GemDT and the Engineering Hub gives the ability for “True Predictive Maintenance” in Aerospace, Defence and Automotive

As part of the GemDT Engineering Hub, Digital Twin solution, working with Inora we are able to offer our clients the ability to offer "True Predictive Maintenance" for the first time, sensors will generate 100% accurate data to enable clients to make 100% accurate predictions on maintenance cycles.

The ability to repair something when it needs to be repaired, not when it is guessed to be repaired. Utilising data from the "Smart Factory" "The Hangar of the Future" and the "Workshop of Today" using the Inora technology on the GemDT Engineering Hub, clients can now do 100% accurate predictive maintenance.

Pure Data Control & Certainty -- At the Edge

Advanced Sensor Data Management

We are at an exciting stage in the development of information and knowledge. With more data about everything becoming available to us, often streaming-in at high speeds, there seems to be no limit to the potential for us to evaluate the state of our world. As we increase our understanding of these data and make better decisions, we are producing many important human and economic benefits, including -- health & safety enhancements, higher operating efficiencies, manufacturing advances, improved defence capabilities, and a cleaner environment, just to name a few.

Fundamental to this new reality is that we now have sensors that measure just about anything we can imagine – revolutionising the way we collect information about everything from temperature, pressure and frequency, to energy flow, heart rhythms and power usage, and thousands of other key metrics. In fact, with the deployment of IoT thinking throughout the scientific and business communities, the number of sensors and the amount of data being collected is absolutely mind-blowing.

And, beyond the sheer volume of data they collect, sensors are also playing critical roles in emerging technologies like autonomous vehicles where the capture of accurate information and its interpretation in real-time is absolutely essential.

All that said, the most significant challenge with sensors today is their reliability, which can be categorised into two primary areas:

1) Catastrophic failure – due to power loss, connectivity issues or severe damage.

2) accuracy deterioration – due to any number of factors, including, degrading hardware and/or electro-magnetic interference.

There are a number of solutions emerging for battery monitoring and replacement, and for backup power provisions in the area of catastrophic failure. Similarly, there are many emerging strategies to protect sensor connectivity and to prevent sensor damage in harsh environments. However, as challenging as these may be to resolve, the consoling good news is that when a catastrophic failure occurs for any reason, one thing is certain – users know that something is wrong, assuring that decisions will not be informed by bad data.

Unfortunately, the same thing is not true in the case of a sensor that has its accuracy deteriorating over time. Because, as a sensor suffers accuracy fall-off, the lack of knowledge about its efficacy and usability most often goes undetected. This means that data quality is seriously affected, but this not known when the data is being processed and interpreted. Consequently, any downstream decision making will continue to be made confidently, while it is being spuriously informed by bad data. Garbage in, Garbage out.

Beyond this, sensors also cause significant problems when intermittent failures occur. For example, vehicles are notorious for triggering problem codes in their electronic diagnostic systems because of intermittent sensor failures. Customers see a light on their dashboard and bring their car to the dealership for repair. Unfortunately, these are often “false positives” that the technician cannot repeat or confirm, and the diagnostics routines they run cannot locate the issue. This leads to unhappy customers who will return in the future with the same issue, poor brand reputation, and directly leads to billions of dollars of cost each year for the automotive industry.

Ellipsoid Analytics* (EA) from GemDT introduces the opportunity to move beyond uncertainty about a sensor’s accuracy at any point in time, and in doing so, also moves us into a new era of ultra-lean computing at high speed.

EA’s unprecedented, proprietary technology autonomously determines the holistic geometry and weight relationship of raw data with absolute numerical certainty – mathematically validated, on simple computing devices, in real-time. This means that there is immediate access to a true-value within any data-stream that can be used to know if each new data point is an anomaly, or not, as soon as it emerges from the sensor.

In addition, it also means that at any instant in time, there is an absolute understanding of a sensor’s capability – with any fade in accuracy also autonomously calculated and available to assess at every moment. This allows the user to have full control of decisions to replace sensors or continue to operate with their diminished accuracy – creating true, fact-based, well-informed decision making.

EA is bringing this advanced sensor data management capability to market with another important feature as well – one that is called, Real-Time, Edge Scrubbing (RTES). This new capability allows sensor data to be fully cleaned and processed immediately upon collection with lean computing hardware at the edge. In addition to the metrics that clearly identify the sensors current capability, the output data will be simultaneously purified of all anomalies/outliers, so that only quality data moves forward for use.

In addition, RTES also replaces any time-series data assessment tools (e.g. ARMA or ARIMA) with processing capability up to 15,000Hz. It should also be noted that RTES is a tool that works just as effectively with multi-dimensional sensors as with unidimensional sensors.

By having 100% accurate data, the ability to introduce Automation into the process is now here, accuracy never seen before enabling innovative organisations to "Inovate". Improving productivity, reducing errors, and increasing margins and customer satisfaction.

Whether you require sensor data management in critical business operations to drive operating efficiencies, or whether you are designing autonomous vehicles that must drive themselves, access to certainty about sensor performance and the reliability of sensor data is paramount. With the unprecedented capabilities now available from EA, you can not only control the quality of your data, but also vastly increase the speed of data processing right where you need it – at the edge.

https://gemdt.co.uk/the-engineering-hub/

*NOTE: EA technology is an invention that has its roots within the highest levels of academic excellence in Europe -- in development for more than 35 years. The invention is a specific new mathematical, combinatorial procedure that is referred to as “numerical balancing”. The technology has the ability to determine the exact structure of any data set without the presence of a pre-determined functional model, and simultaneously, calculate the “opposite face” of the structure as well -- resulting in perfect numerical equilibrium. While providing access to this new level of structural information, it is also important to understand that all EA procedures are carefully controlled by strict, primal-dual validation – meaning that every step in the process is numerically confirmed, so that the user can have the high confidence that only comes with absolute mathematical proofs.