The acronym MRO (maintenance, repair and overhaul) has long been part of the aerospace vernacular. But only recently have service providers adopted automated tools to streamline the upfront process of capturing and digitally reconstructing physical parts that require repairs or redesign.
The central issues surrounding MRO—speed of repair, accuracy, reliability and quality—are relevant worldwide. In the United States, it is a central issue with adherence to FAA parts manufacturer approval (PMA); similar certification processes are in place throughout Europe and Asia. In all cases, the combination of market demand and certification make MRO a prime application for 3-D point-cloud processing.
Increasingly, MRO starts with 3-D point-cloud processing: digitally capturing physical objects and automatically creating accurate 3-D models for downstream design, engineering analysis, inspection, and custom manufacturing.
Applications for MRO
Point-cloud processing is suitable for MRO projects for four reasons:
• It captures and reconstructs parts as they exist, enabling engineers to analyze damage and its underlying causes.
• It enables engineers to accurately re-create parts for which no computer-aided design (CAD) data or documentation are available.
• It enables engineers to analyze situations where every part is different, due to factors such as manufacturing variability and tolerances, wear patterns, and stress.
• It improves asset management by speeding part reconstruction and inspection of as-built vs. as-designed parts, helping aircraft to return to flight much sooner.
Unlike CAD, 3-D point-cloud processing does not require an existing part or assembly to be created again from a blank screen.
Design, engineering, and quality inspection processes spring from what already exists in the physical world.
Let’s look at a few examples of MRO applications, with an eye on how 3-D point-cloud processing could be applied to speed the process and ensure quality.
Commercializing a new part—A company wants to commercialize a new part, but the original equipment manufacturer (OEM) does not release drawings or CAD models. The company needs to be able to capture the part, digitally re-create it, and send data to a CAD program. From there, a model can be created from which the newly fabricated parts can be manufactured with the same form, fit, and performance of the OEM parts.
Structural check—A cargo plane from a shipping company goes in for a maintenance check, often called a “C-check.” As part of the maintenance, the airframe must be checked for cracks and possible repairs. This requires comparing the actual airframe against the nominal CAD geometry to identify areas that are out of tolerance.
Engineers need a way to perform a complete 3-D analysis of all the surfaces, quickly compare them against the CAD model, and identify out-of-tolerance areas for repairs or part replacement. The shipping company must minimize downtime and get the airplane back in service as soon as possible.
Conversion from passenger to cargo plane—A company has been contracted to convert a passenger plane into a cargo plane. It does not have access to the original passenger plane drawings, and because the plane was first built during the late 1960s, there are no CAD models associated with it.
The entire existing environment, including electronic cables, hydraulic systems, and other internal aspects, need to be captured so that engineers can design the new cargo areas. Engineers have four months to remove all passenger equipment, install the main cargo door, and integrate the new cargo system.
Fixing a crack in a fighter plane—A fighter plane undergoes routine maintenance, and a crack in the structure is discovered. Repair requires a stainless-steel doubler fitting that is riveted into place, bridging the gap caused by the crack. Obviously, there is no CAD model of the crack.
The traditional repair process requires that an impression of the crack be made with dental putty, measuring the mold manually with calipers, and creating a CAD model that will be sent to a computer numerical controlled (CNC) machine to make the doubler fitting.
Typically, it takes five or six prototypes to meet the fit tolerance required, grounding the plane up to six weeks. By contrast, if an engineer could capture the crack with a laser scanner, it would take only a few days to create a CAD model of the bracket to repair the crack, machine the bracket, scan the manufactured part, and compare it to the CAD model to ensure that it meets tolerances, and install the new bracket.
Redesigning a passenger plane lavatory—An aerospace company decides to remodel the lavatory of an airliner to increase passenger satisfaction and update accessories. Because the plane first took flight in 1981, there are no CAD models available. New components must fit into the current space while accommodating fixed infrastructure such as water and power systems.
The designer must start from existing conditions, taking out old parts and installing new ones within the same envelope of space. And, of course, this has to be done in a few weeks because the airliner needs to return to flight as quickly as possible to reduce revenue loss.
Why use this process?
There are two central similarities in these diverse cases: There is a need to capture existing conditions rather than create new models, and time-to-market, or getting back into service, is crucial for asset management.
Both of these requirements are suited for 3-D point-cloud processing. Companies throughout the world are already using 3-D point-cloud processing for MRO applications, including aerospace companies such as Pratt & Whitney, NASA, and Tinker Air Force Base.
To get a better picture of how 3-D point-cloud processing can streamline the MRO process, let’s first review how CAD models are used in downstream applications. Then we will take a look at the traditional process for creating those models for MRO parts and see how it compares to 3-D point-cloud processing.
The importance of CAD models
Accurate CAD models of parts and assemblies are required for a variety of downstream processes, including:
• Generating 2-D drawings for documentation and dimensioning
• Providing the model for CAE studies such as stress, thermal, or fluid dynamics analysis
• Creating a numerical control path that drives a CNC machine for manufacturing a new part
• Producing physical prototypes to test parts and make sure they meet specifications
• Inspecting the as-designed part against the as-manufactured part
The traditional process
The traditional process used to create a digital representation of an actual object begins with a manual measurement tool such as a caliper or a coordinate measuring machine (CMM) to obtain dimensions. In both cases, users need to have the expertise to know which dimensions will be important for reconstructing the digital shape.
With the manual process, the dimensions are in 2-D. With a CMM, dimensions are captured in 3-D but only a small number of predefined points are collected. Collecting a large number of points requires a significant time investment.
The output from the data capture phase is a list of measured key characteristics that can be used to model a 3-D CAD object. This process requires a great deal of time spent on sketching, extruding and rotating profiles, and blending all the surfaces together. Total time for this process is usually measured in weeks.
Streamlining with point-cloud processing
Point-cloud processing starts with the existing part. Or it could be several examples of the same part because point-cloud processing provides the luxury of being able to easily obtain the average or optimal shape within a series of parts. Scanners used in point-cloud processing are capable of capturing millions of points in minutes to accurately represent the physical part.
From those millions of points representing different examples of the part, point-cloud processing software programs can create a new model that is an average of all the samples. Automated tools recognize primitives and profiles from the scan data, repair holes and noise, and wrap the point cloud to form a polygon model. The polygon model is then converted into a surface model that can be directly transferred into popular parametric CAD software.
The points-to-parametric-CAD process typically takes a couple of days. Since it is based on the detailed capture and automated reconstruction of the actual part, it increases accuracy, removes ambiguity, and ensures consistency between the CAD model and the manufactured part.
Add new comment