Sentient Science’s DigitalClone software has been extending component life within the wind energy and military aviation industries for over a decade, and now the company is ready to bring its long-term forecasting functionality to commercial aviation and MRO.
The software, which uses physics-based modeling and materials science to predict the life expectancy of complex mechanical systems, was initially developed in 2001 through the U.S. government’s Small Business Innovation Research (SBIR) program. The program provides funding to small businesses for development of technology with both military and commercial applications. Sentient Science has been working with Department of Defense programs to apply its software technology to the long-term prediction of mechanical system failures on fixed-wing aircraft and rotorcraft, such as the Boeing AH-64 Apache and Sikorsky UH-60 Black Hawk.
According to Jason Rios, VP of Aerospace at Sentient Science, DigitalClone’s approach differs from the other predictive big data solutions in the industry. “It’s based more on a fundamental understanding of the materials that make up components and how those materials react when the components are put under operational loads and stresses—because those components break when their materials fail,” says Rios. “So to be able to understand when those materials are going to fail, that’s where our materials science expertise comes into play.”
Rios says the analogy Sentient Science likes to use is that of predicting and preventing heart attacks. The medical technology has evolved from hooking somebody up to an EKG to see if they are actually experiencing a heart attack to a more predictive method of collecting information about factors such as blood pressure, cholesterol and body fat index to look for factors consistent with heart attack risk—which Rios says is more of a big data-like approach. “It also depends on those measurements being outside the norm before they can really offer a good indication, kind of like waiting for a vibration sensor on an aircraft to show an indication,” he adds.
The next step, he says, is approaching someone in their 20’s and analyzing their lifestyle to see if there are factors that could lead to future heart failure and helping them change things now to extend life expectancy. “It’s kind of the same principle that we apply to mechanical systems—to be able to identify, based on the operating profile of those systems, ways that our customers can prolong the life of those systems before any of those failures start to occur. We’re looking a bit further out,” says Rios. “We’re looking from the perspective that we’d like to give our customers information before their systems are starting to get into those early stages of failure.”
DigitalClone creates a model of a mechanical component’s design based on factors such as base materials and the types of manufacturing processes and lubricants used. The software is able to identify where critical parts of the system are located and applies high-performance computing resources to run thousands of advanced simulations looking at how materials will respond to stresses and how much stress they can take before they start to initiate failures. According to Rios, DigitalClone is designed to accept the vast amount of data collected from aircraft and system usage, which feeds into its models. This allows the software to analyze an asset’s usage profile and make adjustments to its life expectancy based on the actual usage conditions it is seeing.
According to Sentient Science, DigitalClone’s long-term forecasting can predict failures much earlier than current monitoring systems, which enables enhanced maintenance planning and provides benefits in terms of design and supply chain optimization. “In our previous development programs with the government, we’ve seen a 30-40% reduction in the cost and a 50-75% reduction in the schedule from what they would normally expect for design optimization for new designs,” says Rios. “We’re able to allow them to iterate on those designs faster—to be able to look at the failure modes associated with the designs to determine whether the initial design will be robust enough. [This] allows them to get to prototyping faster than the conventional approach.”
Operationally, Rios says DigitalClone’s benefits are mainly related to supply chain optimization. “Our technology enables the capability to look 12-24 months out to be able to see gearbox failures in the fleet. That allows the supply chain to do more effective demand planning,” he says. “With all the safety stock that these operators keep because of the uncertainty related to when those items are going to fail, if they can be more confident that they have better demand signals, it allows them the flexibility to reduce a lot of that safety stock,” he adds.
The company has been looking into applying DigitalClone’s functionality to additive manufacturing. “There’s a lot of investment and development activity going on within industry around how to leverage that technology for something like aerospace,” says Rios. “With aerospace, the key consideration is qualifying the parts for airworthiness using that advanced manufacturing technique. So we’ve been pretty heavily engaged in that part of the industry over the past three years, mainly through the SIBR program.”
Sentient Science says it is currently working with the U.S. Army to put together a program to validate DigitalClone’s technology for both government and aerospace industry applications. Although it is not yet able to publicly share information about specific commercial aerospace customers, Sentient Science says it is working with OEMs and operators to deploy the software internally across their fleets. Rios says DigitalClone is expected to be ready in its final form within the next six-to-eight months, which will be followed by a validation process. He estimates the software will be ready for commercial release within two years.
The company is still doing market research with its aerospace customer base to figure out the appropriate price point, but DigitalClone will be offered as a cloud-based software-as-a-service subscription model as it is currently offered within the wind industry. Rios says the model will be based on price per asset.