White Paper June 2018

IT@Intel

Developing a Scalable Predictive-Maintenance Architecture Executive Overview To gain and maintain a competitive advantage in today’s world, manufacturers from all industries are turning to Industrial Internet of Things (IIoT) technology. IIoT can help identify excursions1 in standard processes, prevent unscheduled downtime due to tool failure, decrease overall maintenance costs, and improve productivity. But implementing IIoT solutions can be challenging.

In early pilots, we reduced unscheduled downtime due to fan filter unit (FFU) failure by 300 percent over manual inspection.

Robert Colby Principal Engineer, Intel IT Infrastructure Stephanie Cope Yield Engineer, Technology and Manufacturing, Intel Steven J Meyer Sr. Principal Engineer, Intel Manufacturing IT Shai Monzon Industrial Manufacturing Engagement Manager, Intel IT Ravinder Singh Strategic Engagement Manager, Industrial Solutions Division, Internet of Things Group, Intel Eric Xiaozhong Ji Engineer, Technology and Manufacturing, Intel

Intel has used IIoT in our factories for years, and our solutions are often built on existing tools and processes. As we continue to learn, we now focus on connecting the unconnected and interconnecting intelligent devices to reduce strain on our network. In addition, we develop autonomous functionality that moves capabilities to the edge. Our goal is to design solutions that achieve interoperability and scalability that meet our needs today and in the future. When we designed a new use case to monitor the health of our fan filter units (FFUs) with predictive maintenance, our goals included minimizing data volumes across the network, defining baselines for FFUs, alerting technicians to potential problems in real time, and more accurately predicting maintenance needs. We partnered with GE Digital to develop a solution based on Intel® Internet of Things (Intel® IoT) Gateways and GE’s Predix* platform that performs analytics at the edge, while collecting summary-level data for long-term trend analysis in the cloud. Successful implementation of our FFU solution includes these benefits: • Improved operations on factory floor • Fewer excursions • Better visibility • Less unscheduled downtime FFUs represented a process small enough in scope, and large enough in impact, to demonstrate the positive return on investment (ROI) of predictive maintenance and the scalability of the solution. In early pilots, we reduced our unscheduled downtime due to FFU failure by 300 percent over manual inspection. The solution also provides a foundational platform that we can scale to more complex implementations in the future. 1

Excursions indicate undesired changes in the manufacturing process that result in significant damage to materials.

IT@Intel White Paper: Developing a Scalable Predictive-Maintenance Architecture

Contents 1 Executive Overview 2 Background 3 Solution 4 Benefits 5 Lessons Learned 6 Architecture 7 Next Steps 7 Conclusion

Acronyms AI

artificial intelligence

FFU

fan filter unit

IoT

Internet of Things

IIoT

Industrial Internet of Things

MQTT Message Queuing Telemetry Transport PaaS platform as a service ROI

return on investment

94 % traFFiC reDuCtiOn IoT analytics and data summarization can reduce cloud traffic by nearly 94 percent.2 2 2

Results based on Intel internal testing. Results based on Intel internal testing.

2 of 7

Background Intel’s factories are some of the most advanced in the world, with thousands of process and automation clients on the factory floor, hundreds of analytics servers in the data center, and hundreds of thousands of feet of fiber for transmitting data. Manufacturing semiconductors entails a degree of precision that not every factory requires. But all modern manufacturers can benefit from better understanding the health of their tools and processes. Industrial Internet of Things (IIoT) solutions can help overcome these common problems: • Excursion. An excursion in the manufacturing process, due to failing tool components, can impact yield and forecasting, making it more difficult for the factory to meet goals and objectives. • Unscheduled downtime. Without visibility into the health of all parts of a tool, it is difficult to predict maintenance needs, resulting in unscheduled downtime and less productive workers. • Higher cost of maintenance. When tools unexpectedly fail, in addition to the loss of productivity, repair costs increase due to expediting parts and the impact of secondary failure modes. • Labor-intensive processes. Relying on the labor-intensive process of having humans visually check tool health can be costly over time. In the past, technicians placed a hand on the fan to determine if it was spinning differently from the previous day. The lack of data made it difficult to identify anomalies and increases the risk of error. Implementing IIoT solutions can be challenging. Intel has been using IIoT for years, for a variety of processes, but we needed a new process to monitor the health of our fan filter units (FFUs)—which are ubiquitous in the factory (see Figure 1)—to better understand trends and predict failure. We also wanted to maintain a high degree of security and move analytics to the edge to reduce the amount of data transmitted across the network. We partnered with GE Digital to co-develop a highly scalable solution using Intel® Internet of Things (Intel® IoT) Gateways and GE’s Predix* platform.

Fan Filter Units

Excursion

Manufacturing Equipment

Inaccurate health assessment of FFUs due to poor visibility

Unscheduled Downtime Higher Maintenance Costs Manual-Intensive Processes

Figure 1. Manufacturing equipment often includes fan filter units (FFUs), which draw in fab air and push it through high-efficiency particle filters. This keeps the air inside the machine cleaner by filtering it and reducing particles in the machine’s clean chambers. Detecting FFU health is often a manual process, which makes predicting failure difficult. Share:

IT@Intel White Paper: Developing a Scalable Predictive-Maintenance Architecture

3 of 7

Solution At Intel, we have learned the importance of connecting the unconnected to enable data generation and collection, driving edge analytics to deliver real-time insights, and using IIoT and artificial intelligence (AI) solutions to push filtered data to the cloud. Intel IT selected FFUs as an excellent use case for predictive-maintenance IIoT using edge-to-cloud technologies to store manufacturing data off-premises. FFUs represent a single process that is small enough in scope, yet large enough in impact, to demonstrate return on investment (ROI) in the factory. We used Intel IoT Gateways with GE’s Predix platform to achieve this. Our goals included the following: • Edge processing. We implemented edge analytics to limit the amount of data transmitted across the network and to the cloud. • Reporting. We needed to identify the long-term behavior and trends of each FFU through traditional reporting—statistics, graphs, and charts— based on summary data. Essentially, our data collection and research needed to provide fan vibration fingerprints in various states for analyzing changes.

25 % MeMOrY utiliZatiOn 70 % CPu utiliZatiOn Our edge workload utilizes 25 percent memory and 70 percent CPU, with network traffic of 15 GB per day.3

• ROI. We selected FFUs because they represent a stand-alone process and are exceptional for demonstrating the ROI potential of an edge-computing and cloud-based IIoT predictive-maintenance solution. We evaluated several products, and GE’s Predix platform met the requirements of our solution. It also provided a high degree of security that met our criteria for storing confidential intellectual property data. The FFU solution included the following: • Sensors. We installed an accelerometer at the top of each FFU to measure variations in the fan’s function, creating a baseline for each fan that allowed us to compare behavior across the tool and fleet. • Edge integration. We integrated the accelerometers with Intel IoT Gateways and edge applications using GE’s Predix platform. We then developed an edge architecture compatible with running machinelearning algorithms. This process created a baseline performance for each FFU, measured changes, and generated alerts for anomalies and potential problems. Alerts were sent to the cloud, and from there technicians were notified. • Cloud-based visualization. We transmitted summary data to GE’s Predix platform in the cloud to give tool owners a view into baselines and longterm trends in traditional, easy-to-use report formats. We use our gateways for edge analytics to reduce the amount of data we send to the cloud. Software on the gateways use the algorithm inference to identify signals that predict failure. We send only these signals to the cloud to alert the technician that the fan needs to be replaced. Our edge workload utilizes 25 percent memory and 70 percent CPU, with network traffic of 15 GB per day.3 3

Share:

Results based on Intel internal testing.

IT@Intel White Paper: Developing a Scalable Predictive-Maintenance Architecture

4 of 7

We also used this design to provide technicians and tool owners with real-time summary alerts and statistical reports for long-term planning (see Figure 2). We used the trend data we collected from the FFUs to predict failure more accurately and help avoid unscheduled downtime, increase worker productivity, and reduce the cost of replacement parts through better planning. The potential value of automatically detecting failures in other common sub-assemblies spans the entire manufacturing process, and the FFU solution has allowed us to validate the output and demonstrate positive ROI. The scalable, interoperable platform can be applied to more complex implementations in the future. Manufacturing Equipment A

Edge Analytics Sensor

Sensor

Sensor

Sensor

Sensor

Sensor

Sensor

Sensor

Intel® IoT Gateway GE Predix* Platform Summary Data

Reports

Edge Analytics Sensor

Sensor

Sensor

Sensor

Sensor

Sensor

Sensor

Sensor

Intel® IoT Gateway GE Predix* Platform

Manufacturing Equipment B

Figure 2. We installed sensors on each fan filter unit (FFU) to measure vibration and transmit that data to the edge application. We then employed an edge architecture compatible with running machine-learning algorithms to generate real-time alerts based on inconsistencies and potential problems.

Benefits Our IIoT solution for FFUs provides the following benefits: • Improved factory floor operations. Understanding the health of equipment is critical to energy efficiency, forecasting, yield, and worker productivity. We have increased FFU uptime by over 97 percent by ordering replacement parts upon potential failure detection and proactively replacing them before the failure occurs. • Decreased excursions. Detecting FFU inconsistencies helps avoid one of the leading causes of excursion in the process. Our solution has allowed us to order replacement FFUs in advance, and in those instances, we have experienced no excursion and no impact to product. • Improved visibility. In the past, failures were manually identified. Now, our technicians and tool owners can use summary data to more quickly identify changes in FFU health. • Reduced unscheduled downtime. Better visibility allows us to predict maintenance needs before FFUs fail, helping to avoid unexpected downtime and loss of worker productivity. We have reduced our unscheduled downtime by 300 percent over manual inspection. Based on the prediction of upcoming FFU failure, the FFU is now ordered days ahead and we can plan for downtime. Share:

IT@Intel White Paper: Developing a Scalable Predictive-Maintenance Architecture

Lessons Learned Intel IT has been implementing IIoT solutions for years, and we have learned what common issues to consider when undertaking a manufacturing solution. Advanced planning can lead to a faster, more flexible deployment, and address the following: • Security. Securing data both on- and off-premises is critical, especially in our ever-changing world. Our FFU solution uses GE’s Predix platform because it meets our overall criteria, including storing sensitive intellectual property data. Intel IoT Gateways also include hardware-level security from edge to cloud that not only protects, but also detects and corrects security issues. • Data volumes. Today’s industry is data driven, and automated solutions generate enormous volumes of data that can slow networks and present storage challenges. It is important to develop a management strategy to handle these data volumes. We initially sent all FFU data to the cloud, but quickly understood that it was too much. By moving compute to the edge, our FFU solution reduces the amount of data transmitted across the network, including summary-level information for long-term analytics. Leveraging IoT edge analytics and data summarization can reduce cloud traffic by nearly 94 percent.4 • Data structures. IIoT solutions collect information from a variety of tools, produced by different providers. Each one can have its own unique data structure, which must be standardized in order to fully understand the processes and health of the tools. We standardize all data collected prior to integration so that our measurements and analytics are consistent across tools and providers. • Data validation. When our FFU solution first uncovered potential problems, technicians were initially skeptical about accuracy. But testing and validation quickly confirmed the accuracy of the solution and gained technicians’ trust in IIoT. IIoT technology for predictive maintenance is still new, and over time, with continuous input and fine-tuning, the accuracy of the results will improve. Technicians and solution architects will learn to trust predictive analytics by measuring and validating the output. 4

Results based on Intel internal testing.

Share:

5 of 7

Defining Measurable Success Criteria With any Industrial Internet of Things (IIoT) solution, defining measurable success criteria is critical. For fan filter units (FFUs), we used the following to evaluate the overall success of the implementation: • Hard failure. Our goal was to detect 100 percent of hard failures within five minutes of failure, while minimizing false positives. • Soft failure. Our goal was to detect 80 percent of potentially faulty fans that had not yet failed (soft failures) within 24 hours, while minimizing false positives. • Baseline. We defined a baseline—or fingerprint—of a healthy FFU that could be used in detecting hard and soft failures. • Sensors. The solution needed to support up to 10 accelerometers per tool with a sampling frequency of 1 kHz (1,000 data points per second). • Scalability. The solution needed to scale to hundreds of tools with thousands of sensors without impacting our infrastructure, such as overloading the network with data transmissions. • Compatibility. Our user interface for data extraction and analysis had to be compatible with approved mobile devices for integration with current factory operations. • Edge processing. The solution needed to analyze data at the edge and single out only signals that predict failure before sending them to the cloud.

IT@Intel White Paper: Developing a Scalable Predictive-Maintenance Architecture

6 of 7

Architecture Intel® technology allows us to reuse existing software and connect tools that may or may not have been designed for IIoT. The open, interoperable platform works with our existing data center software, and enables performance at the edge at a lower cost. Using Intel IoT Gateways and GE’s Predix platform (see Figure 3), the FFU solution offers the following: • Scalability. Standardized data structures, reduced bandwidth requirements, and cloud storage provide excellent scalability to multiple FFUs and other manufacturing processes. • Security. Our high-security solution allows us to protect the data, both in the factory as well as the cloud. Cloud-based platform as a service (PaaS) security decisions require a thorough vetting of over 100 security data management policies. • Performance at the edge. The Intel Atom® processor E3800 product family—systemon-chip (SoC)—delivers outstanding compute, graphical, and media performance while operating in an extended range of thermal conditions compared with previousgeneration Intel Atom® microarchitecture.5 • Interoperability. Using the Message Queuing Telemetry Transport (MQTT) lightweight connectivity protocol, we can connect power-constrained devices. Designed for lowbandwidth networks, it allows smooth connection between devices to the cloud. • Open platform. Open source embedded products allow us to use the latest code and the most recent technologies.

5

intel.com/content/www/us/en/embedded/products/bay-trail/overview

Solution Architecture Intel® Internet of Things (Intel® IoT) Gateway Technology and GE Predix* Platform Edge Analytics

Cloud

Connectivity and Insights

GE Predix* with Failure Alerts and Trend Reporting

Intel® IoT Gateways with Algorithm Software

Digital Twin Modeling and Analytics

GE Predix* Platform

Asset Services

Message Queuing Telemetry Transport (MQTT)

Data Services

Event Services

Orchestration

Industrial Intelligence Dashboards Visibility and Insight

Operations

Data Connectivity and Ingestion

Intel Atom® processor E3800 family Failure Signals

Figure 3. Intel® Internet of Things (Intel® IoT) Gateways and GE’s Predix* platform solution provide scalability, security, interoperability, and performance at the edge.

Share:

IT@Intel White Paper: Developing a Scalable Predictive-Maintenance Architecture

7 of 7

Next Steps

IT@Intel

The scalable, interoperable framework we developed for the FFU process creates a foundation that can be used in a variety of other IIoT use cases in the factory. For example, we plan to scale this solution to include detecting anomalies in electro-mechanical devices and other manufacturing processes. Additionally, Intel is now partnering with GE to work with OEMs and suppliers to learn how a partnership approach, along with external domain expertise, can further enhance the value of this IIoT solution across Intel’s supply chain.

We connect IT professionals with their IT peers inside Intel. Our IT department solves some of today’s most demanding and complex technology issues, and we want to share these lessons directly with our fellow IT professionals in an open peer-to-peer forum. Our goal is simple: improve efficiency throughout the organization and enhance the business value of IT investments.

As Intel looks to scale globally, we will explore leveraging Predix* Asset Performance Management (Predix* APM) solutions to centrally view and manage our tools and assets.

Follow us and join the conversation: • Twitter • #IntelIT • LinkedIn • IT Center Community

Conclusion

Visit us today at intel.com/IT or contact your local Intel representative if you would like to learn more.

When Intel created a new use case to monitor the FFUs and better predict health trends before failure, we needed a solution that maintains a high degree of security, performs analytics at the edge, and minimizes the amount of data transmitted across the network. We were able to achieve these goals by using Intel IoT Gateways and GE’s Predix platform. Selecting FFUs as our first cloud-based predictive-maintenance solution also allowed us to demonstrate ROI with a single, ubiquitous process. Through data standardization, partnering with industry experts, and setting measurable goals, our FFU IIoT solution has allowed us to develop a foundation for edge-to-cloud IIoT computing that can scale to other processes and tools in the factory. This new solution has helped us experience a 300 percent decrease in unscheduled downtime over manual FFU monitoring, improving productivity and optimizing operations.

Related Content If you liked this paper, you may also be interested in these related stories: • Improving Manufacturing with Advanced Data Analytics paper • Horizontal IoT Platform Paves the Way to Enterprise IoT Success paper • IoT Plus Advanced Data Analytics to Improve Manufacturing Results paper • Using Big Data in Manufacturing at Intel’s Smart Factories paper

For more information on Intel IT best practices, visit intel.com/IT. For more information about the partnership with GE, visit intel.com/GE. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to www.intel.com/benchmarks. Benchmark results were obtained prior to implementation of recent software patches and firmware updates intended to address exploits referred to as “Spectre” and “Meltdown”. Implementation of these updates may make these results inapplicable to your device or system. All information provided here is subject to change without notice. Contact your Intel representative to obtain the latest Intel product specifications and roadmaps. Cost reduction scenarios described are intended as examples of how a given Intel-based product, in the specified circumstances and configurations, may affect future costs and provide cost savings. Circumstances will vary. Intel does not guarantee any costs or cost reduction. Intel processor numbers are not a measure of performance. Processor numbers differentiate features within each processor family, not across different processor families: Learn About Intel® Processor Numbers. THE INFORMATION PROVIDED IN THIS PAPER IS INTENDED TO BE GENERAL IN NATURE AND IS NOT SPECIFIC GUIDANCE. RECOMMENDATIONS (INCLUDING POTENTIAL COST SAVINGS) ARE BASED UPON INTEL’S EXPERIENCE AND ARE ESTIMATES ONLY. INTEL DOES NOT GUARANTEE OR WARRANT OTHERS WILL OBTAIN SIMILAR RESULTS. INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL PRODUCTS AND SERVICES. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL’S TERMS AND CONDITIONS OF SALE FOR SUCH PRODUCTS, INTEL ASSUMES NO LIABILITY WHATSOEVER AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO SALE AND/OR USE OF INTEL PRODUCTS AND SERVICES INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. No license (express or implied, by estoppel or otherwise) to any intellectual property rights is granted by this document. Intel, the Intel logo, and Intel Atom are trademarks of Intel Corporation in the U.S. and other countries. *Other names and brands may be claimed as the property of others.

2018 Intel Corporation.

Printed in USA

Please Recycle

0618/WWES/KC/PDF

Developing a Scalable Predictive-Maintenance Architecture White Paper

downtime due to tool failure, decrease overall maintenance costs, and improve productivity. ... data volumes across the network, defining baselines for FFUs, alerting ... our unscheduled downtime due to FFU failure by 300 percent over manual .... each FFU through traditional reporting—statistics, graphs, and charts—.

202KB Sizes 2 Downloads 1088 Views

Recommend Documents

Developing a Scalable Predictive-Maintenance Architecture White Paper
Executive Overview. To gain and maintain a competitive advantage in today's world, manufacturers from all industries are turning to Industrial Internet of Things ...

Developing a Scalable Predictive-Maintenance Architecture White Paper
are exceptional for demonstrating the ROI potential of an edge-computing ... upcoming FFU failure, the FFU is now ordered days ahead and we can plan for ... We installed sensors on each fan filter unit (FFU) to measure vibration and transmit ...

JADE A White Paper
FINALLY, THE MAIN FEATS OF THE JADE APPROACH ARE SUMMED UP, WITH .... ster or Gnutella) or a multiplayer game, require ... of another user: the server is not necessary but it is .... an agent platform and a set of services that should.

white paper
50 percent to 83 percent and there is no cap on the total Federal financial ...... project, the Medicare Participating Heart Bypass Center demonstration, was ..... rules, proceedings, and reports and make them available on a public Internet site. Its

White Paper - GitHub
manual labor was replaced by the machine. The second .... 6.1 M-POS Mode. With the revolution of eth——the introduced pos mode, the pos mode of block chain will also usher in a new spring. HDSS absorbs the advantages and .... architecture and futu

Overview & White Paper - Electroneum
market. This overview paper outlines in detail the concept and planning behind Electroneum ... listed on the London Stock Exchange, Mark .... 3G or 4G data.

White Paper - Neuromation
Oct 2, 2017 - decentralized networks. He is an entrepreneur, visionary, guru, author, blogger, keynote speaker, and thought leader of the global technology.

A scalable service-oriented architecture for multimedia ... - Uni Marburg
development and use of web services for a broad scope of multimedia .... computational cloud, such as the Amazon Elastic Compute Cloud (EC2) (Amazon Web ... The service provider is interested in efficient services in that the response time.

White Paper 2
Jan 7, 2005 - The best tilt angle for any PV array is the one ... Triple-Junction, amorphous silicon PV array would be .... Insolation level (W/m2) or (Whr/m2).

Microvisor: A Scalable Hypervisor Architecture for ...
of CPU cores, while maintaining a high performance-to- power ratio, which is the key metric if hardware is to con- tinue to scale to meet the expected demand of ExaScale computing and Cloud growth. In contrast to the KVM Hypervisor platform, the Xen.

A Scalable Security Architecture Enabling Coalition ...
Abstract- Coalitions between autonomous domains are often formed in real life scenarios in order to enable access permissions to shared objects on grounds of ...

A Scalable FPGA Architecture for Nonnegative Least ...
Each of these areas can benefit from high performance imple- mentations .... several disadvantages are immediately apparent—for each iter- ation there are two ...

A Scalable Security Architecture Enabling Coalition ...
The dynamic nature of coalitions poses new challenges relative to security ... this paper we introduce a robust and scalable solution that enables the realization of coalition .... Multi-domain coalitions are prominent in a big number of emerging ...

A Scalable Security Architecture Enabling Coalition ...
relative to security management and joint administration of resources; therefore ... between interconnected medical information systems; e-. Government is also ...

A scalable service-oriented architecture for multimedia ... - Uni Marburg
BPEL for data-intensive applications and (b) provides tools that further ease the ... management and cluster/grid computing, including multidisciplinary analysis ..... Step 3 After processing the whole input video in this manner, a final list is give

FaSTAPTM: A Scalable Anti-Jam Architecture for GPS
selectable by software, but larger tap lengths can be ac- commodated with minor .... receiver does not need to be integrated with the FaSTAP. AJ module and the ...

A Wire-Delay Scalable Microprocessor Architecture for ...
technology challenges of the coming decade. The Grid Processor architecture (GPA) is designed to address these technology challenges [2]. As shown in Figure. 9.6.3, each GPA implementation consists of a 2-D array (4x4 in this example but scalable to

A Scalable Hierarchical Power Control Architecture for ...
1. SHIP: A Scalable Hierarchical Power Control. Architecture for Large-Scale Data Centers. Xiaorui Wang ... Power consumed by computer servers has become a serious concern in the ... years as more data centers reach their power limits. Therefore, ...

Jiminy: A scalable incentive-based architecture for ...
teracts with a MySQL backend database for storage and retrieval of system/user data. ..... This is part of the reason why files in file-sharing peer-to-peer systems.

Scalable Elastic Systems Architecture
Feb 24, 2011 - vice and the cloud does not provide good primitives for the degree of elasticity you require, ... Free or idle resources become hid- den. If however, each .... of a principal. The top layer is the Elastic Service and Runtime. Layer.

White Paper 1
designed to provide a renewable energy power source to ... DC-Efficiency vs Irradiation. 3.5 kWp Triple Junction (UNI-SOLAR. ®. ) 0. 2. 4. 6. 8. 10. 0. 200. 400.