WHITE PAPER: AI is Powering Growth of Lifecycle Optimization Tools

Author: Dr Ip-Shing Fan, John Maggiore and Professor Anna Smallwood, all at Cranfield University

Subscribe

Dr Ip-Shing Fan, John Maggiore and Professor Anna Smallwood, all at Cranfield University, consider how AI will be applied in asset management for aviation

There is a growing area of Digital Aviation aftermarket solutions which leverage AI, data analytics and other technologies to optimize new and evolving use cases. These can be described as Lifecycle Optimization tools. This article explores these new use cases, and the various creative ways AI and other technologies are applied to solve targeted problems in the market today.

AN EMERGING THEME IN DIGITAL AVIATION – ‘LIFECYCLE OPTIMIZATION’

As mentioned in the Spring edition[1] of this publication, we can describe ‘digital aviation’ in a general sense as: knowledge or insights embodied and delivered via software on, at, or about an aircraft to deliver value to aircraft operations. Operations can be divided into major verticals, including Flight Operations, Technical Operations, Ground Operations and Passenger Operations. Focusing on the Technical Operations vertical, we can categorize the aftermarket digital offerings into several key capability themes as shown in figure 1 below.

Figure 1

Health Management refers to proactive and predictive maintenance methods using aircraft data to make informed maintenance decisions to drive operational efficiency. Maintenance Planning refers to the family of capabilities focused on scheduling and tracking maintenance activities, and the ‘system of record’ for these actions. Electronic Logbook refers to the family of capabilities in which PIREPs (Pilot Reports) and Maintenance Tech Log data are collected and managed electronically to both drive operational efficiencies and to collect key data in a codified manner to ease use of these for analytic analysis. Electronic Documentation refers to the electronic delivery of regulatory-approved maintenance documentation to users, both in the back office and at point-of-use via mobile distribution and delivery. Digital Records refers to the family of tools focused on the collection, typically via scanning and Optical Character Recognition (OCR), of paper digital maintenance records to facilitate digital storage and indexing to increase operational efficiency, to ease the transfer to aircraft and asset ownership, and to maintain the residual value of these, especially for leased equipment.

This article focuses on Lifecycle Optimization, an area with much innovation and new and evolving use-cases.

Here we focus on newer offerings in the market to illustrate the applications, and also discuss how the core underlying technologies are leveraged to deliver the respective decision support information to the different user personas. Example Lifecycle Optimization use cases covered in this paper:

• Optimization of Assets over the Lifecycle;
• Materials Procurement over the Lifecycle;
• Data transparency and portability;
• Reduction of fraudulent parts;
• Back-to-birth parts provenance.

OPTIMIZATION OF ASSETS OVER THE LIFECYCLE

Within the context of asset optimization, assets may be any economically significant rotable parts or subassembly for which economic decisions are made. Engines are the classic example, but assets can include landing gear and other major assemblies. Related decisions can include decisions regarding overhaul, leasing, maintenance plans or scrap/retirement. Financial decisions include choosing between Time and Material maintenance plan vs Power-by-the-Hour (PBH) plans. Data inputs into this process include a mixture of digitized technical records and contractual documents.

KeepFlying[2], based in Singapore and started in 2020, has taken an AI-driven approach to optimizing the economics of assets in MRO. Specifically, KeepFlying leverages generative and predictive AI, Large Language Models (LLMs) and interactive chat bots to help users glean otherwise difficult insights into their aircraft or assets such as engines. Users include airlines, MROs, and Lessors. With the right data sources and adeptly applied AI methods, there are a myriad of potential questions that can be answered. A partial list includes Predicted Scrap Rates, Shop Visit Profiles, Work Scope Predictions, Maintenance vs Fuel Consumption trades, Slot Profitability Forecaster, and asset/aircraft Transition economics monitoring. The overall KeepFlying approach is to create a digital twin of these use cases, referred to as a FinTwin^TM (i.e., Financial Twin), thus unlocking the decision-making analysis in an automated and streamlined way that would not be technically or economically viable to execute, otherwise, especially given the current global labor demand for skilled technical resources. The goal is to create an artificially intelligent aviation platform that becomes a virtual Technical Services Engineer capable of collating huge amounts of technical data and presenting its findings to the user.

According to KeepFlying, the potential to materially optimize these assets is significant. For example, airlines spend over $2 Million USD per Narrow Body Redelivery, making return on investment of even incremental efficiencies pay off. In addition, Engine MROs often miss out on over $0.25 million USD in bottom line profit per engine overhaul by not executing the desired optimization.

Says, Sriram Haran, Founder and CEO of KeepFlying, “We believe the real power of LLMs and their Generative AI counterparts is extracting the best parts of these technologies and unifying them into a practical system that addresses many functional and profit-delivering use cases. Consider a system that can utilize hundreds or even thousands of bots simultaneously to achieve a common goal, understanding user intent and the business rules that exist to automatically complete workflow to make technical and economic decision making truly optimized and scalable. A combination of Generative AI powered by Explainable AI; that’s our goal.”

ADDRESSING THE PROCUREMENT LIFECYCLE WITH AI

With approximately $60 billion spent annually on material in the commercial aviation MRO industry, the potential payback on any efficiencies gained via optimization is vast. Moreover, the aviation supply chain is undergoing new challenges which further challenge the material environment. As Jonas Murby, Principal at AeroDynamic Advisory[3], shared at the 2023 IATA Maintenance Cost Conference, “MROs are at capacity with full shops and long Turn-Around-Times for parts, and are forced to prioritize long term contracts over emergent, ad hoc business. Airlines are seeing groundings due to asset availability constraints, and overstretched purchasing departments. Digitization will be a key underpinning to make the supply chain more efficient and resilient.”

Upon recognizing the vast opportunity to provide procurement optimization in the industry in 2017, SkySelect made a pivot from a broader supply chain SaaS offering to focus on Aviation. They are focused on providing repeatable and scalable procurement via their Procurement AI platform.^[2]. SkySelect uses a real-time demand signal from their customers’ M&E System to initiate a process which automates and streamlines the procurement cycle (figure 2). The objective is to make the procurement process accurate, scalable, and efficient, while aligning with the customer’s business rules and objectives. Customers include airlines and MROs, each with their respective business priorities. Moreover, their Procurement AI platform is geared to serve the needs of both parts — buyers and supplier. Specifically, they leverage Machine Learning (ML) trained against very large numbers of acquisition data and interchangeability documentation. SkySelect focuses on automating the procurement value chain, from demand, sourcing, purchase order, ordering, delivery tracking and invoicing. From the demand signal from the maintenance planning system, SkySelect’s AI provides favorable sourcing options and suggests the best optimum between the price of materials and logistics cost (by optimizing for the lowest total cost of ownership).

Figure 2

SkySelect has informally coined a new expression to describe their AI approach: Large Consumption Models. That is, training their AI on what parts airlines and MROs are consuming.

According to Erkki Brakmann, Skyselect Co-founder and CEO, “For our customers it’s not so much about the technology as about getting a highly automated and scalable process with real-time visibility into the supply chain. For example, an MRO customer was able to cut the purchasing turnaround time, from a part requirement to placing an order, from five days to a few hours. At the same time the buyers are able to process 4-5 times as many part requirements as before. Additionally, they increased their on-time delivery performance of parts by 15%. We think these kinds of results speak for themselves”.

ACCELERATING DATA TRANSPARENCY AND REDUCING FRAUDULENT PARTS IN MRO WITH BLOCKCHAIN

Data ownership and data use rights remain a complex issue stirring much debate. Data is the foundation of all digital aviation endeavors. Moreover, in most efforts, data must be shared between companies. Companies provide data to other companies who they believe will protect their data and provide a valuable work product as a result. Supportive of this goal, the Independent Data Consortium for Aviation[4] (IDCA) was formed in 2023 to provide consensus guidance for data sharing around specific use cases: back-to-birth to disposal, AOG servicing, diagnostics and prognostics, leasing, and airworthiness in crisis.

SkyThread[5] was formed in 2021, in part to address the practicalities of data sharing in the real world. The primary objective was to use technology and an ecosystem business model to optimize data flow via a secure, neutral, and trusted digital network. Each provider of data has their own business rules and objectives which they wish to achieve. In an ideal world, each agent in the data transaction can control what data is seen by who and apply their unique analytics to the transaction.

SkyThread designed their platform to do this, and to provide new levels of visibility to parts status and condition data, and assets they are installed on. This is done through a blockchain-enabled, Utility Layer, focused on the intersection of Planes, Parts, People and Places^[2]. Plane defines an aircraft and catalogues its history of parts configurations from an aircraft’s birth through to its decommissioning. The layer creates a back-to-birth (BtB) record for all parts on existing aircraft and captures induction data for all newly produced aircraft parts. It also defines an aircraft part and records its condition throughout its service life (i.e. back-to-birth records for all existing aircraft parts). It defines technicians (people) and all those who are eligible to certify aircraft and part repairs, and certifications of airworthiness; as well as using geo tracking, locations (places) where the aircraft and parts are produced, printed, assembled, and repaired at any given time. 

Another key driver of enhanced data management and control is to provide the ability to create a systematic record of parts for the purposes of parts traceability, an increasingly important topic to address the problem of fraudulent parts. As Dr Chris Markou, head of technical operations at IATA recently stated^[3], “Creating an immutable ledger to ensure the integrity of information regarding parts is vital for accident and incident investigations and targeted recalls. However, this has to be scalable and easily performed. Using digital technology such as block chain to create the needed records with the required detail and accuracy makes sense”.

The industry has a group of startups that have been formed specifically to make things better and easier in the aircraft parts supply chain. While forming their chain, SkyThread realized that many industrial giants, digital firms, and others would build individual solutions to tackle one or more aspects of the business problem. Chuck Marx, Chief Strategy Officer for SkyThread notes “We built our chain to collaborate with all industry efforts to collect part ID data. This is called our chain of chains. All data are welcome as we tackle this industry issue around data sharing.”

AUTOMATING BACK-TO-BIRTH PARTS PROVIDENCE USING AI

Sales of Used Serviceable Material (USM) has increased in recent years, both driven by cost efficiencies, and to provide a more sustainable option for replacing aircraft parts without manufacturing new components. There are standards that provide recommended practices for the traceability of civil aircraft life-limited parts (LLPs), such as SAE ARP6943[6] (Component Traceability Requirements for Life-Limited Parts), which applies to landing gear. Flight cycles, flight hours, and calendar time is described, along with back-to-birth traceability, to ensure airworthiness of service LLPs.

PrôvenAir[7] has created an AI-driven capability to apply these rules in a real-life setting. Jim Boccarossa, PrôvenAir CEO, explained that this was inspired by the manual nature of compiling the needed BtB records that drove them to look to AI to provide a scalable and repeatable method which also saved on labor investment. ProvenAir provides a SaaS platform that generates consistent back-to-birth trace insights for commercial landing gear and other life limited material. Using ProvenAir, tasks that took weeks can be done in hours.

Their approach system uses proprietary algorithms and artificial intelligence to scan and categorize maintenance records, interpret life limited part (LLP) usage and create trace timelines and exception reports for engines, landing gears and APUs. Says Bocarossa, “We learned firsthand, through other commercial endeavors, that the task of compiling BtB records in a repeatable and efficient manner with a minimum of errors was frankly beyond the capabilities of human actors. While technology for technology’s sake does not work, compiling BtB provenance of complex assets is an excellent example of using AI technology to solve a real-world problem, today.”

SUMMARY

As we think about Lifecycle Optimization within the larger digital aviation tech ops ecosystem, we see that it retains the common thread of providing economic decision support capabilities. Uniquely, it is performed solely in the back office, by several newer digital aviation user personas, rather than on or at the aircraft. As with other digital aviation endeavors, there are direct, and second order efficiency driven benefits such as airline operational performance and sustainability results which flow from Lifecycle Optimization activities[8].

We can see that growth in Lifecycle Optimization is being accelerated by several factors, such as the always present imperative to increase airline and MRO profitability, but also emergent circumstances such as the tightened aviation labour market and supply chain constraints. Added to this are the democratization of data via new exchange and sharing frameworks, and ubiquitous development and adoption of step change technologies like AI which also serve to increase acceleration.

The example Lifecycle Optimization solutions highlighted in this article are powerful tools which leverage AI and other technologies via different approaches for different use cases. As with all digital aviation tools they generate revenue and reduce cost of operations via efficiency gains. While addressing important aspects of maintenance and supply chain decisions in the lifecycle of an aircraft, we observe that, as of yet, there is no comprehensive lifecycle model available. Airline and MRO users must integrate these tools into their unique business and operations and make trades on investing in potential integrations between tools and data sources. The notion of Lifecycle Optimization that optimizes decisions based on a complete model of asset and fleet lifecycle, necessitates rethinking in the industry. To help fill this gap, Cranfield University is developing a virtual Cranfield Airline model to conceptualize the integration of operations, engineering and spares management decision making, and a simulation to evaluate through life impact of the different options.

In summary, we foresee continuation of investment and innovation in the theme of Lifecycle Optimization, with new use cases, expanded user personas and new offerings from existing and start-up companies.

CRANFIELD UNIVERSITY DIGITAL AVIATION RESEARCH TECHNOLOGY CENTRE (DARTEC)

The Digital Aviation Research Technology Centre (DARTeC), located at Cranfield University, the number one university in Europe for aerospace, is a world-class center for the research and development of cross-sector digital integration solutions. DARTeC, co-funded by Research England, an industry consortium of leading aviation organizations and Cranfield University, is a significant investment in state-of-the-art facilities that leverages both the University’s Global Research Airport and its Autonomous Vehicle research facility. DARTeC focuses, within an organizational resilience framework, on five primary research challenges that individually have direct relevance to the digital agenda: Connected systems, Unmanned traffic management, Seamless passenger experience, Distributed airport/airspace management and Conscious aircraft.

Collectively these provide the opportunity to explore and address systems integration challenges through advances in technology, intelligence, regulatory frameworks, and business models. For more information on DARTeC contact Anna Smallwood (Anna.M.Smallwood@cranfield.ac.uk).

FOOTNOTES

[1] https://issuu.com/aircraftit/docs/aircraft_it_mro_v13.1/58

[1] https://www.keepflying.aero/

[1] https://aerodynamicadvisory.com

[1] https://www.dataforaviation.org/

[1] https://www.skythread.aero/

[1] https://www.sae.org/standards/content/arp6943/

[1] https://www.provenair.com/

[1] https://issuu.com/aircraftit/docs/aircraft_it_mro_v11.1/52

---