This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

Data Analytics for Smart Manufacturing

Total: 0 products

  • Choose product to compare

Cancel
Compare with CISCO
Compare with H3C
  • Choose product to compare

Cancel
Compare with CISCO
Compare with H3C
  • Choose product to compare

Cancel
Compare with CISCO
Compare with H3C
  • Choose product to compare

Cancel
Compare with CISCO
Compare with H3C
MENU

Data Analytics for Smart Manufacturing

By Dr. Shi-Wan Lin, Co-Chair, Architecture Task Group, Industrial Internet Consortium and Co-Founder and CEO, Thingswise

CEO and co-founder of Thingswise. Co-chair in technical groups for the Industrial Internet Consortium (IIC) and the National Institute for Standards and Technology Cyber-Physical Systems Public Working Group. A member of the Edge Computing Consortium (ECC) Advisory Committee.

Smart factories are a novel paradigm for manufacturing that involves robotics, integrated product production simulation, additive manufacturing, and 3D printing. The ‘smart factory’ concept revolves around how to apply Information and Communications Technology (ICT) in traditional production environments to deepen the convergence of ICT and Operational Technology (OT) to achieve higher degrees of automation.

Smart factory initiatives include digitalizing production technologies and processes; connecting machines to communication networks for data collection, analysis, and utilization; the integration of OT and IT technologies; and the convergence of production and service systems. These initiatives can be implemented not only inside an enterprise, but also across enterprises, and even entire industry ecosystems.

A core concept of smart factories is the infusion of intelligence into manufacturing systems and management processes by using all-digital technologies for global interoperability. After defining the concept, we can easily conclude that specific architectures may differ in implementation details and points of focus, but can ultimately be complementary.

It’s worth mentioning that building all-encompassing systems for smart factories is often very complicated and generally involves long lead-times. At the beginning of each new project, it is important to follow the core concepts described above to guide the strategic vision and architectural design. As the unique details of each implementation become clear, it is essential to incorporate the core values of the enterprise and then iteratively progress from general principles to complex details to assure that the strategic vision is included in the completed facility.

Handle Business Issues First

Enterprises will encounter many technical and implementation challenges on the way to smart factories.

Industrial Internet of Things (IIoT) and smart manufacturing are characterized by vast ecosystems that involve numerous upstream and downstream stakeholders. In the next five to ten years, relevant IIoT and smart manufacturing technologies will continue to develop and evolve quickly. Smart production will be a continuous process of constant updates, enhancements, and improvements.

While advancing the smart production process, we must not blindly invest in everything for the sake of creating an all-encompassing system; otherwise, the end results will likely lead to inefficiencies or possible failures. We need to avoid technology-dominant initiatives that lack strategic guidance or business value drivers. Additionally, we must not turn a blind eye to or feel at a loss about smart production so that we will not go through delays or miss opportunities.

The rise of the Internet and the increasing dominance of eCommerce are having an overwhelming impact on traditional businesses. The IIoT and smart manufacturing will bring immense growth opportunities to enterprises, while simultaneously posing a challenge to each company’s very survival.

The most pressing task is to handle business issues as guided by the strategic vision, in accordance with business values, and through the use of innovative technologies. To this end, adding intelligence to legacy production systems is a feasible option.

Data Analytics Drive Smart Productivity

Intelligence is a cyclic process that covers perception, awareness, decision-making, action-taking, and goal attainment. Restated more simply, this process includes learning, adjustment, and adaptation.

Data analytics in smart factories help to quickly and accurately ascertain the status of equipment and production operations in order to make the right judgments and decisions accordingly. Just as fuel powers a plane’s engine, allowing it to fly; so too does data fuel data analytics, driving production and operations in smart factories.

Data Applications Shift from ‘Reactive’ to ‘Proactive’

At present, most industry insiders believe that the best utilization of data is to first collect large amounts followed by digging out maximum value afterwards. This post-event data utilization approach is at least one-sided, or worse, incorrect.

A more effective way of utilizing data is when enterprises approach the matter in a proactive, targeted, and purposeful manner. However, a high-priority issue in production operations is the need to intelligently optimize production technologies and processes. What information needs to be collected to achieve such intelligent optimization? To obtain said information, what data is collected and how is the collected data analyzed? Which equipment is needed, and how is the equipment connected to collect the desired data?

In smart factories, the results of data analytics should be used to enhance the operational intelligence of stand-alone equipment and groups of equipment. This requires continuous, near real-time streamlined data analytics.

Traditional batch data mining methods will continue to have their role in smart plants; for example, to develop analysis modules for operational systems that will be applied to post-event scenarios. The point is that batch processing is neither the only nor the primary analytic process.

Platform Components for Data Analytics

Technologies related to data analytics are developing quickly, and different types of platforms are emerging. Some platforms provide technical components for data collection, storage, and/or analytics for cloud computing; users are allowed to flexibly select these components and generate relevant sets of analysis tools. This includes cloud platform components for equipment connectivity — including security and data upload services.

The functions that these platforms provide are primarily focused on equipment operations and management. Additionally, most of these platforms need to be deployed on a cloud in some form, and only a few of them can be installed at local production and operations environments to support edge computing and act as parallel, out-of-band analytics for device or control systems.

If we want to implement data analytics for legacy production equipment, the biggest technical challenge is how to connect to legacy equipment interfaces, such as Programmable Logic Controller (PLC) ports, to be able to access the data over an IP network. If we fail to tackle this technical challenge, we will be unable to collect useful data. In this case, even the most powerful data analytics platforms are useless.

Addressing this particular challenge requires the collaboration of multiple parties. Equipment vendors, customers, and related connection-solution developers need to work closely with each other. The result is the ability to quickly and easily collect equipment data by referencing the interfaces and specification documents provided by the original equipment vendors.

A key factor in implementing a smart manufacturing plant is having the equipment analysis results for running equipment and operations processes returned quickly. Combining the analyzed results with other business information — such as market supply and supply chain data — to take production intelligence to new levels is also important. To achieve these goals, the following three points require emphasis:

  • Equipment runs continuously, and its operation requires endless intelligent data feedback. Therefore, the data analytics system must conduct streamlined analytics on data flows generated by the equipment, provide information for decision-making, and then automatically apply the decisions to the running equipment and operation processes. By contrast, traditional analysis frameworks based on batch and passive data queries are unable to effectively support continuously running equipment and operations. Therefore, streamlined analytics is a primary function that data analysis platforms must provide.
  • For the sake of security, reliability, and effectiveness (e.g. constraints on latency and data traffic), data analytics platforms must provide distributed analytics to deploy analytics capabilities locally at the equipment or production facility sites and support edge computing.
  • The data analytics platforms — whose required technologies must be enhanced and simplified — provide customers with ‘out-of-the-box’ analytics systems that are characterized by easy deployment, customization, and maintenance. With such analytics systems, customers are able to quickly and iteratively evolve their smart plant applications. Currently, however, many manufacturing companies do not have the necessary levels of IT expertise. On their journey towards smart factories, these organizations must be able to acquire and benefit from the latest data analytics techniques (including advanced technologies such as machine learning), instead of being stuck in a quagmire due to complexity and specialized skills requirements of IT professionals.
  • ICT Insights on the go

    Scan for mobile reading

    ""

    All-Cloud Network Extra

    All-Cloud Networks Empower Digital Transformation

    Current catalogue

    Comment

    Features

    Special Report

    Focus

    Ecosystem

    Success

    New ICT

    Last Word