Standard Article

You have free access to this content

Prognostics and Health Management of Electronics

Other Applications

  1. Michael Pecht

Published Online: 15 SEP 2009

DOI: 10.1002/9780470061626.shm118

Encyclopedia of Structural Health Monitoring

Encyclopedia of Structural Health Monitoring

How to Cite

Pecht, M. 2009. Prognostics and Health Management of Electronics. Encyclopedia of Structural Health Monitoring. .

Author Information

  1. University of Maryland, Center for Advanced Life Cycle Engineering (CALCE), College Park, MD, USA

Publication History

  1. Published Online: 15 SEP 2009

1 Introduction to the Prediction of Reliability

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading

The traditional reliability prediction methods for electronic products include Mil-HDBK-217 [1], 217-PLUS, Telcordia [2, 3], PRISM [2], and FIDES [4]. All these methods assume the components of the system have constant failure rates with “pi-factor” modifiers to account for various quality, operating, and environmental conditions. There are numerous problems with these types of modeling approaches, and these have been mentioned in hundreds of papers [5, 6]. The general consensus has been that these methods should never be used, because they are inaccurate for predicting actual field failure events, and they are highly misleading, and can result in poor designs [3].

In the Mil-HDBK-217A documentation published in 1965, there was only a single point failure rate for all monolithic integrated circuits, regardless of the stresses, the materials, or the architecture. Mil-HDBK-217B was published in 1973, with the RCA/Boeing models simplified by the Air Force to follow a exponential distribution. In 1979, Mil-HDBK-217 C was published to “band-aid” the problems. To keep pace with the accelerating and ever changing technology base, Mil-HDBK-217 C was updated to Mil-HDBK-217D in 1982 and to Mil-HDBK-217E in 1986. In 1991, Mil-HDBK-217F became a prescribed US military reliability prediction document [7]. In the meantime, IEEE 1413 standard provided guidance on the appropriate elements of a reliability prediction [8]. The IEEE 1413.1 guidebook provides a summary of the evaluation of the common methods of reliability prediction described in this document. That information should be utilized for determining which reliability prediction method is appropriate in a particular application [9].

The physics-of-failure (PoF) approach and design-for-reliability (DfR) methods have been developed by the Center for Advanced Life Cycle Engineering (CALCE) at the University of Maryland with the support of industry, government, and other universities. PoF is an approach that utilizes knowledge of a product's life-cycle loading and failure mechanisms to perform reliability modeling, design, and assessment. The approach is based on the identification of potential failure modes, failure mechanisms, and failure sites for the product at a particular life-cycle loading condition. The stress at each failure site is obtained as a function of both the loading conditions and the product geometry and material properties. The use of PoF modeling approaches for electronic components and devices, like those used for mechanical systems, are a powerful tool in support of electronic prognostic capabilities. This is because the root cause of almost all electronic devices or component failures is often mechanical—something physically breaks at a subcomponent, solder joint, connection, layer, delamination, etc., level. Solder fatigue models are already under development and show promise [10, 11].

Prognostics and health management (PHM) is a method that permits the assessment of the reliability of a system under its actual application conditions. When combined with PoF models, it is thus possible to make continuously updated predictions based on the actual environmental and operational condition monitoring of each individual product.

Assessing the extent of deviation or degradation from an expected normal operating condition (i.e., health) for electronics provides data that can be used to meet several critical goals, which include (i) providing advance warning of failures; (ii) minimizing unscheduled maintenance, extending maintenance cycles, and maintaining effectiveness through timely repair actions; (iii) reducing the life-cycle cost of equipment by decreasing inspection costs, downtime, and inventory; and (iv) improving qualification and assisting in the design and logistical support of fielded and future systems [12, 13]. The importance of PHM has been explicitly stated in the US Department of Defense 5000.2 policy document on defense acquisition, which states that “program managers shall optimize operational readiness through affordable, integrated, embedded diagnostics and prognostics, embedded training and testing, serialized item management, automatic identification technology, and iterative technology refreshment” [14]. Thus, a prognostics capability has become a requirement for any system sold to the Department of Defense.

2 Modeling of Stress and Damage Utilizing Life-Cycle Loads

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading

Life-cycle loads of a product can arise from manufacturing, shipment, storage, handling, operating, and nonoperating conditions. The life-cycle loads (thermal, mechanical, chemical, electrical, and so on), either individually or in various combinations, may lead to performance or physical degradation of the product and reduce its service life. In the stress–damage prognostics approach, the extent and rate of product degradation depends upon the magnitude and duration of exposure to loads (usage rate, frequency, and severity). The life-cycle loads are monitored in situ, and used in conjunction with PoF-based damage models to assess the degradation due to cumulative load exposures.

In studies made by Ramakrishnan and Pecht [15] and Mishra et al. [16], the test vehicle consisted of an electronic component-board assembly placed under the hood of an automobile and subjected to normal driving conditions in the Washington, DC, area. The test board incorporated eight surface-mount leadless inductors soldered onto an FR-4 substrate using eutectic tin–lead solder. Solder joint fatigue was identified as the dominant failure mechanism. Temperature and vibrations were measured in situ on the board in the application environment. Using the monitored environmental data, stress and damage models were developed and used to estimate consumed life.

Shetty et al. [17] applied the PHM methodology for conducting a prognostic remaining-life assessment of the end effector electronics unit (EEEU) inside the robotic arm of the space shuttle remote manipulator system (SRMS). A life-cycle loading profile for thermal and vibration loads was developed for the EEEU boards. Damage assessment was conducted using physics-based mechanical and thermomechanical damage models. A prognostic estimate using a combination of damage models, inspection, and accelerated testing showed that there was little degradation in the electronics and they could be expected to last another 20 years.

Mathew et al. [18, 19] applied the PHM methodology in conducting a prognostic remaining-life assessment of circuit cards inside a space shuttle solid rocket booster (SRB). Vibration time history recorded on the SRB from the prelaunch stage to splashdown was used in conjunction with physics-based models to assess the damage due to vibration and shock loads. Using the entire life-cycle loading profile of the SRBs, the remaining life of the components and structures on the circuit cards was predicted. It was determined that an electrical failure was not expected within another 40 missions.

Gu et al. [20] developed a methodology for monitoring, recording, and analyzing the life-cycle vibration loads for remaining-life prognostics of electronics. The responses of printed circuit boards (PCBs) to vibration loading in terms of bending curvature were monitored using strain gauges. The interconnect strain values were then calculated from the measured PCB response and used in a vibration failure fatigue model for damage assessment. Damage estimates were accumulated using Miner's rule after every mission and then used to predict the life consumed and remaining life. The methodology was demonstrated for remaining-life prognostics of a PCB. The result was also verified by the real time to failure of the components by checking the components' resistance data.

Simons and Shockey [21] performed a PoF-based prognostics methodology for failure of a gull-wing lead power supply chip on a DC/DC voltage converter PCB assembly. First, three-dimensional finite element analyses (FEA) were performed to determine strains in the solder joint due to thermal or mechanical cycling of the component. The strains could be due to lead bending resulting from the thermal mismatch of the board and chip and those resulting from local thermal mismatch between the lead and the solder, as well as between the board and the solder. Then the strains were used to set boundary conditions for an explicit model that could simulate initiation and growth of cracks in the microstructure of the solder joint. Finally, on the basis of the growth rate of the cracks in the solder joint, estimates were made of the cycles to failure for the electronic component.

Nasser and Curtin [22] applied PHM methodology to predict failure of the power supply. They subdivided the power supply into component elements based on specific material characteristics. Predicted degradation within any single or combination of component elements could be rolled up into an overall reliability prediction for the entire power supply system. Their PHM technique consisted of five steps: (i) acquiring the temperature profile using sensors; (ii) conducting FEA to perform stress analysis; (iii) conducting fatigue prediction of each solder joint; and (iv) predicting the probability of failure of the power supply system.

Searls et al. [23] undertook in situ environment loading, such as temperature measurements, in both notebook and desktop computers used in different parts of the world. In terms of the commercial applications of this approach, IBM has installed temperature sensors on hard drives (Drive–TIP) [24] to mitigate risks due to severe temperature conditions, such as thermal tilt of the disk stack and actuator arm, offtrack writing, data corruptions on adjacent cylinders, and outgassing of lubricants on the spindle motor.

Vichare et al. [25, 26] also conducted in situ health monitoring of notebook computers. The authors monitored and statistically analyzed the temperatures inside a notebook computer, including those experienced during usage, storage, and transportation, and discussed the need to collect such data both to improve the thermal design of the product and to monitor prognostic health. After the data was collected, it could be used to estimate the distributions of the load parameters. The usage history was used for damage accumulation and remaining-life prediction.

The European Union funded a project from September 2001 through February 2005 named Environmental Life-cycle Information Management and Acquisition (ELIMA) for consumer products, which aimed to develop ways of better managing the life cycles of products. It used technology to collect vital information during a product's life to lead to better and more sustainable products [27, 28]. The objective of this work was to provide a basic model for predicting the remaining lifetime of parts removed from products, based on the dynamic data collected by the ELIMA system. The ELIMA technology included sensors and memory built into the product to record dynamic data such as operation time, temperature, and power consumption. This was added to static data about materials and manufacturing. As a case study, the member companies monitored the application conditions of a game console and a household refrigerator. The work concluded that for the remaining-lifetime prediction, in general, it was essential that the environments associated with all life intervals of the equipment be considered. These included not only the operational and maintenance environments but also the preoperational environments, when stresses imposed on the parts during manufacturing, assembly, inspection, testing, shipping, and installation might have a significant impact on the eventual reliability of the equipment. Stresses imposed during the preoperational phase were often overlooked.

Tuchband and Pecht [29] presented the use of prognostics for military line replaceable units (LRUs) based on their life-cycle loads. The study was part of an effort funded by the Office of the Secretary of Defense to develop an interactive supply chain system for the US military. The objective was to integrate prognostics, wireless communication, and databases through a web portal to enable cost-effective maintenance and replacement of electronics. This study showed that prognostics-based maintenance scheduling could be implemented into military electronic systems. The approach involves an integration of embedded sensors on the LRU, wireless communication for data transmission, a PoF-based algorithm for data simplification and damage estimation, and a method for uploading this information to the Internet. Finally, the use of prognostics for electronic military systems enabled failure avoidance, high availability, and reduction of life-cycle costs.

3 Canary Devices Approach

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading

Canary devices mounted on the actual product have been used to provide advance warning of failure due to specific wearout failure mechanisms. The word “canary” is derived from one of coal mining's earliest systems for warning of the presence of hazardous gas using the canary bird. Because the canary is more sensitive to hazardous gases than humans, the death or sickening of the canary was an indication to the miners to get out of the shaft. The same approach, using canaries, has been employed in PHM. Canary devices were integrated into a specific component, device, or system design and incorporated failure mechanisms that occur first in the embedded device. These embedded canary devices (also called prognostics cell) were noncritical elements of the overall design providing early incipient failure warnings before actual system or component failure [12].

Mishra and Pecht [30] studied the applicability of semiconductor level health monitors by using precalibrated cells (circuits) located on the same chip with the actual circuitry. The prognostics cell approach was commercialized by Ridgetop Group to provide an early warning sentinel for upcoming device failures [31]. The prognostic cells were available for 0.35, 0.25, and 0.18 µm CMOS processes. The time to failure of these prognostic cells could be precalibrated with respect to the time to failure of the actual product. The stresses that contributed to degradation of the circuit included voltage, current, temperature, humidity, and radiation. Since the operational stresses were the same, the damage rate was expected to be the same for both the circuits. However, the prognostic cell was designed to fail earlier owing to increased stress on the cell structure by means of scaling. For example, scaling could be achieved by controlled increase of the current density inside the cells. With the same amount of current passing through both circuits, if the cross-sectional area of the current-carrying paths in the cells was decreased, a higher current density was achieved. Both the structure and the loading could be scaled. Further control in current density could be achieved by increasing the voltage level applied to the cells. Higher current density led to higher internal heating, causing greater stress on the cells. When a current of higher density passed through the cells, they were expected to fail faster than the actual circuit [30]. Currently, prognostic cells are available for semiconductor failure mechanisms such as electrostatic discharge (ESD), hot carrier, metal migration, dielectric breakdown, and radiation effects.

The extension of this approach to board-level failures was proposed by Anderson and Wilcoxon [32], who created canary components (located on the same PCB) that include the same mechanisms that lead to failure in actual components. Anderson et al., identified two prospective failure mechanisms: (i) low cycle fatigue of solder joints, assessed by monitoring solder joints on and within the canary package; and (ii) corrosion monitoring using circuits that are susceptible to corrosion. The environmental degradation of these canaries was assessed using accelerated testing, and degradation levels were calibrated and correlated to actual failure levels of the main system.

Goodman et al. [33] used a prognostic cell to monitor time-dependent dielectric breakdown (TDDB) of the metal-oxide semiconductor field-effect transistor (MOSFET) on the integrated circuits. The prognostic cell was accelerated to failure under certain environmental conditions. Acceleration of the breakdown of an oxide could be achieved by applying a voltage higher than the supply voltage, to increase the electric field across the oxide. When the prognostics cell failed, a certain fraction of the circuit lifetime was used up. The fraction of consumed circuit life was dependent on the amount of overvoltage applied and could be estimated from the known distribution of failure times.

4 PoF-Based PHM Implementation Approach

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading

The general PHM methodology is summarized in CALCE and shown in Figure 1. It includes two main parts: virtual life assessment and real prognostics assessment. Design data, expected life cycle, failure modes, mechanisms, and effects analysis (FMMEA), and PoF models can be the input for the virtual life assessment to get a better reliability assessment based on the historical and current data. The next step is to get the in situ sensor data, bus monitor data, diagnostics data, and maintenance records for prognostics assessment in real product life cycles. Three methodologies for in situ prognostics assessment include (i) the use of expendable devices, such as “canaries” and fuses that fail earlier than the host product to provide advance warning of failure; (ii) the monitoring of parameters and the extraction of features that are precursors to impending failure [34]; and (iii) the use of stress and damage models that employ life-cycle loads (e.g., usage conditions, temperature, vibration, radiation). Approaches 1 and 3 are PoF related [16], so in this article, the focus is on detailed PoF-based prognostics, which is discussed in a later paragraph.

thumbnail image

Figure 1. CALCE PHM methodology.

The PoF methodology is founded on the premise that failures result from fundamental mechanical, chemical, electrical, thermal, and radiation processes. The objective of the PoF methodology in the PHM process is to calculate the cumulative damage accumulation due to various failure mechanisms for a product in a given environment. The approach to implement PoF into PHM can be based on the FMMEA, which is shown in Figure 2. This approach consists of design capture, identification of potential failure, and reliability assessment [15].

thumbnail image

Figure 2. FMMEA-based PHM approach.

Design capture is the process of collecting structural (dimensional) and material information about a product to generate a model. This step involves characterizing the product at all levels, i.e., parts, systems, as well as physical interfaces. The potential failure identification step involves using the geometry and material properties of the product, together with the measured life-cycle loads acting on the product, to identify the potential failure modes, mechanisms, and failure sites in the product. This task is best performed through virtual qualification, which is a simulation-based methodology used to identify and rank the potential failure mechanisms. The reliability assessment step involves identification of appropriate PoF models for the identified failure mechanisms. A load–stress analysis is conducted using material properties, product geometry, and the life-cycle loads. With the computed stresses and the failure models, an analysis is conducted to determine the cycles to failure and then the accumulated damage is estimated using a damage model. Actually, the PoF methodology can provide a systematic approach to reliability assessment early in the design process.

4.1 Failure Mode, Mechanism, and Effect Analysis

A failure mode is the effect by which a failure is observed to occur [35]. It can also be defined as the way in which a component, subsystem, or system could fail to meet or deliver the intended function. All possible failure modes for each identified element should be listed. Potential failure modes may be identified using numerical stress analysis, accelerated tests to failure (e.g., HALT), past experience, and engineering judgment. The failure mode needs to be observable directly by methods such as visual inspection, electrical measurement, or other tests and measurements.

A failure cause is defined as the specific process, design, and/or environmental condition that initiated the failure, whose removal will eliminate the failure. Knowledge of potential failure causes can help identify the failure mechanisms driving the failure modes for a given element. One method of looking for causes is to review the life-cycle loads item by item to evaluate whether any of the loads there can cause the failure.

Failure mechanisms are the physical, chemical, thermodynamic, or other processes that result in failure. Failure mechanisms are categorized as either overstress or wearout mechanisms. Overstress failure arises because of a single load (stress) condition, which exceeds a fundamental material strength. Wearout failure arises because of cumulative damage due to loads (stresses) applied over an extended time or number of cycles. Within current technology, PHM can only be applied in the wearout failure mechanisms. Typical wearout failure mechanisms for electronics have been summarized in Table 1 [36].

Table 1. Failure Mechanisms, Relevant Loads, and Models in Electronics
Failure mechanismsFailure sitesRelevant loadsFailure models
  1. Δ, cyclic range; V, voltage; T, temperature; S, stress; ∇, gradient; M, moisture; J, current density; H, humidity.

FatigueDie attach, wirebond/TAB, solder leads, bond pads, traces, vias/PTHs, interfacesΔT, Tmean, dT/dt, dwell time, ΔH, ΔVNonlinear power law (Coffin–Manson)
CorrosionMetallizationsM, ΔV, TEyring (Howard)
ElectromigrationMetallizationT, JEyring (Black)
Conductive filament formationBetween metallizationM, ∇VPower law (Rudra)
Stress driven diffusion voidingMetal tracesS, TEyring (Okabayashi)
Time-dependent dielectric breakdownDielectric layersV, TArrhenius (Fowler–Nordheim)

Failure models help quantify the failure through evaluation of time to failure or likelihood of a failure for a given geometry, material construction, environmental, and operational condition. For wearout mechanisms, failure models use both stress and damage analysis to quantify the damage accumulated in the product.

When using the canary devices PHM approach, the geometries or material properties of the prognostics cell can be scaled to accelerate the failure under user conditions, on the basis of potential failure mechanisms. When using the modeling of stress and damage approach, environmental and usage load profiles are captured using sensors. Sensor data is then converted into a format that can be used in the failure models.

In the life cycle of a product, several failure mechanisms may be activated by different environmental and operational parameters acting at various stress levels, but, in general, only a few operational and environmental parameters, and failure mechanisms, are responsible for most failures. High priority mechanisms are those with high combinations of occurrence and severity. Prioritization of the failure mechanisms provides an opportunity for effective utilization of resources.

4.2 Life-Cycle Loading Monitoring

The life-cycle environment of a product consists of manufacturing, shipment, storage, handling, operating, and nonoperating conditions. The life-cycle loads (thermal, mechanical, chemical, electrical, and so on), either individually or in various combinations, may lead to performance or physical degradation of the product and may reduce its service life [1]. The extent and rate of product degradation depends on the magnitude and duration of exposure (usage rate, frequency, and severity) of such loads. If one can measure these loads in situ, the load profiles can be used in conjunction with damage models to assess the degradation due to cumulative load exposures. The typical life-cycle loads have been summarized in Table 2 [12].

Table 2. Life-Cycle Loads
LoadLoad conditions
ThermalSteady-state temperature, temperature ranges, temperature cycles, temperature gradients, ramp rates, heat dissipation
MechanicalPressure magnitude, pressure gradient, vibration, shock load, acoustic level, strain, stress
ChemicalAggressive versus inert environment, humidity level, contamination, ozone, pollution, fuel spills
PhysicalRadiation, electromagnetic interference, altitude
ElectricalCurrent, voltage, power

4.3 Data Reduction and Load Feature Extraction

Experience has shown that even the simplest data collection systems can accumulate vast amounts of data quickly, requiring either a frequent download procedure or a large-capacity storage device [37]. The main reasons for using data reduction in life-consumption monitoring are reduction of storage space; reduction in data-logger CPU load; and alignment with life-prediction models. The efficiency measures of data reduction methods should consider gains in computing speed and testing time; the ability to condense load histories without sacrificing important damage characteristics; and estimation of the error introduced by omitting data points.

The CALCE group has studied the accuracy associated with a number of data reduction methods such as ordered overall range (OOR), rainflow cycle counting, range-pair counting, peak counting, level crossing counting, fatigue meter counting and range counting.

Embedding the data reduction and load parameter extraction algorithms in to the sensor modules as suggested by Vichare et al. [25] can lead to a reduction in on-board storage space, lower power consumption, and uninterrupted data collection over longer durations. As shown in Figure 3, a time–load signal can be monitored in situ using sensors, and further processed to extract (in this case) cyclic range (Δs), cyclic mean load (Smean), rate of change of load (ds/dt), and dwell time (tD) using embedded load extraction algorithms. The extracted load parameters can be stored in appropriately binned histograms to achieve further data reduction. After the binned data is downloaded, it can be used to estimate the distributions of the load parameters. This type of output can be readily input into fatigue damage accumulation models.

thumbnail image

Figure 3. Load feature extraction.

In Vichare's study [26, 38], the temperature data was processed using two algorithms: (i) OOR to convert an irregular time–temperature history into peaks and valleys and also to remove noise due to small cycles and sensor variations and (ii) a three-parameter rainflow algorithm to process the OOR results to extract full and half cycles with cyclic range, mean, and ramp rates. The approach also involved optimally binning data in a manner that provides the best estimate of the underlying probability density function of the load parameter. The load distributions were developed using nonparametric histogram and kernel density estimation methods. The use of the proposed binning and density estimation techniques with a prognostic methodology was demonstrated on an electronic assembly.

4.4 Damage Assessment and Remaining-Life Calculation

Temperature and vibration are the most common causes of electronics failure [38]. The PoF models used to calculate the damage caused by temperature and vibration loadings are summarized in Figure 4. Damage caused by temperature can be calculated in time domain using Coffin–Manson's model. This approach has been demonstrated by Mishra et al. [16], Vichare et al. [39], and Cluff et al. [40]. Damage caused by vibration can be calculated in both time and frequency domains. The time domain, which has been demonstrated by Gu et al. [20], can use Basquin's model. The frequency domain, which has been demonstrated by Mishra et al. [16], can use first-order Steinberg's model [41].

thumbnail image

Figure 4. Damage calculation approach for temperature and vibration data.

4.5 Uncertainty Implementation and Assessment

The PoF model can be used to calculate the remaining useful life. However, it may still not be possible to make logistics decisions with certainty. Hence, it is necessary to identify the uncertainties in the prognostic approach and assess the impact of these uncertainties on the remaining-life distribution to make risk-informed decisions. Uncertainty analysis for prognostics implementations gives the prediction more meaning. With uncertainty analysis, a prediction can be expressed as a distribution rather than a single point. The prediction can be expressed as a failure probability.

Gu et al. [42] implemented the uncertainty analysis of prognostics for electronics under vibration loading. Gu identified the uncertainty sources and categorized them into four different types: measurement uncertainty, parameter uncertainty, failure criteria uncertainty, and future usage uncertainty. Then, the approach for implementing the uncertainty analysis was presented and shown in Figure 5 [42]. It utilized a sensitivity analysis to identify the dominant input variables that influence the model output. With information of input parameter variable distributions, a Monte Carlo simulation was used to provide a distribution of the accumulated damage. From the accumulated damage distributions, the remaining life was then predicted with confidence intervals. A case study was also presented, which used an experiment with an electronic board under vibration loading and a step by step demonstration of the uncertainty analysis implementation. The results showed that the experimentally measured failure time was within the bounds of the uncertainty analysis prediction.

thumbnail image

Figure 5. Uncertainty implementation for prognostics.

5 Application of PoF Implementation for PHM

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading

Prognostics based on the PoF model can be used in different areas, such as new products, legacy systems, and storage reliability prediction. Examples are discussed in the following paragraphs, along with a comparison of PoF-based PHM and data-driven-based PHM.

If the new product has not been manufactured, it is impossible to use the data-driven method since there will be no data available for training the algorithm. The PoF method normally only needs to change the material properties or geometries to model the new products, since most new products are not brand new and they have several previous versions or similar products can be referenced. Then it can save the time for performing prognostics for new products. On the other hand, the PoF approach makes it possible to give guidance for new product designs. The PoF approach incorporates reliability into the design process by establishing a scientific basis for evaluating new materials, structures, and electronics technologies. In addition, it can give feedback (such as failure site and failure mode) from life-cycle monitoring. Therefore PoF-based PHM is suitable for new products, and it reduces the design margin.

The legacy system continues to be used because of the prohibitive time and/or cost of replacing or redesigning it, though it is often less competitive and less compatible with modern equivalents [43]. Legacy systems can be found in both military and commercial sectors. PoF-based PHM to legacy systems can provide significant benefits. In the meantime, the data-driven approach lacks sufficient data to train the algorithm. Few legacy systems are in existence, and thus there are no extras for training purpose. The PoF-based PHM approach is based on an understanding of the life-cycle conditions of the legacy system and its failure modes and mechanisms. The first step is to use all available information (such as previous loading conditions, maintenance records, and so on) to assess the health status of the legacy system. The second step is to calibrate the health status using individual unit data so that an assessment of individual legacy systems' health can be derived. The third step involves the use of sensors and prognostic algorithms to update the health status on a continual basis to provide the most up-to-date prognosis of the system [44].

Improper storage is one factor that can precipitate failures in electronic products, especially in a hostile environment [45]. The temperature and humidity of storage areas are typically prime environmental factors. The extent and rate of product degradation depends upon the magnitude and duration of exposure (usage rate, frequency, and severity) to such loads. If one can measure these loads in situ, the load profiles can be used in conjunction with damage models to assess the degradation due to cumulative load exposures. Depending on the quality of the storage spaces, environmental factors such as vibration, shock, fungi, sand and dust, and radiation might also come into the picture. Hence, the PoF-based PHM method is ready to be implemented. In the meantime, the limitation for the data-driven approach is that it can only detect failure when near the failure point. It is difficult to assess the remaining life from the beginning or middle of storage. The PoF approach will be more suitable since, for storage conditions, the loading will not change frequently, and dwell loading will become the dominant effect. On the basis of the PoF model, it is possible to assess the product reliability after a period of being stored, to indicate the remaining life, and to determine whether it can survive the next mission.

6 Conclusions

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading

Traditional reliability predictions based on handbook methods are generally inaccurate and misleading. In this article, we have shown that PoF-based PHM is more suitable for reliability assessment. PHM can provide previously unknown information on life-cycle environmental and operational conditions, and previously unknown information on failure modes and mechanisms. It can also help prevent premature failures, and provide information on remaining life. In the future, owing to the increasing amount of electronics in the world and the competitive drive toward more reliable products, PoF-based PHM is being looked upon as a cost-effective solution to improve the reliability of electronic products and systems. Currently, more research should be focused on building physics-based damage models for electronics, obtaining the life-cycle data of product, and assessing uncertainty in remaining useful life prediction to make the PHM more realistic. Along with that, advance sensor technologies, communication technologies, decision making methods, and return of investment methods also need to be investigated.

References

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading
  • 1
    Bowles JB. A survey of reliability prediction procedures for microelectronic devices. IEEE Transactions on Reliability 1992 41(1): 212.
  • 2
    Sinnadurai N, Shukla AA, Pecht M. A critique of the reliability analysis center failure rate model for plastic encapsulated microcircuits. IEEE Transactions on Reliability 1998 47(2): 110113.
  • 3
    Telcordia SR-332, Reliability Prediction Procedure (RPP) for Electronic Equipment, 2001.
  • 4
    FIDES Group, FIDES Guide Issue A: Reliability Methodology for Electronic Systems, 2004.
  • 5
    Wong KL. What is wrong with the existing reliability prediction methods? Quality and Reliability Engineering International 1990 6: 251258.
  • 6
    Pecht M, Kang WC. A critique of MIL-Hdbk-217E reliability prediction methods. IEEE Transactions on Reliability 1988 37(5): 453457.
  • 7
    Pecht M, Nash F. Predicting the reliability of electronic equipment. Proceedings of the IEEE 1994 82(7): 9921004.
  • 8
    IEEE Standard 1413–1998, IEEE Standard Methodology for Reliability Prediction and Assessment for Electronic Systems and Equipment. IEEE, December 1998.
  • 9
    IEEE Standard 1413.1–2002, IEEE Guide for Selecting and Using Reliability Predictions Based on IEEE 1413. IEEE, February 2003.
  • 10
    Hess A, Calvello G, Frith P, Engel SJ, Hoitsma D. Challenges, issues, and lessons learned chasing the “Big P”: real predictive prognostics part 2. Aerospace Conference. IEEE, March 2006, pp. 119.
  • 11
    Gu J, Vichare N, Tracy T, Pecht M. Prognostics implementation methods for electronics. 53rd Annual Reliability and Maintainability Symposium (RAMS). Orlando, FL, 2007.
  • 12
    Vichare N, Pecht M. Prognostics and health management of electronics. IEEE Transactions on Components and Packaging Technologies 2006 29(1): 222229.
  • 13
    Leonard CT, Pecht M. Improved techniques for cost effective electronics. Proceedings of the Annual Reliability and Maintainability Symposium. Orlando, FL, 1991; pp. 174182.
  • 14
    DoD 5000.2 Policy Document, Defense Acquisition Guidebook, Chapter 5.3—Performance Based Logistics, December 2004.
  • 15
    Ramakrishnan A, Pecht M. Life consumption monitoring methodology for electronic systems. IEEE Transactions on Components and Packaging Technologies 2003 26(3): 625634.
  • 16
    Mishra S, Pecht M, Smith T, McNee I, Harris R. Remaining life prediction of electronic products using life consumption monitoring approach. Proceedings of the European Microelectronics Packaging and Interconnection Symposium. Cracow, 16–18 June 2002; pp. 136142.
  • 17
    Shetty V, Das D, Pecht M, Hiemstra D, Martin S. Remaining life assessment of shuttle remote manipulator system end effector. Proceedings of the 22nd Space Simulation Conference. Ellicott City, MD, 21–23 October 2002.
  • 18
    Mathew S, Das D, Osterman M, Pecht M, Ferebee R. Prognostic assessment of aluminum support structure on a printed circuit board. International Journal of Performability Engineering 2006 2(4): 383395.
  • 19
    Mathew S, Das D, Osterman M, Pecht M, Ferebee R, Clayton J. Virtual remaining life assessment of electronic hardware subjected to shock and random vibration life cycle loads. Journal of the IEST 2007 50(1): 8697.
  • 20
    Gu J, Barker D, Pecht M. Prognostics implementation of electronics under vibration loading. Microelectronics Reliability Journal 2007 47(12): 18491856.
  • 21
    Simons JW, Shockey DA. Prognostics modeling of solder joints in electronic components. Aerospace Conference. IEEE, March 2006.
  • 22
    Nasser L, Curtin M. Electronics reliability prognosis through material modeling and simulation. Aerospace Conference. IEEE, March 2006.
  • 23
    Searls D, Dishongh T, Dujari P. A strategy for enabling data driven product decisions through a comprehensive understanding of the usage environment. Proceedings of IPACK'01. Kauai, HI, 8–13 July 2001.
  • 24
    Herbst G. IBM's Drive Temperature Indicator Processor (Drive-TIP) Helps Ensure High Drive Reliability, IBM White Paper, October 1997, http://www.hc.kz/pdf/drivetemp.pdf (accessed Sep 2005).
  • 25
    Vichare N, Rodger P, Eveloy V, Pecht M. Environment and usage monitoring of electronic products for health assessment and product design. International Journal of Quality Technology and Quantitative Management 2007 4(2): 235250.
  • 26
    Vichare N, Rodgers P, Eveloy V, Pecht MG. In-situ temperature measurement of a notebook computer—a case study in health and usage monitoring of electronics. IEEE Transactions on Device and Materials Reliability 2004 4(4): 658663.
  • 27
    Bodenhoefer K. Environmental life cycle information management and acquisition—first experiences and results from field trials. Proceedings of Electronics Goes Green 2004+. Berlin, 5–8 September 2004; pp. 541546.
  • 28
    ELIMA Report, D-19 Final Report on ELIMA Prospects and Wider Potential for Exploitation, April 2005, http://www.ELIMA.org (accessed Dec 2005).
  • 29
    Tuchband B, Pecht M. The use of prognostics in military electronic systems. Proceedings of the 32nd GOMACTech Conference. Lake Buena Vista, FL, 19–22 March 2007; pp. 157160.
  • 30
    Mishra S, Pecht M. In-situ sensors for product reliability monitoring. Proceedings of the SPIE 2002 4755: 1019.
  • 31
    Ridgetop Semiconductor-Sentinel ™Silicon Library, Hot Carrier (HC) Prognostic Cell, August 2004.
  • 32
    Anderson N, Wilcoxon R. Framework for prognostics of electronic systems. Proceedings of International Military and Aerospace/Avionics COTS Conference. Seattle, WA, 3–5 August 2004.
  • 33
    Goodman D, Vermeire B, Ralston-Good J, Graves R. A board-level prognostic monitor for MOSFET TDDB. IEEE Aerospace Conference. Big Sky, MT, 2006.
  • 34
    Zhang G, Kwan C, Xu R, Vichare N, Pecht M. An enhanced prognostic model for intermittent failures in digital electronics. IEEE Aerospace Conference. Big Sky, MT, March 2007.
  • 35
    Pecht M. Product Reliability, Maintainability, and Supportability Handbook. CRC Press: New York, 1995.
  • 36
    Vichare N, Rodgers P, Eveloy V, Pecht M. Environment and usage monitoring of electronic products for health (reliability) assessment and product design. IEEE Workshop on Accelerated Stress Testing and Reliability. Austin, TX, October 2005.
  • 37
    Harris R, McNee I. Physics of failure (PoF) approach to life consumption monitoring (LCM) for military vehicles, part 2. Third International Conference on Health and Usage Monitoring—HUMS. Melbourne, 2003; pp. 5564.
  • 38
    Vichare N, Rodgers P, Pecht M. Methods for binning and density estimation of load parameters for prognostics and health management. International Journal of Performability Engineering 2006 2(2): 149161.
  • 39
    Vichare N, Pecht M. Enabling electronic prognostics using thermal data. Proceedings of the 12th International Workshop on Thermal Investigation of ICs and Systems. Nice, Côte d'Azur, 27–29 September 2006.
  • 40
    Cluff K, Barker D, Robbins D, Edwards T. Characterizing the commercial avionics thermal environment for field reliability assessment. Proceedings-Institute of Environmental Sciences 1996: 5057.
  • 41
    Steinberg D. Vibration Analysis for Electronic Equipment, Third Edition. John Wiley & Sons, 2000.
  • 42
    Gu J, Barker D, Pecht M. Uncertainty assessment of prognostics implementation of electronics under vibration loading. AAAI Symposium on Artificial Intelligence for Prognostics. Arlington, VA, 2007.
  • 43
    Madisetti VK, Jung Y-K, Khan MH, Kim J, Finnessy T. Reengineering legacy embedded systems. IEEE Design and Test of Computers 1999 16(2): 3847.
  • 44
    Tuchband B, Vichare N, Pecht M. A method for implementing prognostics to legacy systems. Proceedings of IMAPS Military, Aerospace, Space and Homeland Security: Packaging Issues and Applications (MASH). Washington, DC, 6–8 June 2006.
  • 45
    Zhang Y, Pecht M, Lantz L. A case study of IC storage failures in Taipei trains. Microelectronics Reliability 1998 38: 18111816.

Further Reading

  1. Top of page
  2. Introduction to the Prediction of Reliability
  3. Modeling of Stress and Damage Utilizing Life-Cycle Loads
  4. Canary Devices Approach
  5. PoF-Based PHM Implementation Approach
  6. Application of PoF Implementation for PHM
  7. Conclusions
  8. References
  9. Further Reading