ApolloQuality
Optimal Condition Monitoring
with
Apollo Next Generation Artificial Intelligence
The more expensive technical equipment is the more important is an optimal maintenance window. Late maintenance leads to heightened costs for servicing and in many instances to prolonged outage, too early maintenance to lower return on investment.
The innovative Apollo Technology of superWise Technologies AG opens up totally new opportunities for optimization, which leads to tremendous economic advantages throughout the whole spectrum of technical equipment.
To determine the optimal point in time is an extremely difficult task; there are various, often only approximately known, parameters to consider and to optimize: grade of wear and tear of various components of different costs, determination of wear and tear, risk analysis of deployment and of downtime and downtime costs etc. These parameters are – at least at the moment – only partly measurable with sensors.
There are two main challenges to optimization:
1. The availability of information about the equipment’s current condition. The innovative Apollo Technology empowers you to intelligently analyze arbitrary physical signals and signal patterns and to compare them to their desired values. In contrast to run-of-the-mill methodologies, Apollo can detect and process similarities in signal patterns, not just signals and patterns that are an exact match or are within a predetermined range.
2. Apollo can adjust and optimize multiple parameters simultaneously, taking manually entered set points into consideration to reach an individualized solution. Multiple fuzzy logic rules can be incorporated easily.
The Technology
Apollo is a next-generation artificial intelligence core product. Role models for its development are the complex functions of the human brain. Its mode of operation was emulated as precisely as possible – with astounding results.
Apollo enables applications without tedious parameter tweaking; also, there is no need for extensive training sessions – as is the case with classic neuronal-based systems. Apollo utilizes “on-the-job” training to learn the patterns presented; it is not necessary to physically define the components.
The example: data about a tunnel boring machine’s acoustic noise and its vibrations were collected: once of a smoothly running machine, and then of a machine close to its maintenance period. These two sets of data are sufficient to classify and evaluate arbitrary signals of this or comparable machines.
Markets
Applications for this intelligent condition monitoring are various and widespread: from construction mega machines to (aircraft) turbines to gearboxes to machine tools, and even to the just evolving concept of “Lifetime Conditioning Monitoring” in the automobile industry.
The latter is to be used during garage stops for routine maintenance and repairs to analyze a car’s visual condition; with Apollo’s visual image recognition capability, divergent conditions – be it of the interior or of the exterior – can be analyzed in great detail.
Background
Apollo is an immediately applicable, highly flexible intelligent kernel; more than 250 man-years went into its development. It’s also patented. Apollo’s unique design enables self-adaptive learning without supervision.
Applications using Apollo Technology are already commercially implemented (e.g. at Bosch, Olympus, Draeger, Volkswagen etc.).
To implement other applications often only requires the development of program and user interfaces, and the definition of patterns and/or the adoption of an applicable knowledge base. This leads to a rapid time to market, which nets the company – among others - reduced costs and extended uptime per application.
Apollo NEXT AI’s performance is extremely fast even when handling huge amounts of data – which eminently qualifies it for use in e.g. image processing and recognition.