases:
Process improvement Spotfire software gives manufacturers the data and insights they need to create the most effective processes and enhance product designs, using deep correlation and covariance modeling, time series analysis and design of experiments (DOE). Semiconductor manufacturers looking to exceed tight process specifications, achieve maximum throughput, reduce waste, increase profitability, and mitigate risk can use the platform’s task-specific and extensible visualizations together with advanced analytics to dial in their processes.
Quality management By analyzing factory processes as they are happening, Spotfire can magnify engineering expertise to uncover irregularities, deviations, and flaws and put in place multivariate monitoring regimes that detect process variations before they become exceptions. Production teams can then quickly adjust the process before sub-standard material lands in the customer’s hand, as well as fix the upstream issues to eliminate defects in future products.
Anomaly detection With Spotfire analytics, experts can analyze production data, understand the status of machinery and equipment, and anticipate breakdowns when critical conditions emerge. Sophisticated models of equipment and process behavior can be configured and continuously evaluated, triggering real-time analyst alerts to resolve emergent problems, reducing both waste and costs.
Supply chain optimization Manufacturers turn to Spotfire analytics to detect issues in the supply chain so leaders can respond to the changing customer and supplier demand. Additionally, analyzing supplier performance makes it easier to respond quickly and make data-driven decisions. Spotfire software also delivers predictive scheduling, material and vehicle routing, linear programming, and genetic algorithms. Custom visualization By using Spotfire Mods, organizations can access extensible pluggable custom visualizations to make it easy for anyone to build, share, and use specific industry related visualizations. For example, Violin Plot compares yield of manufacturing processes, machines or batches.
As Brembo SpA, the world leader and acknowledged innovator of disc brake technology for vehicles, began moving to a smart factory, the IT department realized that they needed a better handle on analytics and statistical analysis. At the same time, they needed to evolve their analytics infrastructure for improved performance and user experience. The team began looking for an analytics platform that worked both
on-premises and in the cloud to access and analyze the data so data scientists could embed R models for deeper analysis of its big data lake. Additionally, they wanted a tool that could be used to analyze data throughout the organization, not just in the manufacturing process.
After researching multiple options, Brembo selected Spotfire visual data science for its ease of use, ability to address the company's needs, and suitability for teams across the organization. The manufacturing team began using Spotfire analytics to cluster the cooling curves of its aluminum foundry and to predict the lifespan and maintenance of machine tools.
Spotfire’s technological approach in manufacturing
Increasing revenue with Spotfire analytics
Moving factory management from reactive to proactive
Disclaimer: The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters. To work with Reuters Plus, contact us here.
Discover more about Spotfire and what it can do for your business here
The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters.
Produced by Reuters Plus for
We are collecting data and showing the reality and life of our machine tools. For operators, it's very useful to understand which tools need to be changed.
Paolo Crovetti, Chief Information Officer, Brembo Group
The team quickly got up and running due to the easy-to-use interface, with business users, data administrators, and others creating their own reports without the IT department. Data scientists can take it to the next level by creating data groups, building visualizations, and creating dashboards. Because of the improved productivity and efficiency of the analytics platform, Brembo's revenue increased 25 percent.
This is a game changer, as Crovetti concludes: "With Spotfire, we are able to discover and rationalize a lot of information, which otherwise would not be possible."
The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters.
Produced by Reuters Plus for
Disclaimer: The Reuters news staff had no role in the production of this content. It was created by Reuters Plus, the brand marketing studio of Reuters. To work with Reuters Plus, contact us here.