Every year, commercial and military aviation spends more than $90 billion to keep 70,000 planes flying, transporting customers, delivering freight, and flying security missions with a level of safety and dependability that is second-to-none among industries.
No one would argue that when it comes to flying, safety and reliability are the highest priority. However, executives who are responsible for maintaining aircraft or those whose operations depend on Tech Ops/MRO (maintenance, repair, overhaul), are frequently faced with many questions in a day-to-day environment. For example:
- As an aircraft operator, how do I know whether or not I'm getting the most out of the investments that I'm making in MRO?
- As a third-party MRO provider, how do I know what service commitments I should make to those operators who depend on my services?
- As an airframe or component manufacturer, how do I know where to find the greatest benefits to life cycle MRO improvements for my products?
These are difficult questions to answer. They're difficult, in part, because the assets have long lives, and the impact of maintenance practices plays out over this history. Much of maintenance is cyclic, with all different lengths of cycles and tremendous diversity of tasks, processes, and embedded data types. Consider, for example, that answering any of the questions cited above requires an implicit trade-off between different kinds of resources, such as aircraft time available, people required, and capital investments in parts, pools, and support facilities. And, that is just at the highest level.
More importantly, these questions are difficult to answer because every operator of aircraft places somewhat different demands on its maintenance organization. What is valuable to one operator may be worth much less to another one. Given changes in the markets and fleets and the life cycles of the airplanes, these are tough questions to answer because it's quite possible that today's answer might be different from tomorrow's.
What Is Maintenance?
Maintenance organizations that manage transportation assets face a triple challenge. They must keep mission critical assets available and on-schedule, do so with support people and parts spread out across a network, and respond to a wide spectrum of maintenance problems. The diversity of maintenance tasks is substantial. Maintenance can range from frequent, routine inspections to repairs completed on the flight line, and from complex troubleshooting involving the removal and replacement of components, to complete overhaul of components, engines, and even entire aircraft.
Aircraft maintenance is a combination of scheduled (planned) work and unscheduled (unplanned or non-routine) work. Unscheduled work comes out of discrepancies arising from aircraft operations. Discrepancies can be reported by pilots, discovered by mechanics during routine maintenance (called non-routines), or discovered by ground crews or other personnel. Because there is not always the time or capability to fix discrepancies at the moment and at the location they're discovered, maintenance personnel have a system of assigning criticality to these discoveries, which determines the mission capability of the aircraft. Both military and commercial operators have decision matrices that determine performance capability of aircraft in the presence of deficiencies (known in the military as fully or partially mission capable and known commercially as Minimum Equipment List and Deferred Maintenance Items).
Scheduled work can consist of anything from a routine inspection of a particular component occurring every x hours, to a complete rebuilding of an entire aircraft (occurring every five to 10 years), or any of a wide variety of activities in between. Scheduled maintenance occurs almost every day on aircraft, with somewhere between 50% and 75% of all work carried out on this basis. Scheduled work is carried out strictly in advance of the hours or cycles (landings) limits, unless engineering waivers have been obtained.
Aircraft are like people – they have long lives, and they age during the course of those lives. They also acquire new capabilities during their lives, and, as a result, have ever-changing maintenance requirements that are dependent not just on age, but on environment, levels of usage, even payload, pilots, and maintenance providers.
Aircraft are operated under widely varying circumstances, with varying missions, which introduce variables in how the aircraft perform and undergo wear and tear. All maintenance work ultimately involves people who examine the resulting maintenance needs, make troubleshooting decisions, and implement various repair strategies. Inevitably, when people are involved, there will be differences of approach, judgment, training, and execution of work.
The net results are that there are perennial differences between the planned work and actual work, and between the planned resources and the actual resources used. Every day, players in MRO make decisions that affect not just maintenance, but also operations and marketing in a delicate trade-off with multiple variables. Every one of those decisions leaves a trail of information from which a more intelligent MRO can evolve.
More Intelligent MRO
MRO is an inherently process driven business. When an aircraft is first acquired, it comes with a maintenance program, built around the design criteria and expected lifespan of the parts. From this maintenance program flows a set of maintenance planning activities staged out over the operating life, flowing down into scheduling and execution over the months, weeks, and days, constantly fine-tuning to adjust to the planned and unplanned events. This is illustrated in Figure 1 under the Process Management side of the diagram.
Historically, maintenance programs were heavily oriented around preventive maintenance, with components maintained and renewed to new status at regular intervals, based on estimated lifetimes and anticipated wear based on anticipated usage patterns. This was easy to manage, but resulted in substantial out-of-service time for the assets. It also resulted in unnecessary maintenance and even some induced maintenance.
As maintenance experience accumulated, MRO providers realized that there was a way to use the accumulated knowledge and information about the maintenance history of particular components. With more information, it was possible to create a less resource intensive maintenance discipline, which was not based on fixed time frames, instead more based on the reliability of the components in actual usage in actual conditions. This became known as reliability centered maintenance (RCM). The processes for moving from fixed schedule work toward an RCM approach are shown under the Analysis and Change Process side of Figure 1.
Under the RCM philosophy, maintenance actions are determined more by aggregate experience than by the original design criteria. Whereas RCM has become a significant discipline unto itself, it's now giving way to yet new approaches based on improved monitoring technology and maintenance technology. This emerging philosophy, which is termed predictive maintenance, rests on monitoring individual components in actual operation, anticipating requirements for maintenance, and executing this maintenance just in advance of critical points.
Maintenance is all about Maintenance Events
The foundation for moving from preventive to predictive maintenance is the availability of maintenance information – both historic, as well as current status. Advances in computing technology have led to greater availability of information gathering on equipment mechanical status, and greater capabilities for collecting and analyzing detailed history on maintenance needs. All this makes it possible to move into this maintenance philosophy.
The payoff for the incorporation of this information in the maintenance approach comes from the tremendous cost of maintenance and out-of-service assets. With high opportunity costs of out-of-service assets, high activity volumes, and so many specialized support resources, MRO organizations have considerable opportunities to find business improvement opportunities. The challenge that MRO organizations face is the ability to see the impact of their maintenance approach on the cost of MRO by correlating the impact on aircraft utilization, people, and other operating resources, with capital assets and, particularly parts inventory, all at the same time.
The key to managing these trade-offs is recognizing that maintenance is the management of a collection of maintenance events. Ultimately, all maintenance activity is about resolving these maintenance events, and it's specifically either scheduled or unscheduled events. Maintenance management requires capturing and using knowledge about maintenance events.
Scheduled events consume maintenance resources more predictably than unscheduled events, but this predictability comes at a cost – some of this maintenance will ultimately prove unnecessary (work that is performed in advance of when needed, if it is, indeed, ever needed). Scheduled maintenance is performed when the asset is least required for revenue service, and the resources, such as parts and labor, can be positioned ahead of time to be available in optimal fashion.
On the other hand, unscheduled maintenance means that maintenance is performed only when it's needed. But, it's more challenging to plan and may also have consequential operational impacts if it occurs at an inconvenient time (e.g., when the asset is most required for revenue service). Unscheduled maintenance requires planning to position parts and labor at locations where this unscheduled maintenance may be required. The process of anticipating this unscheduled maintenance depends on the quality of information available and the flexibility in addressing the unscheduled maintenance requirements. However it's done, it is likely to require many more resources positioned ahead of time than scheduled maintenance.
The less scheduled maintenance there is, the more detailed information about maintenance history, configurations of equipment or wear patterns, or operating circumstances is required. The further we move from preventive to RCM to predictive maintenance, the more we've moved away from aggregate use of information. Preventive maintenance establishes fixed repair intervals, and predictive considers the specifics of a particular situation, then anticipates and intervenes just before failure.
Since maintenance events consume resources, the information to find waste, identify underutilized resources, and find hidden asset time is buried in the individual maintenance events. To understand how to manage in this emerging environment, it's helpful to review the essence of MRO activity, which is process driven.
A 360-degree View of Maintenance Events
MRO organizations diligently capture detailed information at many different places in the MRO service chain (with each link of the chain storing its own data). Line mechanics, for example, keep the data about work required to support a particular piece of equipment; the work data are different from the failure data that are collected by engineers. They will again be at a different level of detail than the removals data that are kept by materials managers for maintaining inventory levels and dealings with vendors.
This approach, with each function keeping its own data for its own purposes, works well for supporting specialized workgroups. At the same time, it creates a situation that makes linking the maintenance event detailed data across the MRO chain very difficult. Often, they cannot analyze data across the maintenance service chain or correlate it across time or resource types to ask more detailed questions or probe for root causes. Why is this?
MRO is a process driven business, and each maintenance event creates a stream of maintenance transactions as it touches each aspect of MRO. These touches don't always need to occur in the same time sequence or at an event level (e.g., some of these maintenance data are aggregated for convenience). For example, every component removal results in overhaul activity, and every component removal also has an impact on inventory. This is not to say that the next item to be overhauled is the one just removed, or that inventory is adjusted every time there is a removal, but removals have a direct impact on both, and the exact way depends on all the circumstances associated with that removal.
The maintenance event cycle is a visual way of displaying the sequence of transactions generated by each and every maintenance event, as shown in Figure 2. The activities don't always occur in the precise counterclockwise sequence as they're laid out in the figure, but the MRO processes generally require these activities, which we will call the maintenance event cycle. See Figure 3 for brief descriptions of the transactions around the maintenance event cycle.
The maintenance event cycle can be used to define business improvement opportunities (BIOs) by grouping the various maintenance transactions into five major categories, according to the activity and outcome, as we illustrate in Figure 5. For example, moving counterclockwise on the maintenance event cycle, the first three tasks comprise a set of activities logically grouped because they directly affect operations, they must be decided quickly and generally involve time critical MRO coordination. To gain insight into this opportunity, which we call the Schedule/ Mission Readiness BIO, we can visualize the need to measure aircraft out-of-service hours, mechanical troubleshooting consistency, availability of parts, and effectiveness in completing assigned work.
We can gain specific insight into any one of the BIOs by scrutinizing the transactions within that part of the maintenance event cycle. However, at the same time, there's no doubt that there may be related, but hidden costs and benefits whose full impacts cannot be measured without considering all the other aspects of the maintenance event cycle. For example, aircraft out-of-service hours are dependent on the number of mechanical defects. These may be driven by overhaul or maintenance program practices or purchasing decisions made completely outside the scope of line maintenance activity.
This interdependence of the BIOs actually follows from the definition of the maintenance event: maintenance events kick off a sequence of transactions each of which is linked to each other in the process of maintaining aircraft. So, the full development of the Schedule/Mission Readiness BIO requires a 360-degree view of each maintenance event (i.e., all the way around the maintenance event cycle, capturing each interaction for defining its true business impact).
Driving Business Improvement – Answering Tough Questions
It is these interrelationships among BIOs that form the essence of the MRO analytic challenge. Tapping into MRO savings requires an information flow that parallels the physical linkage and supports asking the right questions. Performing the correct analyses must be supported by the correct maintenance event detail all the way around each part of the maintenance event cycle. An example will help illustrate this, starting again by examining the Schedule/ Mission Readiness BIO in more detail.
Tough Question #1: Am I keeping aircraft ready to fly efficiently?
The Schedule/Mission Readiness BIO deals with critical flight line maintenance, routine inspections and checks, and overnight maintenance of deferred items. The result of these activities is aircraft hours available for service, the minimization of delays and cancellations for maintenance reasons, and many lesser metrics that are critical for operational integrity.
The starting point for an efficiency analysis might be conventional metrics, such as dispatch reliability percent and labor hours used for maintenance tasks. With some basic figures for costs of flight delays and cost of labor hours, it would be possible to come up with a rudimentary efficiency assessment. Having trend line information and basic geographic and fleet breakdowns would provide the minimum information necessary for management.
But, when it comes time to look for improvement strategies, there may be complicating factors below the surface. If we look closely, one factor that may be affecting labor hours is the time mechanics spend waiting for parts. By another measure, we may find that troubleshooting occurs quickly when the parts are readily available. Therefore, can we make some inferences about the inefficiencies in maintaining airplanes?
With proper analyses, the answer might be quite surprising. It may turn out, for example, that the hours spent waiting for parts are for parts that are not stocked on account of inventory cost. This tradeoff for inventory savings versus labor costs has been made in a different department already.
It may be that troubleshooting goes quickly when the parts are available, but actually results in a lot of parts being removed and replaced that are subsequently determined to be "no-fault found". No-fault found (NFF) is a very significant industry problem that arises from troubleshooting practices that do not resolve the original problem.
The NFF problem points out several of the critical linkages between the Schedule/ Mission Readiness BIO and the maintenance event cycle. NFF affects the Reliability BIO by distorting the meantime between failures (MTBF) of a particular component. NFF affects overhaul by sending more parts back to the shop than are actually in need of repair. NFF affects logistics by shipping more parts back to the shop and then requiring them to be repositioned back to the line. And NFF affects inventory and purchasing optimization because while these parts are being removed unnecessarily, there are fewer parts on the shelf to support line operations. Activities that result in NFF may look efficient within the Schedule/Mission Readiness BIO, but present hidden costs in all the other BIOs. And there are examples that demonstrate the opposite impact as well.
Using the same logic, any business improvement opportunity will depend on activities in all other parts of the event cycle – and that's exactly what makes MRO such a challenging business to manage. We'll review some of the other tough questions that arise in the other BIOs precisely because of these deep interrelationships. Then, we'll propose an analytic environment for managing MRO data to capture the business value.
Tough Question #2: Have I made the correct reliability investments?
Reliability is all about time between failures and eliminating unproductive maintenance tasks. With thousands of components to monitor for reliability, sheer volume alone makes this one of the most difficult questions to answer.
But as hard as it is to collect basic data for reliability management, the calculation of the payoff is totally dependent on the maintenance event cycle. Reliability investments affect improved Schedule/ Mission readiness, some through flight line reliability and some through workload reductions. Part of the payoff is because there is less inventory to support fewer removals, and partly because of lower overhaul costs and related logistics costs. Every evaluation must use the correct cost and opportunity cost information, some of which may be sensitive to the maintenance context. And a related tough question will always be, "What was my expected payoff, and what is the actual payoff, when measured after the fact? Why might there be a difference? Is it because of assumptions or changes in operations or maintenance?"
Tough Question #3: Is my inventory working optimally for me?
Purchasing and inventory management is all about keeping capital cost down. It's hard enough to make provisioning decisions for complex assets, and even more so when the time frame is measured in decades and when the variables are likely to change many times in that time frame. But to really answer the question of whether inventory is working optimally, you must ask, "Is the inventory being used often enough to cover its costs?
"Which parts of the inventory were actually used to support schedule readiness? Is inventory available on the shelf when mechanics request it, or is most of the inventory time consumed in the shop or in transit? Do I know to sell inventory when it appears to be becoming surplus to usage patterns? Do I know what inventory to keep that may not be required now, but may be useful to protect against a later time when these parts may be out of production?"
Tough Question #4: Am I making the most effective use of overhaul?
Overhaul (remanufacturing) differs from manufacturing because of the variable nature of the work. The actual work performed must be compared to the anticipated or planned work. There are also decisions to be made about the levels of maintenance work to be completed, sometimes as it relates to minimizing life cycle costs, enhancing performance, or other discretionary work. Deciding what work to do, as well as how to measure the actual work accomplished is difficult enough, without asking the question whether it is actually effective.
The quality of work performed in overhaul intimately affects schedule readiness. The more comprehensive the reliability engineering and management of the maintenance plan, the fewer surprises there will be in overhaul. Overhaul is a big consumer of inventory, and any variances between anticipated and actual work will have impact not only on inventory but also on overhaul production rates and turnaround time. The greater the variability of overhaul work performed, the greater will be the impact on logistics, both inbound and outbound.
Tough Question #5: Is my logistics chain in tune with the requirements of maintenance events?
Aviation operates over a network, and maintenance needs occur all around this network. Supply chain visibility is exceptionally difficult to maintain in the presence of so many locations and so many different types of needs. This is especially true because of fast moving inventory and, frequently, urgent needs. Logistics is a delicate tradeoff between greater storage costs and greater distribution costs. The more storage locations, the easier the distribution decisions, and parts are more likely to be where they're needed. However, it makes for greater inventory investment, parts spend more time in transit, and there are other challenges associated with greater inventory, such as obsolescence, loss, or tracking inaccuracies. Keeping inventory concentrated at a central location makes it simpler to manage the logistics of getting parts to the exact location where they're needed. However, this might require expedited delivery at a much higher cost, and often it may mean that the parts are not immediately available where the most critical operational impacts are occurring. Decisions in the logistics chain affect trade-offs among aircraft availability, level of inventory, cost of expedited distribution, productivity of mechanics, and expensive assets waiting for parts.
Event-based Maintenance Management – Teradata's Robust Framework for Business Analysis
Maintenance activities that are discrete yet intimately interconnected through the maintenance event cycle – that is the essence of the MRO management challenge. While discrete BIOs can be identified, each BIO depends heavily on relationships with all other parts of the maintenance event cycle.
Driving business benefits in MRO organizations requires capitalizing on these interconnections, and first and foremost, that means being able to see the interconnections. In an environment where each of the discrete maintenance activities operates with its own data, the visibility to the connections is greatly reduced. Myriad specialized MRO functions each generate different data streams amounting collectively to millions of maintenance transactions annually per aircraft. There are different levels of data timeliness, history, and detail required by different MRO functions. Functional roles as different as line mechanics, reliability engineers, and purchasing agents have traditionally demanded purpose-built information islands, each with their own business rules. This falls short when it comes to addressing the interdependent nature of the business.
Visibility to detailed maintenance event data is similar to the effect a prism has in front of a beam of white light. Without event detail, the aggregate collection of data in any one of the business improvement areas produces the equivalent of white light. While efficiency and basic metrics can be calculated with aggregate data for any of the BIO areas, there is very little insight as to the drivers behind the performance. And there is no insight at all into the benefits that might originate in business areas outside of the one at which you're looking.
Teradata MRO customers recognize that the highly interconnected nature of specialized MRO functions means an interdependent data stream as well. They know that a part removed on the flight line is the same one inventoried by the materials manager, and the one sitting in shop overhaul. Maintenance managers seeking to understand their business have come to know that business analysis complexity is reduced, and business improvement decisions made faster, when the data mirror this simple fact.
Teradata MRO customers have also prepared for a future where demand on their data infrastructure is constantly growing due to new data types, business changes, and more urgent queries. New generations of more complex aircraft introduce new data types; old aircraft are upgraded with new configurations; interfaces with outsourced MRO providers demand cleaner data exchange on parts history and maintenance support; and global operations compel higher levels of mission readiness. By making a commitment to storing event-level maintenance data, exactly as the business creates it, Teradata MRO customers preserve their options for new analyses as business needs emerge.
Teradata MRO customers have come to know that they need an analytic environment that reflects the maintenance event cycle. Such an analytic environment preserves the discrete details of each individual maintenance event and its connections to all other maintenance transactions. An MRO analytic environment of necessity needs to be a single environment. This is in stark contrast to multiple data stores that have aggregated away some of the detail, have eliminated some of the connections, or have otherwise failed to preserve some of the details associated with the need for maintenance. An analytic environment needs to be a single way for users to access the data, regardless of which specialized MRO functions they are a part. An analytic environment means that the same answer will be obtained for the same question regardless of who within the MRO organization asks it. Ultimately an MRO analytic environment will also become a collaborative platform to share between suppliers, airframe manufacturer, and operator.
Tighter linkages between MRO transactions and MRO suppliers and maintainers will facilitate further movement toward predictive maintenance. Data streaming for condition based maintenance (CBM), data mining routines searching for failure patterns and alerting on scheduled tasks will cull out waste associated with unnecessary maintenance and reduce the opportunity costs of resources. The differentiator in such an analytical environment is the ability to perform ad hoc queries that target a mix and match of streaming data with detailed history and context data.
Such a streaming data and analytic environment is already a reality for selected parts of the aircraft, with engines being a good example. As we gain experience with such environments, and learn to build queries based on current and historical data, the MRO analytic environment will come to encompass something that Teradata calls active enterprise intelligence. Active enterprise intelligence presumes the near-real-time availability of all the data needed for analytic queries, in context with any supporting historical data and experience. This is a significant enhancement of MRO that extends the power of the ERP environment. It effectively enhances each and every decision in the process environment with analytic queries that moves closer and closer to providing predictive maintenance.
Today, MRO providers struggle to balance competing business needs of high aircraft utilization, keeping down investment in parts inventories and supporting assets, and high productivity. Detecting and stopping problem trends in reliability, anticipating impending parts supply shortages, and pinpointing productivity issues, before they cost millions, requires putting your data to work.
What the business sees as an integrated flow of tasks, oftentimes supported by ERP systems, the analytic environment should see as an integrated set of data about the impact of maintenance events on the MRO enterprise. MRO organizations have come to understand the need for integrated process flows as a way to streamline the handoff of tasks among myriad specialized MRO functions. MRO organizations now must see the need for event-level detail as being essential to tap into the potential MRO business benefits. Extracting business improvement potential from an MRO process requires knowledge of many more interrelationships, a much harder job than just keeping the MRO process running. The preservation of detail is first and foremost among MRO data requirements.
When detailed maintenance event data are readily available, the benefit of potential business improvements becomes plainly analyzable in the data. With event-level visibility, the MRO business can move steadily from correcting problems to anticipating problems – by better understanding the precise circumstances that lead to maintenance needs, and using active enterprise intelligence to eliminate them just in time. Teradata has the power to deliver a holistic view of detailed maintenance data to maintenance practitioners – providing visibility over both time and organizational boundaries to put data to work to achieve the goal of high aircraft readiness at a lower cost.
About the Author
Peeter Kivestu is Director, Global Industry Solutions, Travel, Transportation, and Government for Teradata Corporation. He is responsible for driving operations management initiatives, including technical operations, maintenance, repair, and overhaul in aviation and transportation.
Peeter joined Teradata in 2004 with more than 27 years of airline, information technology, and supply chain expertise. His work has spanned most of the airline major operating functions, including finance, marketing, operations, information technology, corporate development, and cargo.