Condition Based Monitoring Computational Processes
Navy SBIR 2015.1 - Topic N151-027
NAVSEA - Mr. Dean Putnam - [email protected]
Opens: January 15, 2015 - Closes: February 25, 2015 6:00am ET
N151-027 TITLE: Condition Based Monitoring Computational Processes
TECHNOLOGY AREAS: Sensors, Electronics, Battlespace
ACQUISITION PROGRAM: PMS 450, VIRGINIA Class Program Office.
OBJECTIVE: The objective is to develop computational processes that employ both deterministic and stochastic processes for correlating known transient functions with operational values to access an accurate state of conditions for monitoring health and predicting life-cycle.
DESCRIPTION: The US Navy has a current need for in situ non-destructive sensing and condition monitoring at the node for component system characteristics in the time, frequency or modal domains. A cost effective method for monitoring health is having an early warning system for machine failure. Algorithm based detection systems suited for machinery failure modes beginning at high frequencies can be triggered as a result of operational transients. To enhance prediction of failure modes, algorithms for tracking damage accumulation with continuous monitoring from acceleration disturbances provide essential information for helping to improve component design and scheduling of maintenance functions. Limitations on hardware storage and power efficiency sometimes prevent health monitoring sampling rates to be consistent and enabling high bandwidth transient capture is problematic given the available technology. Many approaches trending towards the use of statistical methods is highly recommended to solve missing data issues where gap filling is needed when predictive analysis is necessary for measuring the conditions at the node. Detection is necessary in situations where there are machine failures in real situation such as unexpected resonance conditions or deviations caused by operator error, improper maintenance or unexpected events. To aid in the detection of health monitoring, wireless sensors are required at the node or at the component level to capture transient or steady-state response. This is particularly important as a transfer function is necessary to establish a frame of reference from which to capture pattern recognition characteristics of the component. To this end data from sensors can be measured by a handful of instruments including accelerometer, strain gage, thermo, temperature, etc., and converted for analysis purposes from time to frequency domain using Fast-Fourier-Transform (FFT) techniques or to modal domain [1, 2, 3, 7, 8]. Diagnostic functions reduce troubleshooting and maintenance times, prevent fault misdiagnosis, and avoid incorrect part repair or unnecessary replacement. Wireless devices are packaged to allow installation or mounting to components with the intent of eliminating cabling runs and the maintenance and installation costs associated with cables [4, 6].
PHASE I: The company will develop logical processes necessary to write algorithms to digital signal process sensor data including transient and steady-state input for the purpose of identifying and predicting faults and failures. Use of statistical methods is paramount for successful health monitoring and trending analysis for maintenance actions and predictive failures. Input sensor and transient data shall be from similar shipboard component, reasonable likeness or modeled to simulate system functions including known and predictive failure modes. Components shall be shipboard systems requiring health monitoring such as, but not limited to, actuation of valves that operate by electric or hydraulic based linear and rotatory dynamics, pumps, motors, compressors, etc. Consideration for algorithm processing shall execute code on processors that will be designed to capture data at the node where the component resides for the purposes of allowing wireless hardware to transmit secure output. Algorithms will also need to track the damage accumulation with continuous acceleration data (and gap filling for missing data). Example cases of machine failure modes in real situations (especially off-design conditions that commonly lead to failures, such as unexpected resonance conditions or deviations caused by human error, improper maintenance or unexpected process conditions). The company will develop methods for collecting and processing data and define requirements necessary to apply to specific applications where diagnostics and prognostics can be performed to apply condition based monitoring techniques. The Navy needs will be meet as company demonstrates component systems are tested to validate successful detection from monitoring various intentional input failure modes and providing symptom/effect as a result of the fault and failure for trending and life-cycle predictions. The company will prepare a development plan for Phase II, which will address technical risk reduction associated with developing algorithms for condition based monitoring based on sensor, transient and steady-state data, as well as performance goals of detecting and validating failure modes and key algorithm and statistical development milestones.
PHASE II: Based on the results of Phase I and the Phase II development plan, the small business will be implementing the algorithms developed in Phase I within a comprehensive wireless infrastructure. The wireless infrastructure includes array of sensors common to a component node, wireless interface and self-contained power transmit and receive. The infrastructure is considered a managed network for a condition monitoring system. The company will follow a system development methodology already in progress that includes a requirements definition wherein performance, interface, security and reliability requirements are defined within a system specification. Component and software requirements specifications are developed based on the system requirements. The company will incorporate the algorithms for data processing into the infrastructure to add the diagnostics and prognostics necessary to allow for condition based monitoring at the node compatible within the wireless network. The entire infrastructure is considered a prototype and as such the company shall demonstrate that the algorithm will execute successfully and process various sensor input data defined within the system requirements and detect transient and steady-state waveforms for digital signal processing. It will be necessary to process data in the time, frequency and modal domain and allow for conversion from time domain to frequency domain using Fourier transform techniques. It will be necessary to demonstrate ability for correlation techniques of processing input data of a transfer function from operation to a known transfer function or criteria as reference. The differential will be used for deterministic as well as stochastic approach to measuring component health. Similarity sensor data will be auto-correlated to determine the solution for health status leading to actionable information. The successful system will be performance tested in situ either in a shipboard environment or laboratory depending on resources available. The testing will evaluate the analytical or model of simulated failure modes and data from actual component over a range of operational parameters to represent deployment cycles. Evaluation results will be used to refine the algorithms as necessary to meet Navy requirements. The company will prepare a Phase III development plan to transition the algorithms for Navy use.
PHASE III: If the Phase II is successful, the company will be expected to support the Navy in transitioning the algorithms for Navy use should a Phase III award occur. Based on the Phase II results, the company will integrate the algorithms into a next generation wireless infrastructure for installation on a ship to conduct testing as an effective tool for diagnostics and prognostics as part of a new era in shipboard wireless processing of sensor data for machinery condition based monitoring. The processes used for the algorithm will be defined and incorporated into the wireless infrastructure requirements specification and used for procurement purposes. The company will continue to support the Navy for revision to the algorithm to improve processes necessary to monitor various systems and variants as improvements are made to enhance at node machinery condition based monitoring techniques.
PRIVATE SECTOR COMMERCIAL POTENTIAL/DUAL-USE APPLICATIONS: The technology developed under this topic could be used in a wide range of industry applications, providing specific system performance information, and a method for collecting and analyzing such information, to enable a breakthrough in wireless at node condition based monitoring for shipboard environment.
2. Goff, C. I., McNamara, C. L., Bradley, J. M., Trost, C. S., Dalton, W. J. and Jabaley, JR., M. E. (2011), "Maximizing Platform Value: Increasing VIRGINIA Class Deployments". Naval Engineers Journal, 123: 119�139. https://www.navalengineers.org/Hamilton_Award_Papers/2011/Goff.pdf
3. Byington, Carl S., Michael J. Roemer, Gregory J. Kacprzynski, and Thomas Galie. "Prognostic Enhancements to Diagnostic Systems for Improved Condition-Based Maintenance", 2002, DTIC Document Accession Number: ADA408880. http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA408880
4. Yick, Jennifer, Biswanath Mukherjee, and Dipak Ghosal. "Wireless sensor network survey." Computer networks 52.12 (2008): 2292-2330. http://ahvaz.ist.unomaha.edu/azad/temp/ali/08-yick-wireless-sensor-network-localization-coverage-survey-good.pdf
5. Byington, Carl S., Michael J. Roemer, Gregory J. Kacprzynski, and Thomas Galie. "Prognostic Enhancements to Diagnostic Systems for Improved Condition-Based Maintenance", 2002, DTIC Document Accession Number: ADA408880. http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA408880
6. Loverich, Jacob J., Jeremy E. Frank, and Richard T. Geiger. "Self-powered Wireless Sensors for Condition Based Maintenance on Ships", 2009 Intelligent Ships VIII Proceedings, May 20-21, 2009. Drexel University in Philadelphia, PA.
7. Sinha, Jyoti; Elbhbah, Keri. "A future possibility of vibration based condition monitoring of rotating machines", 2013, Mechanical Systems and Signal Processing, Volume 34, 231-240.
8. Chattopadhyay, Aditi, Mark Seaver, Antonio Papandreou-Suppapola, Seung B. Kim, Narayan Kovvali, Charles R. Farrar, Matt H. Triplett, and Mark M. Derriso. "A Structural Health Monitoring Workshop Roadmap for Transitioning Critical Technology from Research to Practice", 2012, DTIC Document Accession Number: ADA554786. http://www.dtic.mil/dtic/tr/fulltext/u2/a554786.pdf
KEYWORDS: Electric, hydraulic, actuator, wireless condition based monitoring; sensor, nodes, algorithm, diagnostics, prognostics, transient functions, damage accumulation, health monitoring, failure modes, detection and life-cycle
Offical DoD SBIR FY-2015.1 Solicitation Site: