Detecting & Predicting Automotive Anomalies—Mapping New Routes of Excellence

Detecting & Predicting Automotive Anomalies—Mapping New Routes of Excellence

September 13, 2018 0 Comments
Detecting & Predicting Automotive Anomalies_870x450

The IIoT is having a dramatic impact on the auto industry. How can manufacturers—of any kind—benefit from a future of machine learning and anomaly detection?

The Industrial Internet of Things (IIoT) is changing everything. From the food we eat to the machines we use to the cars we drive. Everything is rapidly transforming, and for the better. With more things getting connected to the internet, the IIoT is swiftly extending its tentacles to the factory floor—the center stage for all the industrial drama.  

The automotive industry is among the highest impacted by the IIoT. According to Mary Barra, CEO, General Motors, “The auto industry is poised for more change in the next five to ten years than it has seen in the past 50.”

In fact, a study revealed that the number of networked cars will rise 30% a year for the next several years. This means by 2020, one in five cars will be connected to the Internet.

For more than a century now, the automotive industry has created competitive advantage mainly through engineering excellence. Going forward, this will no longer be sufficient.

The future will be an age where parts can monitor and evaluate their own performance and even order their own replacement if required. Real time analytics and cognitive systems will detect and predict anomalies, and turn data from components and systems into valuable insights.

These can then be used by automotive manufacturers to increase the reliability of cars and offer new value-added services to customers. In other words, technologies such as Machine Learning and Anomaly Detection can extend quality assurance beyond the factory doors. However, for all this to happen, it is very critical to set the foundation right.

The biggest challenge for any automotive manufacturer today is the catastrophe caused by unplanned downtime. One leading auto manufacturer estimates unplanned downtime in a factory can cost as much as $20,000 per minute. Huge losses in matter of seconds!

Most often, these situations are the result of production machine failures that could have been avoided if data from the machine was successfully analyzed to detect abnormalities, anticipate the failure and plan a repair.

Deconstructing Roadblocks—The Challenges

  • Unplanned production stoppages- Unexpected failures increase maintenance costs dramatically because of difficulty of reactive and unplanned maintenance. Being able to predict system failures beforehand increases machine availability, boosting productivity and decreasing downtime
  • Analyzing Petabytes of Sensor Data - Sensor equipped machines automatically capture and log data on equipment performance and health. These insights can accelerate efforts towards achieving all-round operational efficiencies. In fact, these sensors (up to 1000+ sensors per vehicle) generate valuable data, such as MPH, MPG, RPM, oil pressure, water temperature, engine temperature, tire wear etc. The flood of data generated has the potential to flag abnormal events quickly and suggest proactively corrective actions. However, due to the immense quantity of this industrial data, automotive manufacturers find it extremely challenging to get a grip on it.
  • Reactive Maintenance Strategies - A popular study revealed that most manufacturers spend 40% of their time on reactive maintenance, 45% on preventive maintenance and only 15% on proactive maintenance strategies. Most often this happens due the complexity of equipment and the lack of accurate data-driven asset management. However, in an automotive set up, any incident involving a machine failure which is not resolved within three minutes could subsequently lead to a production stoppage. Reactive maintenance will not solve the issue of unplanned breakdowns and the irrecoverable losses they cause.
  • Absence of a reliable Asset Failure Management Plan - Most automotive manufacturers do not have a viable automotive asset risk management plan in place. According to Reliable Plant, automotive manufacturers spend nearly 90% of their time on emergency breakdown repairs. These precious productive hours can be easily put to good use by getting visibility into assets which might break down in the near future. Unfortunately, automotive engineers spend most of their time chasing obvious signals of breakdowns, which can easily overshadow other inconspicuous but critical data (called the dark data). By creating a solid Asset Failure Management strategy, potential abnormalities can be detected. This dark data can then be used to predict failures and reduce maintenance costs.

Detecting Anomalies from Automotive Sensor Data

Anomaly detection continues to play a vital role in most industries across the globe. Analyzing automotive sensor data to detect and predict anomalies can be extremely useful to mitigate risks of unexpected downtime. According to a report by MarketsandMarkets, the size of the global anomaly detection market is estimated to grow from $ 2.08 billion in 2017 to $4.45 billion by 2022 at a compound annual growth rate of 16.4%.

It is critical for automotive manufacturers to ask the below questions:

  • Is there a valid Asset Failure Management strategy in place to detect and predict anomalies from time-series data to mitigate the risks of unplanned downtime?
  • Is there a foolproof system which can detect potential Anomalies deep seeded within the auto manufacturing processes?
  • How are anomalies in scrap/defects being identified to enhance the quality of the product and bring down the recall incidents?
  • How is the volume of sensor data being analyzed to detect and predict anomalies that can impact production capabilities?

It is time to accelerate productivity by monetizing this data. Even a small hike of 1% in efficiency gains can yield astounding results over a period of time. A 1% increase in production efficiency can create a value of $30B in fuel cost savings.

With accurate Anomaly Detection at the right time, your IIoT technology investments can be converted into a powerhouse of infinite value.

Sign up for a free trial here to experience the power of automated machine learning to detect and predict anomalies from your time-series data.

Start a Trial Today

Anita Rajasekaran

Anita Rajasekaran

Anita Raj is a Product Marketing and Growth Hacking Strategist on the Progress DataRPM team, with over 10 years of experience working in the field of big data, cloud and machine learning. She brings a deep expertise in running growth marketing leadership experience at multi-billion-dollar enterprise companies and high growth start-ups. 

Comments
Comments are disabled in preview mode.
Topics
 
 
Latest Stories in
Your Inbox
Subscribe
More From Progress
New_Mobile_Dev_Ebook_Progress_Website_Thumbail
The New Mobile Development Landscape
Download Whitepaper
 
IDC Spotlight Sitefinity Thumbnail
Choosing the Right Digital Experience Platform to Improve Business Outcomes
Download Whitepaper
 
TheFastestWayToBuildMobileAppsArtboard-2
The Fastest Way to Build Mobile Apps With Cloud Data
Watch Webinar