The volume of road load data is equivalent to over 100 CDs per vehicle. Land Rover has deployed ICE-flow to handle the data processing and management challenges...
During the development of a vehicle, enormous amounts of field test data need to be collected, processed, and distributed to the various departments that are waiting for them. Testing is needed to provide ‘real-life’ loads for the virtual product development process; to provide feedback on components as they are developed, and is required for final ‘sign-off’ of the early prototypes and final product. And the volume of data is vast, especially during Road Load Data collection.
A typical road load test program combines runs on proving grounds, tests on public roads, and, in the case of Land Rover, extended off-road trials. They also take place around the world, for example in desert and Arctic conditions. The proving ground involves driving around long circuits at various speeds and over many specific events, such as gravel beds, cobblestones and potholes. Public tests usually follow specified courses of highway, town and off-highway usage, all at pre-defined speeds and vehicle loads. Data can be acquired for just a few seconds or for a couple of hours, while the test program itself can extend for weeks or months. Given the cost of testing, and the critical role that testing plays in the development process, it’s important to get as much information as possible. Usually between 150 and 200 channels of data are collected: analog signals include strain, acceleration, displacement and temperatures; digital signals from the vehicle bus are also recorded to provide information on how the vehicle is actually being driven, for example engine speed, gear selection and so on. Altogether, a typical six-month test program can generate between 15GB and 120GB of data - and that can be when the challenges start and the opportunities arise!
The data need to be validated, ideally before leaving the field, to ensure that they do not suffer from overloads, underloads, drift, or spurious readings. Does everyone use the same validation process? Or indeed the same testing configuration in the first place? The fact that the time histories are so huge, easily large enough to swamp the Internet bandwidth, creates its own problems. How can hundreds of channels of data, each megabytes long, be quickly inspected and cleaned while still on the trackside? Things that were so obvious in the field, mundane facts such as the actual vehicle configuration, or what it was driving over at the time, need to be shared with the people back at the office. However, such information (known as metadata) is often inconsistently appended to the measurement data – or simply lost. How can a technician in the field enter this type of information? All too often, CDs get archived in desks and cupboards, and engineers lose time searching for the right disk, and then wading through the mass of data to find the right track. The data need to be processed and reduced, and since nobody wants to overwrite the raw data, that means the data volume can multiply dramatically - which only compounds the problem. Furthermore, what guarantees that the data are processed consistently from person to person? Finally, data need to be traceable at any point in the process, so that everyone knows what is happening and who did what to the data – especially when tests are carried out from anywhere in the world and are to be used by others in widely separated locations.
The Road Load Data Process is a prime candidate for change. Tests need to be better prepared; more information needs to be gathered in the limited time available; data need to be quickly validated, reduced, analysed and delivered. Above all, everything has to be faster than it is today. Its not about more algorithms, ‘breakthrough approaches’, or starting again from scratch – it’s simply about supporting the process in a better, more efficient and consistent way. And getting the right information to the right people every time. It’s a challenge that Land Rover addressed with the ICE-flow system from nCode International.
Combining Integration, Communication, Engineering, and workflow management tools, ICE-flow integrates hardware and software in the form of testing and analysis solutions to provide a comprehensive end-to-end process solution for Road Load Data Analysis. The goal of ICE-flow is to help companies move from ad hoc testing procedures - to an integrated process where managers can design multiple event testing programs at base, download them to the test department, and monitor the ongoing test status at anytime. It contains all the algorithms to automatically detect anomalies, clean the data, and to discover which channels are important for durability engineering purposes at the end of each test run. It helps everyone back at base know which data sets have not yet been validated, which are suspect, and which can be made available to the departments that are waiting for them. It provides the advanced data management tools and process audit trails that help everyone work more professionally. And it has been configured to match established Land Rover processes and IT infrastructures.
ICE-flow is a modular system combining an open architecture, a range of web-based and desktop tools, and powerful process-oriented approaches. At its heart is the central server, which connects to the user ‘clients’ over the Internet or company Intranet. The server comprises modules for process compiling, test preparation, data validation, data processing, and reporting. ICE-flow also accesses local databases for process and resource management, for the huge volumes of raw and processed data, and for all associated test information. Rather than spending time transferring massive amounts of raw data over the Intranet, the central server does all the work and just sends display information to the client. The system is therefore very fast and interactive: users can search for appropriate datasets, pan across long time histories and down through hundreds of channels, zoom in, use cursors…just as if they were connected to the data directly.
A common ICE-flow system is used in both Jaguar and Land Rover; however the data are stored in independent physical storage systems in the respective companies to minimize data traffic. As an integrated road load management system, ICE-flow controls the complete test program: from project initiation, finding out which equipment is available and calibrated, to noting where it physically can be found, who used what – and when, to managing a hundred of other time consuming yet important details. Information from the vehicle, measurement instrumentation, cabling, and transducers can be captured automatically using barcodes or similar approaches. Manual data entry is only performed once. The ICE-flow environment even provides change control notification that can help enforce corporate procedures. The result: everything runs faster and more smoothly, there is less room for mistakes, and everything is traceable – just in case. A new test project starts by defining the vehicle and its associated build – and from this point each set-up, every change, and all incoming data is automatically associated to it. Managers can enter general details such as the date the test is required, ownership information, who wants the data…as well as specific details such as tyre specifications, vehicle loads and so on. In fact as much or as little information as is needed at the time - extra fields can easily be added later ‘on the fly’ by any operator. Each measurement is associated with an individual transducer / cable and is mapped to a specific channel on the recording system. These are also associated with the specific location on the vehicle component and the measurement direction. Real channels, virtual channels and sub-channels are easily set-up, as is grouping by name, transducer type, critical location, rosette, list of channels…
Turning the deluge of information into a controlled stream of useful knowledge is a critical part of the process. ICE-flow imports data from a variety of sources, including MEGADAC and nCode’s own SoMat eDAQ. After testing, the driver is able to perform many basic data validation tests by the trackside to ensure that tests do not need to be repeated. It is also possible for engineers back at base to run more sophisticated checks, such as anomaly checks using cross channel statistics. Raw data are first stored into a protected area, then the formal data validation and processing phases can begin. Data reduction used to be a time consuming and potentially error-prone process – where anything but the simplest processes had to be programmed by an expert. ICE-flow provides an interactive way of defining and building the most sophisticated analysis process – and everything has been designed from the ground up for multichannel, multi-event, multi-test processing. Processes can be developed back at base and stored on the server. Anyone from anywhere can then simply upload and use a specific analysis routine – saving a lot of time and eliminating the potential for errors. At this point, the data are known to be clean, processed, and ready for distribution to the other departments.
The raw data are always preserved and cannot be accidentally ‘lost’ or overwritten. Every change is logged by the system, which records the person that made the entry, the time it was made, and the nature of the change. If needed, analyses up to a given point in the audit trail can be re-run to correct for an error, or to take into account new circumstances. Anyone involved in the data chain can easily append additional ‘metadata’ to the raw or processed data to identify aspects of the test that will be important downstream. Only authorized users can log in to the system, and the system administrator can assign each a different permission level to view data, to modify tests, to add metadata, and so on. It all adds up to improved system security, process integrity, and data protection. In short, the best data by far.
For more information on ICE-flow, click here
For more information on nCode, contact your sales office