Big Data and Aircraft Health Management

Earlier the aviation industry had to depend on regular inspections and part replacements to ensure that aircrafts were safe for flight. To fully understand the implications for managing the health of an aircraft, an analyst must combine the data collected with complementary information–such as the current aircraft configuration, manufacturing history, service records, and other information that may help characterize the current operating performance of the aircraft. To take action, the analyst can add information about available service personnel, equipment, bay capacity, inventory, and logistics capabilities. Combining all of this information and understanding its impact create a huge information and technology challenge.

As the aviation industry adopts a proactive approach for maintenance, repair, and overhaul, it requires an architecture that captures data at the point of creation and moves it where it can be stored and analyzed. With it, the system can trigger an immediate response, support planning activities, or serve as the basis for new data driven offerings. Many companies have already started on this journey and have many of the pieces in place. We see three areas where companies face challenges in this model.

The capabilities of the information technology architecture include the following:

Sense:

On the aircraft, some of these sensors simply gather data about various components and operating environment and pass it on, but others can be programmed to store and analyze it. Ideally, this software can be updated over time to enhance the algorithms that process streaming data. Ground based business systems collect data about flight characteristics, service history, and ground operations.

Connect:

Sensor data must be routed to a central point on the aircraft and downloaded either in-flight or once the aircraft is on the ground. Limitations in bandwidth and connectivity affect when and how data is transmitted to central locations. As more types of data are gathered, companies require a scalable infrastructure to integrate and secure the data received from all the different devices and components.

Store:

When combining flight data including records from engineering, manufacturing, logistics, and service applications, the volume of data is overwhelming. 

Decisions have to be made about where it will be stored and how accessible the raw data will be to operational and analytic applications. Data warehouses can serve as consolidated repositories, but it creates challenges around latency, duplication, synchronization, and reconciliation. Engineered systems can provide the extreme performance required to handle these volumes of data.

Cleanse:

Designations for parts, customers, suppliers, and sites can be inconsistent depending on the source of the data. For example, a vendor or customer with the name ACME might be listed as ACME Parts or ACME Corp, making the data associated with that company inconsistent and inaccurate. With data cleansing capabilities, inconsistencies are eliminated and entities are consolidated, enabling the ability to build corporate hierarchies. Structured sensor data may have gaps, outliers, anomalies, or duplications that also need to be addressed.

Analyze:

Once clean data is available, analytic tools help identify trends, anomalies, relationships, and interactions among data elements. The results can be used to generate algorithms that enable real-time diagnostics and prognostics. Tools are often available for analyzing data from a single source, but challenges multiply when there are interrelationships among data from multiple sources.

React:

Raw data and the conclusions derived from them should be shared across the ecosystem. In the aviation industry, many parties have an interest in the raw data and the analysis that is generated. Security is important because much of the information is proprietary, particularly as value-added services are enabled through the use of data. Participants will want to limit access between competitors while providing visibility along the value chain.

Share:

Raw data and the conclusions derived from them should be shared across the ecosystem. In the aviation industry, many parties have an interest in the raw data and the analysis that is generated. Security is important because much of the information is proprietary, particularly as value-added services are enabled through the use of data. Participants will want to limit access between competitors while providing visibility along the value chain.

• Analytics is critical to understand causal relationships or generate responses based on the information available. Identifying the interactions across data sets is difficult because the analyst has to marry disparate data and understand the implications of the interrelationships. Once analytics identify where business value can be generated, a company can begin to prioritize their attention and investments in other components of the ecosystem.

• If the underlying data is not good, then the cliché “garbage in, garbage out” applies. Industry standards are being expanded to provide a common basis for data sharing. Work is required when industry participants use different designations for supplier codes, customer codes, and other critical fields required for analytics. A company must then determine if it should fix the master data or provide hubs to link common items that use different designations.

• Data architecture becomes important as the addition of sensors and inclusion of more business systems compound the amount of data available for analysis. When business systems with their own data repositories are included, it raises the question of whether data can remain in its legacy location or needs to be moved to a data warehouse. Replication of data with its associated lag times creates reconciliation challenges while it enables analytics that leverages cross pillar data. The best answer depends on the nature of the analytics, the latency requirements, and business challenges being addressed. With such advances happening in big data and business analytics, information technology is a differentiating factor in improving service levels and creating data-driven revenue streams for airlines and aircraft manufactures.