September 2008
Special section: Data Acquisition & I/O

High winds tester

High-speed wind tunnel distributed data system precludes programming

FAST FORWARD

  • Aerodynamic test facility upgrade helps restart service.
  • High-speed data acquisition system to the rescue.
  • Results: Per-test configuration and execution without reprogramming.
 
By Kent Lowrance and Cynthia Whitley

When it is time for an old aerodynamic test facility to go back into service, it is important to restore all data acquisition systems capabilities within the facility.

The National full-scale aerodynamic complex (NFAC) is a large-scale test facility at the NASA Ames Research Center at Moffett Field, Calif. The facility has two test tunnel sections (40ft by 80ft and 80ft by 120ft) capable of different low-speed atmospheric test conditions. In 2002, NASA announced it was closing the facility, and the last customer test was completed in February 2003. In 2004, the Department of Defense (DOD) announced NFAC operations and maintenance would return to service, funded as part of the DOD budget. One part of the NFAC reactivation project was to restore data acquisition systems capabilities within the facility.

The task was to design, develop, install, check out, and commission NFACs new high speed data acquisition systems for the two tunnels as part of the facility's reactivation effort. We also provided test support, training, and maintenance services for the data systems.

The solution was a flexible distributed data system for large wind tunnel testing that allows per-test configuration and execution without the need for reprogramming. The system is capable of high-speed data acquisition, processing, and storage for a large channel count system by distributing the work among several different computers.

The new NFAC data acquisition system (DAS) provides capabilities for acquiring, processing, displaying, recording, and reducing test article measurements. The processing involves converting raw data to engineering units and deriving second-level parameters.

In this system, we combine currently available data measurement hardware with today's faster computers to provide a functional system that is easily upgradeable. Operations are easier with a flexible interface for users to share information simultaneously while testing is in progress.

System architecture

The NFAC DAS is composed of four types of acquisition systems. The dynamic DAS (DDAS) is for high speed strain or acoustic signals. The basic DAS (BDAS) is for strain or acoustic signals. It includes multiple computer systems and provides for acquisition of model balance signals, tunnel balance signals, and strain gages at medium speeds. The safety of flight DAS (SoFDAS) (or black box recorder) acquires a subset of the BDAS and DDAS channels and provides limit checking functions to ensure the health of both the unit under test and test apparatus systems.  The basic DAS2 (BDAS2) includes a single computer system interfaced to all test support hardware and provides acquisition for pressures, temperatures, and other low speed interfaces.

Each DAS puts its processed data into a reflective memory (RFM) network for distribution to local displays and to switched central distribution and recording hardware. Additional DAS support computers include a data storage server computer with redundant array of inexpensive disks (RAID), a real-time calculations computer, and a data distribution computer. The data distribution subsystem makes the data available to display computers via a standard Ethernet network.

A real-time calculation subsystem provides data processing and analysis for displays and other near real-time needs. A data storage system archives data and provides it to other external computers for further processing and analysis if required. The data storage systems also function as an asset management system server to provide test unit and system configuration to the various subsystems and to the external processing computers.

Each wind tunnel has four operation stations: the master operator console (MOC), two test engineer stations, and a safety engineer station. Four additional customer stations provide data display and analysis.

Network

The DAS is a distributed computing system connected by an RFM real-time network and switched Ethernet, local area network (LAN) within the facility. The RFM network is the key component of the distributed system. The network provides a common data area for all attached computer processors. This allows any processor to place data in RFM and make it available to all other processors with no processor overhead.

The data transmission rate is about 170 megabytes per second. RFM not only provides high bandwidth and low-latency data transfer, it simplifies the software required for the DAS. It also permits easy consolidation and distribution of software modules during system implementation.

If a data system's processor overloads and cannot handle the required number of channels, an added second processor will divide channel capacity between the two new processors. You will need minimal software development to add the second processor, since both systems are doing exactly the same functions. The DAS also uses a more traditional LAN to transmit system configuration and status data. The LAN allows transmission of near real-time data to multiple customer computers using TCP software. The LAN uses a multilayer, backbone switch capable of supporting 1 gigabit/sec Ethernet, 100 megabit/sec Ethernet, or 10 megabit/sec Ethernet.

Data1

Data resampling

A critical requirement of the DASs is the capability to acquire timing pulses from the rotorcraft encoder along with time-based measurement data for use in resampling algorithms that correct data to accurately reflect what occurred at a rotor position at high speeds with high accuracy. This accuracy is typically less than 0.1 degrees of rotor angle while operating at speeds up to 2,000 RPM.

The position corrects in real time for the 256 DDAS channels at 68 KHz/channel and the 240 BDAS channels at 9 KHz/channel when acquiring rotorcraft data. The system as a whole can acquire, reprocess, and store to disk almost 20 million samples of data every second at a continuous rate.

Interface software

The master operations console (MOC) is the main control station for the distributed data architecture. From this location, an operator can control all aspects of testing. An operator may select and load test configurations, change data system modes, calibrate subsystems, perform diagnostics, and control data storage. From the MOC and all engineer stations, the software allows users to perform configuration management, display current data, view and report on stored data, and monitor for alarm conditions. A user configurable work space with dockable windows, an explorer tree-like configuration user interface, a drag-and-drop user display builder, and management of configuration information via a relational database engine make this software a powerful tool for configuring and executing tests.

System explorer

You can navigate the DAS hardware subsystems, sensors, tags, displays, calculations, alarms, and trips through the system explorer window, which implements a tree structure similar to Microsoft's Windows Explorer. From the tree structure, you can add, remove, and configure data as needed to meet testing requirements.  Make changes to hardware subsystems, sensors, and tags in editor windows.

Sensor database

A single sensor database for all configurations ensures use of the latest calibration. For those subsystems that provide data in the form of counts or millivolts, the sensor database is used to configure the engineering unit conversions, units, sensor identification information, sensor calibration data, etc., for each sensor. Conversion support for up to the 14th order polynomial, resistance temperature detector (RTD), and B, E, J, K, N, R, S, and T type thermocouples is provided.

Calculations

More than 15,000 custom and standard calculations for the DAS were coded.  Custom calculations include statistics, safety of flight, tunnel, RPM, balance, and scale calculations. We calculate statistics (average, minimum, maximum, standard deviation, and half peak-to-peak) for every tag and perform them across each rotation or data point. Safety-of-flightlight calculations include mean stress levels, vibratory values, vector monitoring, rotor balance flexure stress, and rotor balance force and moments. Tunnel calculations include tunnel conditions such as barometric and atmospheric pressure, tunnel temperatures and pressures, uncorrected dynamic pressure, static plate values, tunnel centerline pressures, tunnel Mach number, Reynolds number, Q and wind direction. RPM calculations include rotor conditions such as rotor tip speed, rotor Mach number and advancing Mach number, tip speed ratio, and rotor force coefficients.

Alarms, trips

Four levels of windowing alarms and trips with optional operator acknowledgment, two levels of rate of change alarms and trips with optional operator acknowledgment, and alarms and trips based on the variance of a group of analog type parameters are available for each test. Operator messages generate when an alarm or trip condition occurs and when an alarm or trip condition clears. The test sequencer can also use alarms or trips to cause automatic jumps to shutdown or execute other sequences in the event of a fault condition.

Custom displays

Custom displays provide windows for viewing real-time test data. You can build these free-format displays on the fly with a drag-and-drop format that has many types of controls including tabular, gauges, charts and graphs, and text boxes. You can create displays prior to or during testing by selecting the option from the system explorer. Arrange displays into groups, and each group can contain multiple displays. Security is such that anyone can view displays from a group, but only the user that created the group can modify and save it.

Data storage, analysis

During tests, the software generates data in two basic formats. The first format is current value data, which is averaged raw data acquired at specific intervals, converted to engineering unit data values, and made available for real time use in displays, alarms, calculations, and the like. The current value data is available at all times. The second format is buffered raw data, which contains each individual sample of data every subsystem has collected. When data stores to file, the data storage routine collects the buffered raw data and separates it into multiple files, depending on source, and stores the data to the RAID system. After data storage completes storing the test data, a post point routine analyzes the raw data and writes engineering unit and calculated files for all stored data. You can also recalculate data if conversion coefficients or calculations change post-test.

ABOUT THE AUTHORS

Kent Lowrance is a project manager with Jacobs Technology Inc., a technical services consultant specializing in commercial, industrial, and government markets in Tullahoma, Tenn. Cynthia Whitley is a section manager at Jacobs Technology.

RESOURCES