Photo Courtesy of Visit Detroit
Sponsored by the IEEE Reliability Society
In recent years, Prognostics and Health Management (PHM) has revolutionized the perception of product reliability and has resulted in a broad range of applications. Digital electronics are found in numerous facets of modern day life where consumers of all types of electronic utilization from cell phones to automobiles have come to depend on their reliability to operate effectively in both professional and private life. Furthermore, the commercial and military markets demand even greater reliability from electronics-based systems where a failure could produce catastrophic results. As a result, there is a growing need for a major change in reliability towards system health (degradation) assessment, fault diagnostics, and prognostics (the real-time prediction of reliability and the remaining useful life) of electronics-rich systems. There is also a special need to address soft faults and intermittent failures which are much too common degraders of system reliability. The computer and telecom industry leaders such as Dell, HP, Cisco, Ericsson and many others are investing in prognostics and system health management research as a viable cost-effective solution in response to the demands for more affordable and reliable systems suitable for complex life cycle conditions. Cloud Computing and Computer users expect high reliability and availability from their computers / computer systems, and the goal of PHM is to minimize the possibility of losing data due to faulty operation, degradation of the electronics in a computer / computer system, no fault found (NFF) problems, and unavailability of computer for critical applications. The Manufacturing Industry applies PHM to manufacturing processes and equipment management as a key part of a proactive maintenance thrust that began several years ago. This was a natural outcropping of preventive maintenance to performance-based maintenance, and finally to proactive maintenance. The cost/benefit advantages are quite striking savings in maintenance and spares costs, elimination of line stoppages, and end product quality/reliability improvements. PHM has been applied to Aerospace and Military systems for more than 20 years. The key to applying PHM to Avionics and Military systems is the combination of technology for highly reliable and accurate sensors, algorithms and software design for diagnostics, prognostics, and health management, and definition and development of the integrated environment that uses and matures the PHM capability with the end result being a much more operationally reliable and cost effective system. PHM could have enormous benefits for Power Smart Grid and Green Energy Systems like wind turbine devices in terms of production, reliability and maintenance. The objective of PHM for Wind Energy is the development of an autonomous health management system to control components in wind turbines, especially for a Programmable Logic Controller (PLC). It consists of intelligent in-situ monitoring, fault isolation and location (diagnostics), performance degradation prediction, and remaining useful life assessment (prognostics). PHM can also help ensure LED Lighting Systems availability through extension of maintenance cycles; lowering life cycle costs by reducing inspection costs, repairs, downtime, and inventory; reducing the frequency of intermittent failures; and to provide advanced warning of system failures; enabling pre-planned and proactive maintenance to forestall failures, extend operational life, and to obtain load history data for future design improvement purposes.
Aerospace and Military Systems: Prognostics and Health Management, in one form or another, has been applied to Aerospace and Military systems for more than 20 years. Early on, the concept was simply to detect and report failures basic Built-In-Test. That grew into more sophisticated technologies and methods for evaluating the health of equipment and for using the results to direct and support maintenance. Today, the newest weapons systems are integrating sensors at the lowest level of mechanical, electrical, and electronics for integrating the data to provide an accurate assessment of the current state of the systems in near-real time. They are using the information to determine the future state of systems based on changes in system sensor parameters and usage, and using that diagnostic and prognostic knowledge to determine the best way to execute missions and to enhance and support maintenance and logistics. The ability to modify mission plans based on system health helps ensure mission success in the face of system functional degradation. The ability to plan material logistics, pre-positioning spares, and scheduling maintenance based on current and predicted system health creates a much more efficient and effective environment. PHM is central to this concept. The key to applying PHM to Avionics and Military systems is the combination of technology for highly reliable and accurate sensors, algorithms and software design for diagnostics, prognostics, and health management, and definition and development of the integrated environment that uses and matures the PHM capability with the end result being a much more operationally reliable and cost effective system.
Automotive Electronics: The automotive sector is the largest vehicle electronics industry on the basis of manufacturing volume and total sales. The electronics in modern automobiles have evolved into complex computer systems performing functions like fuel injection and emission control, anti-skid braking, active suspension, and electronic transmission control. Micro-controller based engine and transmission controllers enable automobile manufacturers to meet fuel efficiency and pollution levels mandated by government regulations, while providing the customer with improved driving characteristics. Other customer preference features such as active suspension, climate control, entertainment, and navigation systems also require sophisticated electronics. Transition is underway to next generation vehicles powered by fuel cells or featuring hybrid internal combustion and electric propulsion. It is expected that the electronics content of automobiles will continue to increase and reliability of these increasing complex systems will be even more critical. The current form of diagnostic systems provides little insight about the system level reliability or the remaining useful life of the system. Prognostic health management systems have the potential to enhance the safe-use service life of long-life systems without compromising operational readiness. PHM technologies have wide applicability within the extreme environments for automotive applications.
Structural: Structural degradation, whether engendered by inherent material aging due to the environment and service loads (fatigue, corrosion, radiation etc) or by unpredictable external events (earthquakes, hurricanes, bird impact etc), is an inevitable part of life. Conventionally, the continued reliability of structural components has been ascertained through nondestructive inspections carried out at preset intervals. This approach is currently giving way to the concept of Integrated Structural Health Management to minimize the possibility of catastrophic failure of safety-critical structures. The ISHM approach includes initial assessment of a structure, followed by continual non-intrusive in-service assessment of the structure using on-board sensor systems, and intrusive off-line inspections and repairs as necessary when structures are close to failure. The ISHM approach requires integration of several emerging and some mature sub-fields of science and engineering: sensors, smart structures and materials, damage and failure mechanics, structural and reliability analysis, and nondestructive evaluation. ISHM represents a shift away from routine inspections, which are currently scheduled based on ensemble statistics obtained on similar structures, to inspections that are scheduled based on the actual state of the structural health of individual structures. This is expected to lead to improved safety and result in increased efficiencies in terms of reduced cost of inspections, and minimized disruption.
Nondestructive Evaluation (NDE): Nondestructive Testing (NDT) is defined by ASTM as “the development and application of technical methods to examine materials or components in ways that do not impair future usefulness and serviceability in order to detect, locate, measure and evaluation discontinuities, defects and other imperfections; assess integrity, properties and composition; and measure geometrical characteristic.” The words NDT and NDE are often used somewhat synonymously, with the term NDE increasingly applied to methods in which quantitative interpretation of the test results in terms of characteristics of the material or defects is possible. NDE / NDT traditionally has involved a number of non-invasive measurement modalities such as ultrasound, eddy current, or x-ray, in which a component is interrogated by sensors which are scanned over the component at discrete times, which is often removed from the structure during maintenance. Structural Health Monitoring (SHM) is generally associated with the use of fixed sensors which provide information continuously about the state of a structure. However, these operational differences in time and space conceal many commonalities. Both NDE and SHM fall within the broad ASTM definition as technical methods to examine materials and components to assess integrity without impairing future usefulness and serviceability. Much of the physics controlling the energy / structure / defect interactions are similar for those techniques that involve some sort of wave propagation. There are a number of common issues in interpreting the data (always incomplete) obtained in terms of the future serviceability of the system. There is a growing body of thought that the most effective strategies to assure the future serviceability of components and structures will involve the combination of both types of information. As one example, one could view fixed sensors as “canaries” to indicate when some sort of change is occurring that calls for a more detailed examination by other means. Many other strategies are under consideration. There clearly is great advantage to be gained by using these modalities in complementary ways.
Manufacturing Process / Equipment: Prognostics and Health Management ties to manufacturing processes and equipment as a key part of a proactive maintenance thrust that began several years ago. This was a natural outcropping of preventive maintenance to performance-based maintenance, and finally to proactive maintenance. The cost/benefit advantages are quite striking savings in maintenance and spares costs, elimination of line stoppages, and product quality/reliability improvements. Sensor readings are fed into a central control that monitors vibration levels, flow rates, temperatures, chemical concentrations, and strains to provide real-time trends to the operators. These in turn can help determine when adjustment, removal/replacements, or other actions can be taken to avoid unplanned line failures. For proactive maintenance to work, the potential failure modes must be prioritized, physics of failure understood, precursors to failure monitored, algorithms developed to tie sensed value behaviors to actions, and displays/training to assure operating personnel can use the information to take action, when needed. The key to this is understanding the production/equipment maintenance processes to set-up an effective proactive maintenance methodology and to monitor trends and issues to facilitate continuous improvement of those processes and the maintenance process, itself.
Computer Industry: Computers have become an indispensible part of our daily life, affecting such diverse areas as communication, transportation, health care, trade, finance, security, defense, research, entertainment and social networking. A variety of platforms such as desktops, laptops, tablets, and various hand held computing devices have become common necessities, thus creating a diverse set of operational and environmental stresses, especially as functionality has also continued to increase. Personal laptops, desktops see different operational loads as compared to those used in server farms, industrial settings, and defense applications. Operating a computer in a more stressful environment, such as an industrial location with higher temperatures, higher humidity, vibration, and corrosive gases, will increase the probability of failure. The failures that are generated will depend upon the specific failure mechanism that is initiated due to the specific environmental conditions. The computer industry is investing in prognostics and system health management research as a viable cost-effective solution in response to the demands for more affordable and reliable systems suitable for complex life cycle conditions. Computer users expect high reliability and availability from their computers, and goal of PHM is to minimize the possibility of losing data due to faulty operation, degradation of the electronics in a computer, no fault found (NFF) problems, and unavailability of computer for critical applications. Researchers are developing techniques to monitor system parameters as well as environmental data to estimate the degradation to the components of the computer and predict remaining useful life. Researchers at SUN Microsystems (now part of Oracle) have been using continuous system telemetry along with real-time pattern recognition for improving the reliability and availability of servers. Another example is in the disk drive industry, where companies such as Compaq, Seagate, IBM, Western Digital Corporation, Maxtor and Quantum Corporation are using SMART (Self-Monitoring, Analysis, and Reporting Technology) capability to monitor and compare specific disk drive characteristics to “alert levels,” for actionable responses. SMART can be helpful in predicting and scheduling maintenance to maintain high availability. This conference invites papers in PHM technologies and applications for the computing industry.
Cloud Computing: Cloud Computing (CC) is the next wave in the computing revolution. The general idea of CC is “computing as a commodity” service, similar to water, gas, and electricity. CC is based on 5 key principles: (1) On demand self service, (2) Rapid elasticity, (3) Resource pooling, (4) Measured service, and (5) Broad Network Access, and because of the manner by which it is being sold as a service, requires large amounts of monitoring. The core concerns about CC stem from data security and privacy, however there are equivalent concerns with reliability, availability, and performance. In a public cloud environment, a consumer has no hands on control of their data or applications. Consumer data and applications exist in cyber-space. A consumer’s only hope for having some level of trust in the remote-access public cloud model will occur if monitoring and sensing mechanisms are employed. Cloud Computing and their customers expect high reliability and availability from the complex computer system of systems. The goal of PHM is to minimize the possibility of losing data due to faulty operation, degradation of the complex computer system as well as the supporting power and cooling systems, no fault found problems, and degradation or loss of computer usage for critical applications. Here, PHM principles of fault tolerance, exception handling, sensing, and monitoring are the sole means by which a consumer can receive assurances that their data is secure and private, and the computing power that they require is reliable, always available, and network accessible.
Nuclear Power: Globally there has been great
interest in using nuclear power to provide sustainable,
carbon-free energy, and in the USA there is again interest in
expanding the use of nuclear power. As part of these activities
there are major initiatives focused on “life extension” for
existing light-water reactors beyond the initial 30 or 40 year
lives and in the USA extensions from 60 to 80 (or even 100)
years is being considered. To enable longer term operation a
range of advanced diagnostics methods that are suitable for
on-line, continuous, in-plant monitoring over extended time
periods (months to years) are being considered. A related issue
is then, based on a condition assessment or degradation trend,
to have the ability to estimate the remaining useful life of
components, structures and systems based on the available
condition information. Integration of these diagnostics and
prognostics technologies into the plant instrumentation and
control system are necessary for safe longer term operations.
The use of such diagnostics and prognostics technologies is a
part of the philosophy of proactive management of component
lifecycles that is becoming important for life management and
extension. At the same time, advanced designs for nuclear power
plants (such as small modular reactors and high temperature
reactors) are being considered to boost the generating capacity
and improve efficiencies. It is expected that advanced designs
will function with fewer refueling outages and longer operating
cores, for instance, potential plants with single life-time
cores that function for 20-30 years as compared to the current
1-2 year refueling cycles in light water reactors. These plants
will have new concepts of operations, and automated and enhanced
functions using digital instrumentation and control systems. As
with the existing fleet of nuclear power plants, advanced
diagnostics and prognostics tools are expected to play an
important role in ensuring the safe long term operation of the
next generation of nuclear power plants.
Medical Equipment & Devices: Reliability and predictability are two highly desired features in electronic medical equipment and devices. These systems have evolved from simple data acquisition devices to more complex components of much wider healthcare information systems. PHM opens the way for advances in reliability and predictability in order to extend product life and reduce the total cost of ownership for the customer. These advances help improve design and maintenance planning by providing prognostics and predictive diagnostics knowledge. However, PHM is a relatively new concept in medical equipment and device industries. Further research and development are needed to better understand and develop standards that improve the overall product life cycle
Prognostics and Health Management (PHM): Prognostics and Health Management (PHM) is a technology to enhance the effective reliability and availability of a product in its life cycle conditions by detection of current and approaching failures and by providing for mitigation of the system risks. Prognostics is the real-time enhancement of reliability and availability and the prediction of the remaining useful life of the product by assessing the extent of deviation or degradation of a product’s monitored parameters from its expected normal operating conditions. Prognostics can yield an advance warning of impending failure in a system, thereby enabling more efficient and effective maintenance and corrective actions. Prognostics helps in preventing catastrophic failures and can reduce unscheduled maintenance expenses. The outputs of a prognostic assessment of a product are the failure risk, time to failure, remaining useful life, and a prognostic distance within which maintenance and repair actions can be planned, in order to minimize their impact on product availability. Health Management is the process of using diagnostic and prognostic information to intelligently manage the use and maintenance of a system. The ultimate result is increased effective reliability, availability, and safety with reduced logistics and support costs.
Financial Benefit / Return on Investment: Commitments to implement and support Prognostics and Health Management within systems cannot be made without the development of supporting business cases. Performing cost-benefit analyses and/or return on investment (ROI) estimates for the implementation of PHM in systems depends on careful determination of the costs of developing, implementing and supporting the PHM approaches weighed against potential future cost avoidances created by PHM. Potential cost avoidances include: failure avoidance, minimization of remaining useful life thrown away when performing scheduled maintenance, logistics footprint reduction, repair cost reduction through improved diagnosis and avoidance of collateral damage, reduction in no-fault-founds, reduction in redundancy, warranty verification, waste stream reduction, and improvements in design and qualification of future systems. Additionally, with the advent of outcome-based contracting, specifically Performance Based Logistics (PBL) and Product Service Systems (PSS), where the system manufacturer/supporter is compensated in part based on the availability achieved by the customer, trading off life cycle costs against availability for systems that incorporate PHM becomes tantamount.
Logic / Algorithms / Reasoning Approaches: System monitoring requires sensing and logging real-time data, techniques that interpret the meaning of such data, and reasoning and learning systems to filter large volumes of logged data into human interpretable results. Computer Science has advanced over the past 4 decades in algorithms for reasoning, and learning about large quantities of data, even data that is not quantifiable. As one example, recent work in machine learning algorithms have resulted in the development of string kernels, which provide various similarity measures for strings of symbols. The ability to determine when two strings are similar or dissimilar is important in adapting traditional system health-monitoring techniques for anomaly detection using complex learning algorithms. These learning systems can also be used to detect failures more rapidly as they learn how to build new warning indicators.
Condition-Based Maintenance (CBM): The objective of CBM is to maintain the correct equipment at the correct time. CBM is based on using real-time Prognostics and Health Management data to prioritize and optimize maintenance resources. By observing the state of the system (condition monitoring), the system will determine its health, and act only when maintenance is actually necessary thus minimizing the remaining useful life of the equipment that is thrown away. Using CBM, maintenance personnel are able to decide when the right time to perform maintenance is. Ideally CBM will allow the maintenance personnel to do only the right things, minimizing spare parts cost, system downtime and time spent on maintenance.
Predictive Maintenance (PdM): In the evolution of maintenance strategies, PdM makes use of PHM type monitoring methodologies, but with a more aggressive goal – to use the known status of the monitored equipment, combined with physics-of-failure models and statistics (vs. data trending) to provide the maintenance function with a measure of risk vs. time-to-failure. This time-to-failure may be measured as “degradation” or “consumed/remaining safe or useful life.” This methodology may be called “CBM+” or “Proactive Management of Materials Degradation (PMMD.” These approaches all are intended to provide maintainers with a “look-ahead,” so that the degradation or remaining life can be weighed against other maintenance tasks, planned or unplanned.
Use of Sensors and Canaries for PHM: Sensor systems are needed for PHM to monitor environmental, operational, and performance-related characteristics. The gathered data can be analyzed to assess product health and predict remaining life. A generic sensor system will typically have sensing elements based on a suitable transducer technology, onboard analog-to-digital converters, onboard memory, embedded computational capabilities, data transmission, and a power source or supply. Sensor systems with multiple sensing abilities, miniature size and light weight, low power consumption, long range and high rate data transmission, large onboard memory, fast onboard data processing, low cost, and high reliability are specifically advantageous to PHM applications. Recent PHM research is also focusing on the concept of “canaries” for early warning of degradation and impending failures so that risk mitigation options can be rapidly implemented. Like, canaries in a coal mine, these are sacrificial non-functional elements in the design, and are tailored to degrade in an accelerated manner in response to different stress sources. The degradation mode in the canary is typically an accelerated version of the degradation expected in the functional product, and the design is usually based on fundamental physics-of-failure concepts. Canaries can significantly reduce the uncertainties in the remaining useful life estimates because each canary is tailored to a known failure mechanism and the materials, manufacturing processes, and life-cycle stresses can be tailored to “match” the behavior of the functional elements that are being monitored. Canaries can be calibrated to provide sufficient advance warning of failure (termed ‘prognostic distance’ or ‘remaining useful life’) to enable timely maintenance and replacement actions. System level research is also needed for integrating networks of sensors and canaries; collecting, post-processing and storing their outputs, and delivering power to these networks. This conference invites papers on all aspects of sensors and canaries for PHM applications.
Numerical Modeling and Methods in PHM: The robust diagnostic and prognostic techniques in PHM that increase the effective reliability, availability, and safety of a system need to provide quantifiable information, e.g. damage location; size estimation, propagation and prediction, that account for the uncertainties induced by the environments or the systems continuously. The development of well-established and validated numerical modeling and methods can facilitate the understanding of problems with complex geometries, high dimensionality, nonlinear, anisotropic and inhomogeneous material properties for various sensing techniques and systems in PHM. The models are useful in visualization the field-flaw interaction which is important to understand the underlying physical principles. By providing quantitative values of field distribution, numerical models can be used to adjust and optimize the operational system parameters, improve the sensor design and development. When the measured signal is expensive or not adequate to acquire, the output of the numerical models can be substituted as test bed to generate training data and develop the model-based probability of detection (POD) estimation, uncertainty analysis, probabilistic risk assessment (PRA), and model-based inversion techniques. Additionally, the models can help understanding how a complete system can interactively merge the different PHM technologies and advantage the capabilities of each single method, which improve the overall system performance by covering all aspects of health monitoring and management. This conference provides an excellent opportunity for an open and intellectual discussion on numerical modeling in PHM and invites papers on this application topic.
PHM Standards: There are many aspects of the Prognostics and Health Management that need to be addressed before the successful implementation in any product. These include system classifications, methodologies for PHM, sensors and sensor selection for monitoring, data acquisition, data processing, algorithms for fault identification, fault isolation, failure prediction and remaining useful life estimation, software, testing (algorithms, software, sensors), reliability (sensors and systems), decision making for condition based maintenance, logistics, costs, and training of personnel. Successfully implementing PHM in a fielded system will require consideration of each of these components. The purpose of the standards is to provide guidelines for conducting research, development, and implementation of PHM technology. Standards could cover such aspects of the field as terminology, definitions, methodologies, metrics, specifications, testing, software, hardware, monitoring, analysis, modeling, products, systems, reliability, safety, ratings, application, best practices, compliance, management, and training. The standards developed and approved through consensus will be the guidelines by which organizations should proceed with PHM implementation. The ultimate goal of the advancement of PHM worldwide can only be achieved if the standards are developed in consensus, are consistent with international anti-trust laws, and prohibit monopolization. This conference provides an excellent opportunity for an open and intellectual discussion on PHM standards and also invites papers on developing standards and guidelines for different aspects of PHM.