Development of a system for aircraft engines condition monitoring and preventing failures

Development of a system for aircraft engines condition monitoring and preventing failures

The system 

Client delivers several types of aircraft engines for civil planes which collectively fly millions of miles every day. A grounded aircraft causes losses – the failure of one machine affects over 20 other flights, impacting thousands of passengers. The costs for airlines amount to millions of dollars.  

Client, in care of reliability of engines, develops a solution allowing for collection, monitoring and processing of data related to the state of engines (among others, referring to engines states during take-off, flight, landing, and correlation between weather conditions and operation of the engines), that allow maintenance teams to react well in advance. Every few minutes, millions of parameters are processed. By monitoring the condition of the engines, the client has saved millions of dollars. 

The need 

• Improving scalability – for the on-premises system in use, a major challenge proved to be the scalability in terms of the increasing volume of data to be processed, which translates into the excessive costs of maintaining physical infrastructure. The client would like to rewrite the on-premises system and move it to the cloud, which will allow the use of new telecom mediation technologies. We are developing a solution to replace the current one.  
• Minimizing carbon footprint – the client aims to become a zero-net carbon company in the next few years. The customer collects data on fumes and air pollution. Machine monitoring supports using them to minimize the carbon footprint.  
• Maximizing time-on-wing – the company made total care solution available to its clients. The provider makes profits when aircraft with its engines fly. For this reason, maximizing time-on-wing is one of the key goals.  
• Optimizing usage of Azure services – in 2021, the solution was one of the biggest projects on the Azure platform in Europe. The cost of infrastructure maintenance on the Azure platform accounts for several millions of dollars every year. The client’s goal is to optimize costs through skillful management of Azure cloud services, tailored to current business needs.  

Our tasks 

The client’s system is used by engineers and service teams but also by end customers (airlines, and aircraft manufacturers). Currently, 70% of work is focused on the development of the new functionalities, and 30% on bug fixing and working on technology debt. In the project, we use Agile models of work depending on the needs of teams (Scrum and Kanban). Our specialists have freedom in the way solutions are implemented, making sure that architectural principles are met, as agreed with architectural teams. Determining the scope of work is carried out based on the client strategic roadmaps and definitions of quarterly goals which streamlines the prioritization process.  

Data analysis 

We provide solutions to process data from multiple sources, organize it, verify, and ensure that data is not lost in the event of processing failure on the platform. Data processing allows the generation of various types of reports, provides information on errors and recommendations, thanks to which customers’ engineers and operators can react (very often, well in advance). Aircraft engines are the safest among others thanks to such mechanisms working in the background.  

Digital Twins  

One of the solution cornerstones is the Digital Twins approach. These are defined and built digital components of a model representing the engine. Their behavior is then analyzed, considering collected information about the performance values of the real engines and their operating conditions. Based on the simulation data, it is possible to predict what might happen, or optimize the engine’s performance, e.g., through software updates. Such analyses make it possible, on the one hand, to optimize engine operating costs and, on the other hand, to verify with high accuracy the wear and tear of individual components. Fewer defects and visits to garages mean fewer canceled flights, and thus less problems in air traffic.  

Intelligent Engine 

Clients’ engines can send data on a huge number of parameters. The newer the engine is, the bigger the number of parameters. Engines may send data of various volumes and with various frequencies. Multiplying this by the number of aircraft and engines in each result in big data volumes. These need to be processed in a strict timeframe. To meet the requirements, we are taking advantage of the scaling capabilities of the Azure environment, components and techniques related to Big Data processing and IoT integrations with physical devices installed in an aircraft. 

Why Inetum?  

1. Flexibility – client decided to cooperate with an external IT provider because of the need for human resources flexibility and scalability (so that to focus on business development).  
2. Quality Assurance process – we presented a well-described Quality Assurance process that was difficult to implement on the client’s platform.  
3. Competences – the client gained our software development expertise.  
4. Cost – the cost of software development services was also important (lower than when working with engineers from the US or Western Europe). 

Challenges 

• Efficient collection and processing of high frequency data, the number of which will increase with the number of machines.  
• Ensuring consistency and availability of data. Transmitted data might get lost at different stages, or may not be visible (e.g., during data transfer between an aircraft and the Azure platform). 
• Ensuring service consistency / quick disaster recovery in the event of infrastructure or cloud region failure – when some of the Azure components or cloud regions stop working, data and service recovery become a challenge. Azure cloud provides such capabilities but to fully address the need of Client, we developed a dedicated solution for manual initiation of switching between Microsoft Azure regions (user-initiated failover/failback). This is dictated by cost optimization considerations related to maintaining some expensive resources in the backup region.  
• Maintaining domain knowledge with an increasing project scope. 

Plans 

Client trusts in Inetum specialists’ competences. So far, we have been mostly delivering new functionalities (feature-driven development). The new cooperation model, planned for 2024, assumes that Inetum will mostly provide Managed Services and optimize Azure services costs. The project for client started in 2013 and includes long-term cooperation. 

Team  

120 Inetum specialists work in teams of several dozen people. The project involves automation and manual QA, Python and .NET developers, Scrum Masters and entire DevOps and Business Intelligence teams, technical leaders, architects, PMs, test manager, and account manager. 

Technologies 

Microservices (for a quarter), Microsoft Azure (App Services, SQL databases, virtual machines, Azure Storage, Event Hub Namespaces, Service Bus Namespaces, Application Insights, Data Factories, Azure Cache for Redis, Azure Cosmos DB, IoT Hubs, Azure Databricks Services, Service Fabric Clusters, Azure Batch Account), Python, .NET