Industrial Sound Emission: Detecting & Predicting Problems in Industrial Applications from Sound

Competitive edge: From superior hardware towards more intelligent, digital solutions

Industry 4.0 as well as predictive or preventive maintenance are hot topics in today’s industrial sector. Most companies we talk with face an ever-stiffening competition on their physical core products: Competitors, often from abroad, are increasingly catching up in terms of quality and start competing with high-end physical products. This development is not new, but it seems to be ever accelerating. 

For many years, the typical answer to the challenge was moving further up the value chain by offering fully integrated services instead of stand-alone products ("solution provider"). But also here competition is becoming intense. Instead, the intelligence of a product has become the most decisive factor to differentiate it from lower tier competitors: For example, does the product automatically provide a warning before it is at risk of breaking - so major damage, unplanned shutdowns and production loss can be avoided? It is capabilities like these, based on smart sensors and AI-driven analytics, which provide substantial competitive advantages today.

While there is a wide consensus about the increasing value of intelligent digital solutions, we nevertheless see few concrete  applications in the market that add real customer value. In our experience, many companies approach artificial intelligence and advanced analytics from the wrong angle. They ask: “What data do we have and what could we make out of it?”. However, the starting point for any kind of data collection should always be focussed on a tangible customer need instead. Find out more how we approach our projects:

Audio Analytics: “If there is a problem with the machine, I can hear it”

Many of our projects start with the compelling finding that highly experienced specialists within our customer’s workforce are able to hear if something is wrong with their machines or the processes they are responsible for. 

They listen carefully for unusual cracking, rattling, scratching, humming - or any other sound from the machine or the process. Yet it is rather unrealistic for those
specialists to be standing next to a machine 24/7. A fully automated, “smart”
condition monitoring solution would be highly beneficial.

Moreover, even the ears of the best technician are limited to the normal human hearing range between 20 Hz and 20 kHz. In contrast, our systems are able to work with any type of sound: solid borne sound like vibration or acceleration, audible sound, infrasound or ultrasound. 

And if a human ear can already hear a defect - it is a comparably simple task for PeakProfiling: our AI-based solutions are typically able to detect even the most granular differences in machine states.

"If there's a problem with the machine I can hear it"
Worker in factory

Vibration analysis: “If there is a problem with the machine, I can feel it”

We often start projects using audible sound, since this provides a very clear cut use case to many clients. 

However, in later project stages, we typically include solid borne sound as well, since there are several practical advantages: Solid borne sound is much more stable against environmental noise, a
typical challenge in an industrial setup.

Also, most companies are experienced in working with solid borne sound data. Accelerometers are often already installed and
data is being collected on a constant basis. Our advanced analytics can build on data that already exists.

On top of traditional vibration analysis PeakProfiling adds advanced sound features as well as cutting-edge classification approaches using artificial intelligence.

"If there's a problem with the machine I can feel it"
factory machinery

Our speciality: detection of specific defects and predictive maintenance

Typically, our clients ask us to solve two types of problems: We are either tasked to detect a specific faulty component or
process irregularity, or to predict upcoming defects before they actually happen.

Benefits: Cost savings and process reliability

The benefits of detecting or predicting failures in machines or processes are obvious:

  • Major damages and costly repairs can be avoided
  • Unplanned shutdowns are reduced
  • Maintenance routine can be planned more efficiently
  • Spare and wear parts inventory will be optimized
  • Production loss is minimized
  • Operational safety can be improved
  • Power consumption will be reduced

This either benefits our customers themselves, or they increase the intelligence of their products to enhance value for their own clients.

motor-detect-predict

Broad scope: almost any industrial application or process entails sound emission

The range of possible applications is very broad. In fact, any physico-chemical process causing a sound (e.g. cracking, grinding, abrading, etc.) or a vibration (e.g. weight imbalance, misalignment, looseness, wear &tear, etc.) can be detected at a very early stage - even before it becomes visually or audibly noticeable by humans.

Our experience covers diverse fields like electric motors, heating plants, gas tanks, separators, bearings, etc. If you are interested in detecting/predicting defects in your machines or processes, please contact us here.

Power plants
Power Plant
Electric motors
electricmotor
Separators
Multi_Gravity Separator
LPG tanks
tank-783495
Shopping Basket