Parker Aerospace writes innovative safety algorithm on Azure Databricks

Founded in 1917, Parker-Hannifin Corporation is a global leader in motion and control technologies. The Fortune 500 company has divisions dedicated to climate control, sealing and shielding, electromechanical, and perhaps most notably, aerospace. Parker’s Aerospace Group supplies aerospace parts to a diverse clientele, which includes all major commercial and military aircraft manufacturers and operators. As part of Parker Aerospace, the Digital Products team develops digital assets for lifecycle product traceability and models for predictive part performance. Data plays a crucial role in the team’s work and enables them to provide airline operators with actionable insights regarding part and system performance in real-world flight conditions.

Part of the Aerospace Group’s promise is to build parts that perform in all flying conditions. This requires insight into part performance while on-wing of the aircraft—something the Aerospace Group traditionally didn’t have. So, the company partnered with its longstanding customer, Republic Airways, to share operational data from the airline’s fleet. “We wanted to develop monitoring algorithms to improve system performance and reduce unplanned downtime for our customers. We needed flight data to do that,” says Jeff Smith, Head of Digital Product Programs at Parker Aerospace. The first step in the collaboration was to better understand the on-wing life cycles of the Parker Aerospace components used by the airline, a massive data-collection project encompassing more than 40 different products installed across a fleet of more than 220 aircraft.  

Gathering information at this scale meant a massive influx of data. While the partnership with Republic Airways promised to provide Parker Aerospace with never-before-seen access to part performance, the company lacked the infrastructure and personnel to process the data on-premises. “By the time the data was ready for analysis, it was already weeks old,” Smith explains. Parker Aerospace needed a solution to decode, cleanse, and append data to a table accessible to data engineers, scientists, and analysts. The company also needed a solution that could scale to meet demand and provide secure access to Republic Airway’s data.  

Flying in the cloud  

Parker Aerospace decided to adopt Azure Databricks for its ability to allow the team to not only leverage its current processing capabilities, but also seamlessly integrate into the company’s existing infrastructure and tools. “It’s been amazing,” says Jae Lim, Senior Principal Data Scientist for Parker Aerospace. “We’re new to cloud computing environments at this scale, and Microsoft staff were always there to help us out every step of the way,” explains Lim.  

One of Parker Aerospace’s earliest uses of Azure Databricks was to implement a monitoring solution known at this point as the “Leak Detection Algorithm.” Using the extensive flight data supplied by Republic Airways, the algorithm detects abnormal drops in hydraulic fluid quantity so that the airline’s maintenance team can be alerted to potential reliability issues before they affect operations. By monitoring daily changes and comparing them to historical data, the algorithm can identify deviations to help the airline avoid unplanned maintenance events, reduce cancellations or delays, and enhance overall efficiency. “Parker Aerospace performance monitoring exists because of Azure Databricks,” says Smith. 

The company is using Azure Databricks to completely transform its analytics capabilities. Republic Airways sends each day’s aircraft operational data to Parker Aerospace for analysis—a massive and constant influx of raw data. “Now, we can process large new data sets nightly,” explains Lim. By leveraging advanced analytics algorithms, Azure Databricks provides more accurate insights into aircraft performance and maintenance needs, which helps the airline optimize its operations and improve service delivery. Parker Aerospace has also integrated Power BI with Azure Databricks, which enables Parker Aerospace to visualize data in new ways and create report dashboards for Republic Airways. 

The partnership between Parker Aerospace and Republic Airways is mutually beneficial. In addition to supplying Parker Aerospace with insights, the Leak Detection Algorithm integrates seamlessly into an interactive predictive maintenance dashboard in Power BI powered via Azure Databricks, providing the airline’s maintenance team with insights into the performance of their fleet. This dashboard allows maintenance personnel to visualize and analyze data trends, enabling them to make informed decisions about maintenance priorities and resource allocation. The dashboard features predictive analytics capabilities through Azure Databricks, which can forecast potential maintenance issues based on historical data and enhance maintenance planning and efficiency. “Microsoft had all the necessary resources we needed to make the collaboration with Republic Airways possible,” says Austin Major, VP of Sales and Business Development at Parker Aerospace. 

Safer parts, faster  

Azure Databricks has now been fully deployed throughout Parker Aerospace, which hosts the software internally. The company has significantly reduced its processing time for new data. “Before using Databricks, the nightly processing of new data in a batch took over 14 hours,” says Lim. “With Azure Databricks, processing time has been reduced to less than two hours, with the results from the previous day available in the morning.” This reduction in processing time has granted the company timely access to critical insights from the data. “The auto-scaling feature of Azure Databricks allows us to use resources more efficiently,” says Lim.  

Parker Aerospace and Republic Airways have substantially enhanced their ability to make informed decisions and respond promptly to maintenance needs, while the faster processing time has improved overall operational agility. “As the industry realizes the impact and benefit of data analytics, data science and analytics will be an integral part of initial designs to realize business goals using data,” says Major. 

Reimagining aerospace  

Parker Aerospace plans to integrate more Microsoft services in the future, largely focusing on using sensor-rich data to develop prognostic models that predict part failures before they occur. The company intends to collaborate with other airlines to gather diverse aircraft sensor data for improved prognostics. “Solutions like Azure Databricks will play a critical role,” explains Smith. Once successful models are developed, they will be deployed on Azure to provide advanced warnings on various parts and systems, further enhancing airlines’ dispatch reliability. 

As the aerospace industry evolves, data science and analytics solutions like Azure Databricks will help bridge the gap between data and business values in a scalable analytics environment. By evolving along with the technology, Parker Aerospace will be able to continue innovating toward safer skies for years to come. 

Find out more about Parker Aerospace on Twitter, Facebook, and LinkedIn.