Sounce: AI-driven SaaS Solution For Quality Assessments Based on Acoustics

Sound provides essential information about our environment, affects how we engage with each other, and — maybe surprisingly — tells us a lot about the products and technology we’re using. In fact, sound is one of the most original indicators for the functionality and quality of mechanic systems.

Sophie Schwandt, Solution Owner of Sounce, and Simon Weiß, Product Owner of Sounce, explain how artificial intelligence (AI) differentiates good sound from unwanted noise and how it takes the strain off engineers; for example in the automotive industry. Thanks to this new technology and its ability to learn, the power of sound can be leveraged to a new extension.

At Porsche Digital, we have been working relentlessly to revolutionize failure detection through sound

The control and elimination of unwanted sound is an important part of quality assurance. However, the design of effective testing systems to identify the great variety of potential noises is laborious, difficult and complex.

With the increasing complexity of the underlying mechanical systems, engineers need effective and time-saving tools to identify anomalies, defects and their sources. To that end, Porsche Digital has developed an AI system called Sounce, which precisely detects unwanted sounds and supports its root cause analysis through a deep learning approach.

Meet Sounce, the AI-powered sound detector

Sounce enables automatic detection of unwanted noise in real-time. It has been mainly designed for production and development processes: exemplary use cases are component endurance runs or end-of-line test benches for drive systems, closing systems or imbalance testing. We developed the so-called Software-as-a-Service (SaaS) solution together with our colleagues from the development department of Porsche AG. Sounce is an AI system that automatically detects noise and acoustical properties and helps engineers make sure that parts and products sound the way they should: nice and pleasant.

The challenge: noise analysis is a subjective and laborious process

Why do we need an AI to detect sound? Whether in quality assurance or in component qualification, sound and noises provide essential information about product and process quality.

However, detecting and analyzing noise and its sources can be complex and challenging. Therefore, it is sometimes omitted as a source of information. In other cases, it is the work of highly specialized engineers. However sound changes and if the monitoring system does not detect that change, defects can be overseen, and quality suffers. With deep learning, the system learns to adapt and offers an intelligent long-term perspective to quality assurance. Besides its transferability, the reduced effort of the setup is also an advantage of deep learning-based noise detection.

Our solution: AI-powered automatic failure detection

Based on a deep learning approach, Sounce reliably detects failures and unwanted sound. The results are documented, visualized and summarized in detail in a web application. Besides the classic test bench monitoring, the web application allows novel deep learning analysis for an improved root cause analysis.

There are five steps to its AI-based noise detection:

1. Listen & record: The test bench or station is equipped with minimally invasive sensor technology. The data acquisition is started. This process is done in a minimum of time to quickly come to the interesting part: the AI model training.

2. Evaluate: The engineer uses the software to document relevant quality criteria of the noise detection and thus creates the basis for the AI model training.

3. Train: Based on the available data a deep learning algorithm is trained and provided in the cloud.

4. Monitor & Detect: The test bench is continuously monitored, and noise anomalies are detected automatically in real-time. The noises can be visualized, evaluated, and compared.

5. Verify: Through easy interaction with the web application, the engineer provides feedback on the accuracy of the noise detection and optimizes the algorithm in the long term.

Good sound is good business

With Sounce, we want to improve the development and production of components in the automotive industry and beyond. It is an AI project that we have implemented as productive software and, needless to say, we are very excited about it. Our flexible Software-as-a-Service solution comes with a number of benefits, including objective noise evaluation, 24/7 monitoring (even in single-shift operation), remote test bench control, and seamless documentation of detected sounds.

If you’re interested in Sounce, please feel free to explore our website: www.sounce.io.

Sophie Schwandt, Solution Owner

Sophie Schwandt works as Solution Owner for Sounce at Porsche Digital

Simon Weiß works as Product Owner for Sounce at Porsche Digital

Simon Weiß, Product Owner

About this publication: Where innovation meets tradition. There’s more to Porsche than sports cars — we are developing new digital products and services — always with our customers in focus. On our Medium blog, we tell these stories. It’s about our #nextvisions, emerging technologies, and the people that drive our digital journey. If you want to know more, follow us on Twitter, Instagram and LinkedIn.

--

--

--

Next Level German Engineering: Where innovation meets tradition. The Porsche technology hub to create tomorrow.

Recommended from Medium

#IamtheFutureofAI Katrina Ingram

The Pros and Cons of Artificial Intelligence

Slime Robot! Why isn’t anyone talking about it?

Bias In Predictive Modeling

How to Use AI to Make Your Team More Productive

AI Is Smarter Than You

Adventures in hill climbing with AI

Non-Computational Implications of Artificial Social Intelligence

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Porsche Digital

Porsche Digital

Official Account of Porsche Digital | Our mission: Digital engineering to spark excitement and to create value

More from Medium

This is what banking will look like in 10 years

Why Agile is the way to go for Edge AI

in reality projects rarely turn out as planned

tLabel: Talabat AI Labels Restaurant Cuisines

What is Edge computing and why do we need it?