IEEE Power Electronics Magazine - December 2016 - 31

Failure Rate λ (t )

b 1 1). The second part is dominated by random failures
in the useful life of a product, with a constant failure rate
(where b = 1). The third part is dominated by wear-out failures in the end of life of the product, with an increasing failure rate (where b 2 1).
The reliability/unreliability function (or CDF curve) provides a clear picture of the probability of failure as well as
the characteristics of failures developing with time. By this
function, many other reliability metrics can be deduced.
The early failure shown in Figure 3 is more related to the
production capability of a manufacturer, while the failures
in the useful life of a product are more related to random
events/usages. These two types of failures are normally difficult to model and predict.

Early Failure
Random Failure
Wear-Out Failure
Total
β <1
β =1

Time in Operation
FIG 3 The bathtub curve, or failure rate, in the life cycle of a
typical power electronics product.

Mission Profiles
The mission profiles indicate the operating conditions and
functions that a power electronics converter needs to perform during its specified life cycle. Typical mission profiles
for power electronics could be defined as, e.g., the wind
speed/solar irradiance (for the renewable energy production), speed and torque variations of the electric machine
(for the motor drive application), the operational ranges/
codes for the output voltage/current, the usage behaviors of
customers, and also the environmental factors like temperature, humidity, vibration level, etc. The mission profiles are
closely related to the stress/loading of the components,
which is the main source of the failures in the power electronics components. Meanwhile, the mission profiles will
lead to various design solutions for the converter and have a
strong impact on the cost of products. As a result, understanding the mission profiles is also a fundamental step for
the reliability analysis of power electronics.

new approach to assess the reliability
Metrics of Power Electronics
Assessing the reliability metrics of a converter system, which
contains many different kinds of components with complex
operating/loading conditions, is still a challenging task. Today,
most reliability information of power electronics components
has to be collected by statistical analysis of failed products
[7]-[11], and then experience-based handbooks are established as guidance to determine the lifetime of different
devices and designs. This approach has been proven inaccurate, as it is too general and application independent. An
alternative method for assessing the reliability metrics of a
converter is shown in Figure 4 [6], which separates the problem into four different groups of research/modeling activities.
In this structure, the critical components, as well as the major
failure mechanisms, in a converter system need first to be filtered out and identified. Based on the interested failure
mechanisms in the critical components, the stress experienced by these components is translated from the given mission profiles and design solution of a converter. On the other
hand, the reliability of the interested components under different stress levels in practical use is tested and modeled to

β >1

Identification
* Critical Component
* Failure Mechanism

Stress Analysis
Stress of Component
at Given Mission
Profiles and
Converter Design

Strength Modeling
Reliability of Component
at Different Stress Levels

Reliability Mapping
Reliability of Converter
at Given Mission
Profiles and
Converter Design

FIG 4 Analysis flow for assessing the reliability metrics of a
power electronics converter [6].

reflect the strength information of the components. Finally, a
series of algorithms and statistical analysis is introduced to
map the overall reliability metrics of the whole converter system under the given mission profiles and design solutions.
The potential methodologies in the four different activities in
Figure 4 are briefly discussed later.

Identification of Critical Components
and Failure Mechanism
Understanding of the reliability physics of power electronic
components is a starting point to assess reliability of the converter system. The target of this group of analyses is to identify the most critical components/assemblies in a system and
then the major failure mechanisms, as well as the corresponding stressors, which trigger the failure of components.
This information normally has to rely on the investigations
and statistics based on the failed products/components in
the field. Various investigations have been done into the failure causes and distributions in power electronics systems
December 2016

z IEEE PowEr ElEctronIcs MagazInE

Table of Contents for the Digital Edition of IEEE Power Electronics Magazine - December 2016