Sign up for the QMED & MD+DI Daily newsletter.
Reliability Technology for Manufacturers: Engineering Better Devices
Originally Published MDDI October 2001 The medical device industry is among the wide cross section of industries that benefit from reliability engineering. Four methods presented here offer manufacturers concrete techniques to improve the effectiveness of their medical devices.
October 1, 2001
13 Min Read
.svg?width=850&auto=webp&quality=95&format=jpg&disable=upscale "Reliability Technology for Manufacturers: Engineering Better Devices")
Originally Published MDDI October 2001
The medical device industry is among the wide cross section of industries that benefit from reliability engineering. Four methods presented here offer manufacturers concrete techniques to improve the effectiveness of their medical devices.
B.S. Dhillon
The history of reliability engineering goes back as far as World War II, but the serious application of reliability engineering concepts to medical devices is a much more recent development. The latter part of the 1960s are generally regarded as its real beginning, and many publications concerning medical device reliability appeared during this period. 1–3
Since then, several methods and techniques have been developed to analyze and ensure the reliability of engineering systems.4 Many of these methods and techniques can equally be used to ensure the reliability of medical devices and systems. This article presents widely used reliability analysis methods that can be used to improve the effectiveness of medical devices and equipment.
The scope of reliability engineering is extremely wide, encompassing many areas of engineering technology, from ensuring the success of space missions to delivering a steady supply of electric power in a variety of applications. It is useful as a means to improve the reliability and effectiveness of medical devices and systems as well. The four methods of reliability engineering most widely used in the industrial sector are failure rate estimation, fault tree analysis, the Markov method, and failure mode and effects analysis.
FAILURE RATE ESTIMATION
The failure rate estimation method is widely used in the industrial sector to estimate the failure rates of electronic equipment.5 During bid proposal and early design phases—when it is often referred to as the parts-count method—failure rate estimation is used to provide a quick estimate of a system's potential failure rate. The information needed to use the method includes generic part types and quantities, the equipment's use environment, and part quality levels. For a single use environment, the equipment or device failure rate is expressed by
Equation 1
where the following variables apply:
λd is the device failure rate, expressed in terms of failures per 106 hours.
n is the total number of different generic component classifications in the device under consideration.
qi is the quantity of generic part i.
λgp is the generic failure rate of generic part i, expressed in failures per 106 hours.
fq is the quality factor of generic part i.
Tabulated values for λgp and fq of various parts can be found in MIL-HDBK-217, Reliability Prediction of Electronic Equipment.5
As a design matures, more information becomes available, and the failure rates of device parts are estimated individually. Usually, MIL-HDBK-217 is used to estimate the failure rate of electronic parts. The part failure rates are then added to obtain a total-equipment failure rate. This value provides a clearer picture of the actual failure rate of the device under consideration than does one obtained by using equation 1.
An equation of the following form is typically used to estimate failure rates of electronic parts.5
Equation 2
where λp is the part failure rate usually expressed in failures per 106 hours, Θe is the factor that accounts for the influence of environment, Θq is the factor that accounts for part quality level, and λb is the base failure rate normally defined by a model relating the influence of temperature and electrical stresses on the part under consideration.
FAULT TREE ANALYSIS
Fault tree analysis (FTA) is one of the most widely used methods to analyze engineering designs with respect to their reliability in the industrial sector, particularly in nuclear power generation. FTA is event oriented, as opposed to the failure orientation of the failure mode and effects analysis (FMEA) method described subsequently. Furthermore, FTA is more expensive than FMEA.
FTA starts by identifying an undesirable event—called a top event—associated with a system. Events that might cause the top event are generated and connected by logic operators such as AND and OR. The AND gate provides a true (faults) output if all inputs are true (faults). The OR gate provides a true output if one or more inputs are true.
Figure 1. Fault tree symbols: (a) OR gate, (b) AND gate, (c) reluctant event, and (d) basic fault event. |
The construction of a fault tree proceeds by the generation of events in a successive manner until the events (basic fault events) need not be developed further. The fault tree itself is the logic structure relating the top event to the basic events. These relationships are depicted through the use of a large number of symbols.4,6 The four basic symbols used in the construction of fault trees are shown in Figure 1.
In the figure, the circle represents a basic fault event (e.g., failure of an elementary component); the basic fault-event parameters are failure probability, unavailability, and failure and repair rates. The rectangle denotes the resultant event that occurs from the combination of fault events through the input of a gate such as AND or OR.
The basic steps involved in performing fault tree analysis are as follows:
Define the system, the assumptions involved in the analysis, and the events or states that would constitute failure.
Establish a system block diagram indicating inputs, outputs, and interfaces if simplifying the scope of analysis is desirable.
Establish the top-level fault event.
Use fault tree logic and fault event symbols and apply deductive reasoning to identify what could cause the top-level fault event to occur.
Continue developing the fault tree by identifying causes for intermediate fault events (i.e., the fault events that can cause the top-level fault event to occur).
Develop the fault tree to the desired lowest level, that of the most basic fault events.
Analyze the completed fault tree qualitatively as well as quantitatively.
Identify appropriate corrective measures.
Document the analysis and take appropriate measures to rectify problem areas.
To obtain the output probability of failure of the OR and AND gates, the following two equations can be used:
For OR:
Equation 4
and for AND:
Equation 5
where the following are true: F0 is the OR gate's output fault event probability of occurrence; FA is the AND gate's output fault event probability of occurrence; n is the total number of independent input fault events; and Fi is the occurrence probability of the ith input fault event, for i = 1, 2, 3, . . ., n. (The method is described in detail in reference 7.)
THE MARKOV METHOD
The Markov method is a powerful reliability evaluation technique that can generally handle more cases than any other method. One important application is reliability analysis of repairable systems. The technique can also be used when the components are independent, or for systems involving dependent failure and repair modes.
The method proceeds by the enumeration of system states. The resulting differential equations are then solved to obtain various reliability measures. The only serious problem with the method is that as the number of system states increases, the calculations can often become unmanageable.
The Markov method is based on the following assumptions:
All transition rates (e.g., failure and repair rates) associated with the system under consideration are constant.
The transitional probability from one system state to another in the finite time interval Δt is given by αΔt, where a is the constant transition rate (e.g., failure or repair rate) from one system state to another.
All occurrences are independent of each other.
The probability of more than one transition in the finite time interval Δt from one system state to another is negligible—i.e., (αΔt) (αΔt)→0.
FAILURE MODE AND EFFECTS ANALYSIS
Failure mode and effects analysis (FMEA) was developed in the 1950s and is used to evaluate designs at their early stages in terms of reliability.8,9 This criteria is also very useful to highlight both the need for and the effects of design changes. The method involves listing all possible failure modes for each component with their effects on the device subsystems.
FMEA requires that the following steps be performed:4
Define the boundaries of the system under consideration and its associated detailed requirements.
List all system components and subsystems.
Identify and list each component's failure modes, including a clear description.
Assign a failure rate or failure probability to each component failure mode.
List each failure mode effect on the subsystem, system, and plant.
Enter remarks for each identified failure mode.
Review critical failure modes and take appropriate corrective measures.
There are many benefits of performing FMEA: it provides a systematic approach to classify hardware failures, lowers development time and cost, reduces engineering changes, and is easy to understand. It serves as a useful tool for more efficient test planning and highlights safety concerns. Furthermore, this method can improve customer satisfaction and serve as an effective tool to analyze small, large, and complex systems. The method is described in further detail in reference 7.
Perhaps most importantly, FMEA provides a safeguard against repeating the same mistakes in the future and improves communication among design interface personnel.10 The application of this method during the initial stages of medical device design can be very useful.
REFERENCES
1. BS Dhillon, "Bibliography of Literature on Medical Equipment Reliability," Microelectronics and Reliability 20 (1980): 737–742.
2. BS Dhillon, Reliability Engineering in Systems Design and Operation, (New York: Van Nostrand Reinhold, 1983).
3. BS Dhillon, Medical Device Reliability and Associated Areas (Boca Raton, FL: CRC Press, 2000).
4. BS Dhillon and C Singh, Engineering Reliability: New Techniques and Applications (New York: John Wiley, 1981).
5. Reliability Prediction of Electronic Equipment, MIL-HDBK-217 (Washington DC: Department of Defense).
6. RJ Schroder, "Fault Tree for Reliability Analysis," in Proceedings of the Annual Symposium on Reliability (1970), 206–210.
7. BS Dhillon, Design Reliability: Fundamentals and Applications (Boca Raton, FL: CRC Press, 1999).
8. JS Countinho, "Failure Effect Analysis," Transactions of the New York Academy of Sciences 26 (1964): 564–584.
9. BS Dhillon, "Failure Mode and Effects Analysis-Bibliography," Microelectronics and Reliability 32 (1992): 719–731. 10. P Palady, Failure Modes and Effects Analysis (West Palm Beach, FL: PT Publications, 1995).
Copyright ©2001 Medical Device & Diagnostic Industry
You May Also Like
