robin/Robin_PHD
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
07c5788e1e
Robin_PHD/papers/JOURNAL_fmea_sw_hw/sw_hw_fmea.tex

1544 lines
68 KiB
TeX
Raw Blame History

%%% OUTLINE
%\documentclass[twocolumn]{article}
\documentclass{article}
%\documentclass[twocolumn,10pt]{report}
\usepackage{graphicx}
\usepackage{fancyhdr}
%\usepackage{wassysym}
\usepackage{tikz}
\usepackage{amsfonts,amsmath,amsthm}
\usetikzlibrary{shapes.gates.logic.US,trees,positioning,arrows}
%\input{../style}
\usepackage{ifthen}
\usepackage{lastpage}
\usetikzlibrary{shapes,snakes}
\newcommand{\tickYES}{\checkmark}
\newcommand{\fc}{fault~scenario}
\newcommand{\fcs}{fault~scenarios}
\date{}
%\renewcommand{\encodingdefault}{T1}
%\renewcommand{\rmdefault}{tnr}
%\newboolean{paper}
%\setboolean{paper}{true} % boolvar=true or false
\newcommand{\derivec}{{D}}
\newcommand{\ft}{\ensuremath{4\!\!\rightarrow\!\!20mA} }
\newcommand{\permil}{\ensuremath{{ }^0/_{00}}}
\newcommand{\oc}{\ensuremath{^{o}{C}}}
\newcommand{\adctw}{{${\mathcal{ADC}}_{12}$}}
\newcommand{\adcten}{{${\mathcal{ADC}}_{10}$}}
\newcommand{\ohms}[1]{\ensuremath{#1\Omega}}
\newcommand{\fm}{failure~mode}
\newcommand{\fms}{failure~modes}
\newcommand{\fg}{functional~grouping}
\newcommand{\FG}{\mathcal{G}}
\newcommand{\DC}{\mathcal{DC}}
\newcommand{\fgs}{functional~groupings}
\newcommand{\dc}{derived~component}
\newcommand{\dcs}{derived~components}
\newcommand{\bc}{base~component}
\newcommand{\FMMD}{ModularFMEA}
\newcommand{\bcs}{base~components}
\newcommand{\irl}{in real life}
\newcommand{\enc}{\ensuremath{\stackrel{enc}{\longrightarrow}}}
\newcommand{\pin}{\ensuremath{\stackrel{pi}{\longleftrightarrow}}}
%\newcommand{\pic}{\em pure~intersection~chain}
\newcommand{\pic}{\em pair-wise~intersection~chain}
\newcommand{\wrt}{\em with~respect~to}
\newcommand{\abslevel}{\ensuremath{\Psi}}
\newcommand{\fmmdgloss}{\glossary{name={FMMD},description={Failure Mode Modular De-Composition, a bottom-up methodology for incrementally building failure mode models, using a procedure taking functional groups of components and creating derived components representing them, and in turn using the derived components to create higher level functional groups, and so on, that are used to build a failure mode model of a system}}}
\newcommand{\fmodegloss}{\glossary{name={failure mode},description={The way in which a failure occurs. A component or sub-system may fail in a number of ways, and each of these is a
failure mode of the component or sub-system}}}
\newcommand{\fmeagloss}{\glossary{name={FMEA}, description={Failure Mode and Effects analysis (FMEA) is a process where each potential failure mode within a system, is analysed to determine system level failure modes, and to then classify them {\wrt} perceived severity}}}
\newcommand{\frategloss}{\glossary{name={failure rate}, description={The number of failure within a population (of size N), divided by N over a given time interval}}}
\newcommand{\pecgloss}{\glossary{name={PEC},description={A Programmable Electronic controller, will typically consist of sensors and actuators interfaced electronically, with some firmware/software component in overall control}}}
\newcommand{\bcfm}{base~component~failure~mode}
\newcommand{\cf}[1]{\textbf{#1()}}
\def\layersep{1.8cm}
\newboolean{pld}
\setboolean{pld}{false} % boolvar=true or false : draw analysis using propositional logic diagrams
\newboolean{dag}
\setboolean{dag}{true} % boolvar=true or false : draw analysis using directed acylic graphs
% \setlength{\topmargin}{0in}
% \setlength{\headheight}{0in}
% \setlength{\headsep}{0in}
% \setlength{\textheight}{22cm}
% \setlength{\textwidth}{18cm}
% %\setlength{\textheight}{24.35cm}
% %\setlength{\textwidth}{20cm}
% \setlength{\oddsidemargin}{0in}
% \setlength{\evensidemargin}{0in}
% \setlength{\parindent}{0.0in}
% %\setlength{\parskip}{6pt}
% % \setlength{\parskip}{1cm plus4mm minus3mm}
% \setlength{\parskip}{0pt}
% \setlength{\parsep}{0pt}
% \setlength{\headsep}{0pt}
% \setlength{\topskip}{0pt}
% \setlength{\topmargin}{0pt}
% \setlength{\topsep}{0pt}
% \setlength{\partopsep}{0pt}
% \setlength{\itemsep}{1pt}
% \renewcommand\subsection{\@startsection
% {subsection}{2}{0mm}%
% {-\baslineskip}
% {0.5\baselineskip}
% {\normalfont\normalsize\itshape}}%
\linespread{1.0}
\begin{document}
%\pagestyle{fancy}
%\fancyhf{}
%\fancyhead[LO]{}
%\fancyhead[RE]{\leftmark}
%\cfoot{Page \thepage\ of \pageref{LastPage}}
%\rfoot{\today}
%\lhead{Developing a rigorous bottom-up modular static failure mode modelling methodology}
%\lhead{Developing a rigorous bottom-up modular static failure modelling methodology}
% numbers at outer edges
\pagenumbering{arabic} % Arabic page numbers hereafter
\author{R.Clark$^\star$, A.~Fish$^\dagger$ , C.~Garrett$^\dagger$, J.~Howse$^\dagger$ \\
$^\star${\em Energy Technology Control, UK. r.clark@energytechnologycontrol.com} \and $^\dagger${\em University of Brighton, UK}
}
%\title{Developing a rigorous bottom-up modular static failure mode modelling methodology}
\title{Failure Mode Effects Analysis (FMEA) for Software/Hardware Hybrid Systems using a modular bottom-up hierarchical modelling methodology}
%\nodate
\maketitle
\today
\paragraph{Keywords:} static failure mode modelling; safety-critical; software fmea
%\small
\abstract{ % \em
%\input{abs}
The certification process of safety critical products for European and
other international standards often demand environmental stress,
endurance and Electro Magnetic Compatibility (EMC) testing. Theoretical, or 'static testing',
is often also required.
Failure Mode Effects Analysis (FMEA)~\cite{iec60812}, is a bottom-up technique that aims to assess the effect all
component failure modes on a system.
%
It is used both as a design tool (to determine weaknesses), and as a requirement of certification of safety critical products.
FMEA has been successfully applied to mechanical, electrical and hybrid electro-mechanical systems.
This paper discusses the benefits and drawbacks of current
FMEA techniques and then proposes a modular FMEA methodology,
Failure Mode Modular De-Composition (FMMD)~\cite{clark}
that has the advantages modularity, traceable failure modes throughout the model
hierarchy, increase in test efficiency
and has
the ability to model integrated hardware and software systems.
% Work on software FMEA (SFMEA) is beginning, but
% at present no technique for SFMEA that
% integrates hardware and software models % known to the authors
% exists.
% %
%
% %
% %Failure modes in components in say a sensor, could be traced
% %up through the electronics and then through the controlling software.
% %
% %Presently Failure Mode Effects Analysis (FMEA), stops at the glass ceiling of the computer program.
% This paper takes, from the literature, new and emerging methodologies
% for software FMEA, applies them to a simple example system, and then
% reaches conclusions about the effectiveness and failure mode
% coverage of the combined FMEA techniques.
A small, but complete embedded system, worked example is presented to show FMMD applied to an
integrated electronics/software system.
%, the industry standard
%{\ft} signalling loop.
%
} % abstract
\section{Introduction}
%FMEA methodologies trace from the 1940's and were designed to
%model simple electro-mechanical systems.
%
The origins of
FMEA methodologies were formed %originally designed
in the 1940's to
model failure effects in simple electro-mechanical systems.
%
Because those early systems were relatively simple,
%modern FMEA methodologies follow this paradigm and
they traced component failure modes directly to system level failures.
There were no concepts of modularity and, at that time, no need to include
software elements.
%
%This paper explores the historical reasons why FMEA is performed in the way it is currently and
%the new factors placing higher demands upon it.
%
Software generally sits on top of most modern safety critical control systems
and defines its most important system wide behaviour and communications.
%
Currently standards that demand FMEA investigations for hardware(HFMEA) (e.g. EN298, EN61508),
do not specify it for software but instead essentially just specify good practise,
i.e. review processes and language feature constraints.
%
That is to say there id no formal framework for following
failure modes from low level hardware elements through into the software models.
%
This is a weakness.
%
Where HFMEA % scientifically
traces component {\fms}
to resultant system failures, software until recently, has been left in a non-analytical
limbo of best practises and constraints.
%
Software FMEA has been proposed
in several forms~\cite{modelsfmea,sfmea,procsfmeadb,sfmeaauto}.
%
However, SFMEA is always performed separately from HFMEA.
%
% This paper seeks to examine the effectiveness of current and proposed SFMEA
% techniques, by analysing a simple hybrid hardware/software system,
% which is in common use and has mature field experience. %
% %analysing the chosen example, which is well known and understood
% %
% Because the chosen example is well understood it is
% %, this example is
% useful
% to compare the results from these FMEA methodologies with
% the known failure mode behaviour.
% %from years of field experience, and determining how well the HFMEA and SFMEA
% %analysis reports model the failure mode behaviour.
% % %
%If software and hardware integrated FMEA were possible, electro-mechanical-software hybrids could
%be modelled, and so we could consider `complete' failure mode models.
%
%Presently FMEA, stops at the glass ceiling of the computer program: FMMD seeks to address
%this, and offers additional test efficiency benefits.
This paper is a condensed version of the PhD thesis entitled `failure Mode Modular De-compositon'~\cite{clark}. % \today
%\today
\nocite{en298}
\nocite{en61508}
\section{Introduction}
{
%This paper describes a modular FMEA process that can be applied to software.
%This modular variant of FMEA is called Failure Mode Modular de-composition (FMMD).
%
%Because this process is based on failure modes of components,
%it can be applied to electrical and/or mechanical systems.
%
%The hierarchical structure of software is then examined,
%and definitions from contract programming are used
%to define failure modes and failure symptoms for
%software functions.
%
%With these definitions we can apply the FMMD modular form of FMEA
%to existing software\footnote{Existing software excluding recursive~\cite{misra}[16.2] code,
%and unstructured non-functional language.}.
}
\section{FMEA Background}
%What FMEA is, briefly variants...
FMEA~\cite{iec60812} is the process of taking
component failure modes, %and by reasoning,
tracing their effects through a system
and determining what system level failure modes could be caused.
%
%The certification process of safety critical products for European and
%other international standards often demand environmental stress, magnetic susceptibility,
%endurance and Electro Magnetic Compatibility (EMC) testing.
%
% Theoretical, or `static~testing', is often also required.
% %
% FMEA is a tool used for static testing.
% %
% For many types of safety critical system in the European Union, product design testing and FMEA
% is legally mandatory~\cite{en230,en298}.
% %
% %Its use is traditionally only applied to hardware (electrical and mechanical) systems.
% %
% %
% FMEA has its roots in the previous century where simple electro-mechanical systems were the norm.
% %
To perform FMEA, the effects of a component failure mode are examined
with respect to other components in the system; and from this behaviour
a system level failure or effect is determined.
%
That is an experienced engineer, will take a parts list, look up all the failure modes
for the parts and typically list them on a spreadsheet with one failure mode per row.
%
The Engineer will then, using available component data, circuit schematics and perhaps
experimental data, determine the effect on the system for each failure mode.
%
There will typically be attributes for each failures modes outcome.
%
These could include whether the failure is detectable, time to repair and the
perceived severity of the outcome.
%
This report is the end product of an FMEA investigation.
%
% Several variants of FMEA exist,
% but the three in main use are:
% \begin{itemize}
% \item Deisgn FMEA (DFMEA) is FMEA applied at the design or approvals stage~\cite{en298, en230}
% where the aim is to ensure that single component failures (at least) cannot
% cause unacceptable system level events~\cite{~\cite{iec60812}fmea},
% \item Failure Mode Effect Criticality Analysis (FMECA) is applied to determine the most potentially dangerous or damaging
% failure modes to fix, using FMEA in conjunction with severity and failure probability figures~\cite{fmeca,mil1991,fmd91},
% \item Failure Mode Effects and Diagnostics Analysis, is FMEA peformed to
% determine a statistical level of safety.
% This is associated with Safety Integrity Levels (SIL)~\cite{en61508}~\cite{en61511} classification.
% \end{itemize}
\subsection{Concept of `reasoning~distance'.}
\label{reasoningdistance}
%\fmmdglossRD
In order to evaluate the efficiency or performance of a process it is
often useful to count the number of operations required to perform it.
In computer science sorting algorithms are often classified by the order of operations using big `O' notation.
A bubble sort for instance has a worst case sorting time proportional to the number $N$ if elements of $O(N^2)$.
Because of this a programmer is very unlikely to implement a simple bubble sort
if the number of elements $N$ is predicted to be large.
To evaluate FMEA techniques a metric is required.
When analysing a failure mode of a component, it is reasonable to
look at how the failure mode will affect the other components in the system and to put this then
into the context of the systems behaviour.
Components may fail in several ways. European standard EN298~\cite{en298} gives the possible
failure modes for a resistor as $OPEN$ and $SHORT$ for instance.
The term $f$ is defined as the number of component failure modes for a given component.
A system will have $N$ number of components.
For each component then there will be $f \times (N-1)$ other components that could be affected by
it failing.
By counting the number of of checks to make, i.e. failure mode against all other components in the system,
a metric for for evaluating FMEA is defined.
%
This count of checks is defined as `reasoning~distance' ---or in other words is --- the number of stages of logic and reasoning used
in {\fm} analysis to map a failure cause to its potential outcomes; counted
by the number of {\fm} to other component analysis stages made.
%
%The basic FMEA example in section~\ref{basicfmea}
%considered one {\fm} against some of the components in the milli-volt reader.
%
To create an exhaustive FMEA report every
known failure mode of every component
within the system would have to be examined against all its other components.
% %
% `Reasoning~distance', for one {\fm}, is defined as the number of components checked against it
% to determine its system level symptom(s).
%
No current FMEA variant gives guidelines for the components that should
be included to analyse a {\fm} in a system.
%
Were a {\fm} examined against all the other components in a system
this would give a maximum reasoning distance.
%
This is termed the exhaustive FMEA (XFMEA) case for a single {\fm}.
%does not
% The exhaustive~reasoning~distance would be
% the sum of the number of failure modes, against all other components
% in that system.
Thus the exhaustive~reasoning~distance for a particular component
would be to multiply
the number of failure modes it has by the number of remaining components
in the system.
%
The exhaustive reasoning~distance for a system would be the
the sum of these multiplications for all the components it contains.
%
If a small system were to have say 100 components, with three failure modes per component, this
would give an exhaustive reasoning distance---for single failure analysis---of $3 \times 100 \times 99$.
%
The discussion on reasoning distance provides a metric to examine
the state explosion problems associated with FMEA ( and other forward search failure investigation
methodologies).
%
%\fmmdglossSTATEEX
%
It is apparent that the shorter the reasoning distance, the more precisely theoretical examination
can determine failure symptoms.
%
For instance for a very simple small circuit, a better understanding of failure effects is expected,
than for a very large system where there are more variables and potential {\fm} interactions.
%
%.... general concept... simple ideas about how complex a
%failure analysis is the more modules and components are involved
% cite for forward and backward search related to safety critical software
%{sfmeaforwardbackward}
\subsection{FMEA and the State Explosion Problem}
\label{sec:xfmea}
\paragraph{Problem of which components to check for a given {\bc} {\fm}.}
%\fmmdglossSTATEEX
%
FMEA for safety critical certification (i.e. for EN298 and EN61508)~\cite{en298,en61508} has to be applied
to all known failure modes of all components within a system.
%
Each one of these, in a typical report, would be one line of a spreadsheet entry.
%
FMEA does not define or specify the scope of the investigation for each component failure mode.
%
For instance should the signal path be followed, with all components encountered along that, or should the scope be wider?
%
%If we wethe effect of a component {\fm} against all other components
%in a system, this could be said to be exhaustive analysis.
\paragraph{Exhaustive Single Failure FMEA.}
%\fmmdglossXFMEA
%
To perform XFMEA, every possible interaction
of a failure mode with all other components in a system would have to be examined.
%
Or in other words, all possible failure scenarios considered.
%
%to do this completely (all failure modes against all components).
This is represented in the equation below, %~\ref{eqn:fmea_state_exp},
where $N$ is the total number of components in the system, $RD_{single}$ is the reasoning~distance and
$f$ is the number of failure modes per component:
%
\begin{equation}
\label{eqn:fmea_single}
RD_{single} = N.(N-1).f . % \\
%(N^2 - N).f
\end{equation}
%
This means an order of $O(N^2)$ checks to perform
to undertake XFMEA for single failures.
%
Even small systems have typically
100 components, and they typically have 3 or more failure modes each, which would give
$100 \times 99 \times 3 = 29,700 $ as a reasoning~distance.
%
%\fmmdglossSTATEEX
\paragraph{Exhaustive FMEA and double failure scenarios.}
%
%\paragraph{Exhaustive Double Failure FMEA}
For looking at potential double failure
scenarios\footnote{Certain double failure scenarios are already legal
requirements---The European Gas burner standard (EN298:2003~\cite{en298})---demands the checking of
double failure scenarios (for burner lock-out scenarios).}
%
(two components failing within a given time frame) and the order becomes $O(N^3)$.
Where $RD_{double}$ is the reasoning~distance for double failure scenarios:
\begin{equation}
\label{eqn:fmea_double}
RD_{double} = N.(N-1).(N-2).{f}^{2}% \\
%(N^2 - N).f
\end{equation}
%
For a theoretical system with 100 components and a fixed 3 failure modes each, this gives reasoning distance of
$100 \times 99 \times 98 \times 9 = 8,731,800 $. % failure mode scenarios.
%
In practise there is an additional complication; that of
the circuit topology changes that {\fms} can cause.
\paragraph{Reliance on experts for meaningful FMEA Analysis.}
Current FMEA methodologies cannot consider---for the reason of state explosion---an exhaustive approach.
%We define exhaustive FMEA ({\XFMEA}) as examining the effect of every component failure mode
%against the remaining components in the system under investigation.
%
%\fmmdglossSTATEEX
%
%Because for practical reasons,
In practical terms XFMEA cannot be performed for anything other than a trivial system,
reliance is placed upon experts on the system under investigation
to perform a meaningful analysis.
%
These experts must use their judgement and experience to choose
sub-sets of the components in the system to check against each {\fm}.
%
Also, %In practise
these experts have to select the areas they see as most critical for detailed FMEA analysis:
it is usually impossible, for reasons of time to perform the work,
to action a detailed level of analysis on all component {\fms}
on anything but a very small %hypothetical
system.
% \subsection{Component Tolerance}
%
% Component tolerances may need considering when determining if a component has failed.
% Calculations for acceptable ranges to determine failure or acceptable conditions
% must be made where appropriate.
% %
% An example of component tolerance considered for FMEA
% is given in section~\ref{sec:resistortolerance}.
%\section{FMEA in current usage: Five variants}
\section{FMEA in current usage: Four variants}
%\paragraph{Five main Variants of FMEA}
\paragraph{Four main Variants of FMEA}
\begin{itemize}
%\item \textbf{PFMEA - Production} Emphasis on cost reduction and product improvement;
\item \textbf{FMECA - Criticality} Emphasis on minimising the effect of critical systems failing~\cite{fmeca}; % Military/Space
\item \textbf{FMEDA - Statistical Safety} Statistical analysis giving Safety Integrity Levels~\cite{en61508};
\item \textbf{DFMEA - Design or Static/Theoretical} Approval of safety critical systems using FMEA and single or double failure prevention~\cite{en298};% EN298/EN230/UL1998
\item \textbf{SFMEA - Software FMEA} --- Usage not enforced by most current standards~\cite{en298,en230,en61508}. %only used in highly critical systems at present.
\end{itemize}
\nocite{MILSTD1629short}
\subsection{FMEA and modularity.}
Because modern electronics has become more complex the number
of basic components has risen dramatically.
To add to this components used to fulfil common functions are often Integrated Circuits (ICs)..
Typical examples include voltage regulators, op-amps, micro-controllers~\cite{pic18f2523}, memory modules and
protocol handlers~\cite{mcp2515}. To build any of these component from scratch would be very expensive and time consuming,
but these IC `components' have very high internal transistor counts, and each have their own unique
failure mode behaviour.
Thus modern electronics has already become too large in scope to sensibly implement the base component failure mode directly mapped to
a system failure paradigm.
The automotive industry, because of mass production, must make products that have high safety integrity %that are very safe but
% financial pressure keeps their products
but must also be affordable.
%
This leads to specialist firms producing modules, such as automatic braking systems,
that are bought in and assembled to make a auto-mobile.
%
Performing failure analysis using the basic component single failure modes to
system failure mapping, would be very difficult: this would require expert knowledge
of the design behaviour and component types used in each module.
%
Because modern systems have become more complex and now include software elements modularity
of some form, has become necessary to break down the state explosion problems associated with FMEA.
%
Some modular techniques are starting to be formalised, and are described below.
\paragraph{Automotive SIL (ASIL) --- modularisation of FMEDA}
%
The EN61508 variant for automotive use, as defined in standard ISO~26262, is known as Automotive SIL (ASIL)~\cite{Kafka20122}.
%
Because of the modular approach forced on automotive designers
a process has been developed called `ASIL~de-composition'~\cite{6464473}.
%
This allows automotive designers to use pre-certified modules in their designs
and applies broad statistical guidelines to achieving particular safety levels by
use of redundancy and automated diagnostics etc.
%
Note that the ASIL modules are given a relaibility rating which can be enhanced with redundancy.
It does not introduce traceable {\fm} reasoning in its hierarchy.
%%
%% IN SOFTWARE THIS WOULD BE TIGHTLY COUPLED AS OPPOSED TO LOOSELY COUPLED FUNCTIONS.
%
\paragraph{Indenture levels --- modularisation of FMECA}
The US military standard for FMECA~\cite{fmeca}, describes a very broad modularity regime, that
it terms `indenture' levels. Indenture levels are arranged from the top down
and identify finer and finer grained modules. For instance, an aircraft
may be the first indenture level, and the next may be an identifiable module such as
an altitude radar: within that finer grained modules may be identified until
the base components are listed. Note that this is a top down approach to modularisation and
this can introduce errors into the reliability calculations~\cite{MILSTD1629short}.
%
\paragraph{Modularisation in Software}
%
It is interesting to compare the development of FMEA methodologies with software.
Software expanded in complexity faster than electronics,
and to cope with this software languages developed modularity (function call trees, classes and finally distributed processing mechanisms).
%
FMEA has, by necessity, started to include some modular features but none yet
have defined mechanisms for ensuring that all failure modes
from a module must be considered in the analysis of the module(s)
that incorporate it.
\paragraph{Top Down or Bottom-up?}
% Because FMEA is a bottom up technique, applying a top down analysis (as in FMECAs indenture levels)
% cannot guarantee to consider all component failure modes in the correct context.
% %
A top down approach (such as FTA) can miss~\cite{faa}[Ch.~9] individual failure modes of components,
especially where there are non-obvious or unexpected top-level failures.
%
In order to ensure that every failure mode is considered, a bottom-up approach
including every base components {\fms} must be used.
%
Going back to the software analogy, the indenture levels of FMECA are similar to
a software call tree where the highest indenture levels would be leaf functions.
%
There is no equivalent of the software `class'.
%
In the real world however there are.
Off the shelf sensors can be purchased which communicate using standard protocols~\cite{Pfeiffer:2003:ENC:1199616}. % consider CANOpen standard sensors, these are%~\footnote{CANopen sensors...}
%modules connected by an industrial data bus.
%
These not only typically have electrical and mechanical
components, they have a firmware and communication bus aspects~\cite{canspec, caninauto}.
%
These type of modules combine hardware, electronics, software, communications
and distributed programming.
%
Current FMEA techniques struggle with software alone, and also, fail to integrate the analysis of hardware and software
systems~\cite{sfmea, embedsfmea, modelsfmea, sfmeaa}. %, sfmeainterface }.
%
\subsection{FMEA and software.}
In addition to increasing complexity in electronics, modern control systems nearly always have a significant software/firmware element,
and not being able to model software with current FMEA methodologies
is a cause for criticism~\cite{safeware}[Ch.12].
%
Similar difficulties in integrating mechanical and electronic/software
failure models are discussed in ~\cite{SMR:SMR580}.
%
Currently standards that demand FMEA for hardware (e.g. EN298~\cite{en298}, EN61508~\cite{en61508}),
do not specify it for software, but instead recommended computer-architectures, good software practise,
review processes and language feature constraints.
%
%
%Software FMEA techniques have been proposed
%FMMD is a modularisation of FMEA and can produce failure~mode models that can be used in
%all the above variants of FMEA.
\paragraph{Current work on Software FMEA}
SFMEA usually does not seek to integrate
hardware and software models, but to perform
FMEA on the software in isolation~\cite{procsfmea}.
%
Work has been performed using databases
to track the relationships between variables
and system failure modes~\cite{procsfmeadb}, to %work has been performed to
introduce automation into the FMEA process~\cite{appswfmea} and to provide code analysis
automation~\cite{modelsfmea}.
%
Although the SFMEA and hardware FMEAs are performed separately,
some schools of thought aim for Fault Tree Analysis (FTA)~\cite{nasafta,nucfta} (top down - deductive)
and FMEA (bottom-up inductive)
to be performed on the same system to provide insight into the
software hardware/interface~\cite{embedsfmea}.
%
Although this
would give a better picture of the failure mode behaviour, it
is by no means a rigorous approach to tracing errors that may occur in hardware
through to the top (and therefore ultimately controlling) layer of software.
\subsection{Current FMEA techniques are not suitable for software}
The main FMEA methodologies are all based on the concept of taking
base component {\fms}, and translating them into system level events/failures~\cite{sfmea,sfmeaa}.
%
That is there is only one stage of reasoning between the low level component {\fm} and
the system level symptom of failure.
%
In a complicated system, mapping a component failure mode to a system level failure
will mean a long reasoning distance; that is to say the actions of the
failed component will have to be traced through
several sub-systems, gauging its effects with and on other components.
%
With software at the higher levels of these sub-systems,
we have yet another layer of complication.
%
%In order to integrate software, %in a meaningful way
%we need to re-think the
%FMEA concept of simply mapping a base component failure to a system level event.
%
% SFMEA regards, in place of hardware components, the variables used by the programs to be their equivalent~\cite{procsfmea}.
% The failure modes of these variables, are that they could become erroneously over-written,
% calculated incorrectly (due to a mistake by the programmer, or a fault in the micro-processor on which it is running), or
% external influences such as
% ionising radiation causing bits to be erroneously altered.
It is desirable to trace failure modes effects through the hardware and software interfaces.
This is for two reasons.
%
The first is to ensure that the software can detect all possible
hardware failures, and secondly that the software reacts appropriately.
\section{FMEA defeciences and `wishlist'}
%\subsection{FMEA - General Criticism}
A summary of deficiencies in current FMEA methodologies is listed below:
\begin{itemize}
%\item FMEA type methodologies were designed for simple electro-mechanical systems of the 1940's to 1960's,
\item State explosion - %impossible
very difficult/time consuming to perform FMEA exhaustively, %rigorously
\item Difficult to re-use previous analysis work~\cite{rudov2009language},
\item Very difficult to model simultaneous/multiple failures,
\item Software and hardware models are separate (if the software is modelled at all) meaning the software interface may not be correctly modelled,
%\item reasoning distance -- component failure to system level symptom process is undefined in regard
%to the components to check against each given component {\fm},
\item FMEA methodologies are undefined in regard to which components to check against given failure modes,
%
\item Distributed real time systems are very difficult to analyse with FMEA because they typically involve many hardware/software interfaces.
\end{itemize}
Traditional forms of FMEA are no longer % fit for purpose!
of meaningful use for complex modern systems especially those incorporating programmatic elements.
%
They were designed to analyse simple electro-mechanical systems
and even common place high component count analogue circuits (that are usually surface mount and therefore physically small), are
getting too complicated for meaningful analysis using FMEA.
%
% With surface mount technology and increasingly dense integrated circuitry, electronics generally
% has much higher component counts and more complex components than those in use when FMEA
% was designed.
%
From the above defeciencies, a wish list for a better FMEA is presented, stating the features that should exist
in an improved FMEA methodology,
\begin{itemize}
\item Must be able to analyse hybrid software/hardware systems,
\item avoid state explosion (i.e. XFMEA is impractical by hand~\cite{cbds}),
\item encourage exhaustive checking within each modular, %(total failure coverage within {\fgs} all interacting component and failure modes checked),
\item traceable reasoning inherent in system failure models,% to aid repeatability and checking,
\item re-usable i.e. it should be possible to re-use analysis,
\item possibility to analyse simultaneous/multiple failures,
%\item one to one mapping from {\bc} {\fms} to system level failures (see section~\ref{sec:onetoone}),
\item modular --- i.e. usable in a distributed system.
% \item
\end{itemize}
\section{Proposed Methodology: Failure Mode Modular De-composition (FMMD)}
\section{The proposed Methodology}
\label{fmmdproc}
%
%% One line
The basic concept of FMMD is to modularise FMEA from the bottom-up: by choosing groups of components that
work together to perform a given function: the failure modes of the components
are analysed, and a failure mode behaviour for the group determined: this group
can now be used as a component in its own right with a set of its own failure modes.
%
% In essence, this methodology beginning with low level modules (or {\fgs})
% which are analysed and assigned a failure mode behaviour.
% They are then considered as higher level components with
% their own failure mode behaviour. These higher level components
% are then collected to form {\fgs} and so on until a hierarchy is built
% representing the entire system.
%
% Any new static failure mode methodology must ensure that it
% represents all component failure modes and it therefore should be bottom-up,
% starting with individual component failure modes.
\paragraph{FMMD process.}
To ensure all component failure modes are modelled and traceable through stages of analysis, the new methodology must be bottom-up.
%
%This seems essential to satisfy criterion 2.
%The proposed methodology is therefore a bottom-up process
A {\em {\fg}}, is defined as a small collection of components
that interact to provide
a function or task within a system.
%
Starting with {\bcs} small {\fgs} are chosen and each component failure mode considered in the
context of the {\fg}.
%
%% GARK
%
The component failures are termed {\em{\fcs}}. %`test~cases'.
For each {\fc}
there will be a corresponding resultant failure, or `symptom', from the perspective of the {\fg}.
%
% MAYBE NEED TO DESCRIBE WHAT A SYMPTOM IS HERE
%
%From the perspective of the {\fg} failures of components will be symptoms.
It is conjectured that many symptoms will be common. That is to say
that component failures will often cause the same symptoms of failure
from the perspective of a {\fg}.
%
%
A common symptom collection stage is now applied. Here common symptoms are collected
from the results of the {\fcs}.
%Because it is possible to model combinations of failures, criterion 6 is satisfied.
%
With a collection of the {\fg} failure symptoms, a new component, a {\em{\dc}} is created.
The failure modes of this new {\dc} are the symptoms of the {\fg} it was derived from.
%This satisfies criterion 4, as we can now treat {\dcs} as pre-analysed
%modules available for re-use.
%
%
By using {\dcs} in higher level functional groups, a hierarchy can be built representing
the failure mode behaviour of a system. Because the hierarchy maintains information
linking the symptoms to component failure modes (via {\fcs}), reasoning connections from base component failures to top level failures can now be made
by tracing cause and effect though the hierarchy of modules~\footnote{This means that an FMMD model can be used to produce traditional FMEA reports where each {\bc} {\fm} is linked to
a system level failure.}.
%The traceability should satisfy criterion 5.
An advantage of performing FMEA in this modular way, is that the
{\fgs} are small in terms of the numbers of components. This means the $O(N^2)$ effect
of the reasoning distance is greatly reduced for the overall project.
This addresses the state explosion problem of XFMEA.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%FFT%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
In the field of digital signal processing there is an algorithm that revolutionised
access to frequency analysis of digital samples called the Fast Fourier Transform (FFT)~\cite{fftoriginal}.
This took the Discrete Fourier Transform (DFT), and applied de-composition to its
mesh of (often repeated) complex number calculations~\cite{fpodsadsp}[Ch.8].
%
By doing this it broke the computing order of complexity down from having a polynomial %n exponential
%order
to logarithmic order~\cite{ctw}[pp.401-3].
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%FFT%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
It also means that modules are re-usable (analogous to software classes).
%
Where there are repeated sections of circuitry (as in for instance common types of interface)
the analysis for that module may be simply re-used.
%
%
A practical example of a hardware FMEA performed both traditionally and using FMMD may be found in~\cite{syssafe2011}, a hybrid
software and hardware hybrid example is analysed in~\cite{syssafe2012}
and examples of `reasoning~distance' efficiency savings can be found in~\cite{clark}[Ch.7].
%
\paragraph{Integrating software into the FMMD model}
%
%With modular FMEA i.e. FMMD %(FMMD)
%the concepts of failure~modes
%of components, {\fgs} and symptoms of failure have been defined. % for a functional group.
%
A programmatic function has similar attributes to an FMMD {\fg}. % with these concepts. %a {\fg} as defined by the FMMD process.
%
An FMMD {\fg} is placed into a hierarchy, likewise
a software function is typically placed into the hierarchy of its call-tree.
%
A software function calls other functions and uses data sources %via hardware interaction
which could be viewed as its `components':
it has outputs, i.e. it can perform actions on data or hardware.
%which will be used by other functions that may call it.
%
It is shown below that a software function can be mapped to an FMMD {\fg}: its failure modes
are the failure modes of the software components %(other functions
it calls %)
and/or the hardware from which it reads values.
Its outputs are the data it changes, or the hardware actions it performs.
%%
%% Talk about how software specification will often say how hardware
%% will react and how to interpret readings---but they do not
%% always cover the failure modes of the hardware being interfaced too.
%
When a
software function has been analysed---using failure conditions of its inputs as a source of failure modes---its symptoms of failure
can be defined (i.e. how functions that call it will see its failure mode behaviour).
%
%
FMMD is applied to software functions by viewing functions in terms of their failure mode behaviour.
%
That is to say, using FMMD, software functions are treated like {\fgs} of electronic components.
%
%
As software already fits into a hierarchy, there one less analysis decision to make when compared
to analysing electronics.
%
For electrical and mechanical systems, although the original system designers
concepts of modularity and sub-systems in design may provide guidance,
applying FMMD means deciding on the members for {\fgs} and the subsequent hierarchy.
%
\section{Example for analysis} % : How can we apply FMEA}
% %
The example chosen is the smallest meaningful embedded system that the author is familiar with.
A standalone temperature controller.
This consists of both hardware and software.
The hardware elements are sensors (temperature sensing Pt100 resistors),
an actuator (an output to switch on and off a heater)
and some indicator LEDs.
The software reads the temperature from the sensor and applies checks
to detect any failures.
The software then applies a PID~\cite{dcods} algorithm to determine the length/modulation of the pulses applied to the heater.
yourdon context diagram here
\section{Closed Loop Control Hardware/Software Hybrid Example}
It is desirable to model a complete standalone system with FMMD,
not only a standalone system, but ideally a hybrid software/hardware system.
%
Temperature control is typically a first order differential problem, and is often
addressed using the Proportional Integral Differential (PID) algorithm~\cite{dcods}[p.66].
%
Traditionally this was performed in analogue electronics
with trimmer potentiometers providing the P, I and D parameters.
%
Since the introduction of digital computers, it has been possible to
implement PID in software. %pro-grammatically.
%
A PID temperature controller is presented
as a complete example of an electronic/hardware hybrid analysed using FMMD. %would mean an
%analysis of a realistic standalone system without being it becoming an un-wieldingly large task.
% % \paragraph{The PID Temperature Control Algorithm.}
% % PID control starts with a setpoint, or desired value for a process
% % (here the temperature). It reads the process value and determines an error value for it.
% % The aim of the PID controller is to minimise this error term, by setting an output value,
% % which is fed back into the process (in this example the amount of power to supply the heater).
% % The error value is integrated and multiplied by an I constant.
% % A differential of the error value is calculated and multiplied by a D constant.
% % The error value itself is multiplied by a P constant, and all three of these are added
% % to obtain the output required.
% % %
% % A mathematical description of PID with frequency domain modelling (La-Place transforms etc)
% % may be found in~\cite{dcods}[Ch.3.3].
%
\subsection{Design Stage: Implementation on a micro-controller.}
When designing a computer program it is often useful to
start with a system overview.
A structured analysis `Yourdon' context diagram~\cite{Yourdon:1989:MSA:62004} is presented below, see figure~\ref{fig:context_diagram_PID}.
%
\begin{figure}[h]+
\centering
\includegraphics[width=400pt]{../../submission_thesis/CH5_Examples/context_diagram_PID.png}
% context_diagram_PID.png: 818x324 pixel, 72dpi, 28.86x11.43 cm, bb=0 0 818 324
\caption{Yourdon Context Diagram for a standalone micro-processor implemented PID Temperature Controller.}
\label{fig:context_diagram_PID}
\end{figure}
%
Using figure~\ref{fig:context_diagram_PID} the system in terms of its data flow is reviewed, starting
with the data sources (the Pt100 temperature sensor inputs) and the data sinks (the heater output and the LED indicators).
%
There are two voltage inputs (see section~\ref{sec:Pt100}) from the Pt100 temperature sensor.
%
For the Pt100 sensor, the voltages it outputs are read and %for
this requires an ADC and MUX.
%
%\fmmdglossADC
%
For the output, a Pulse Width Modulator (PWM) can be used (this is a common module found on micro-controllers
facilitating
variable power output~\cite{aoe}[p.360]).
%
PWM's ADC's and MUX's are commonly built into cheap micro-controllers~\cite{pic18f2523}[Ch.15].
%
%and add more detail, see figure~\ref{fig:context_diagram2_PID}.
\begin{figure}[h]+
\centering
\includegraphics[width=400pt]{../../submission_thesis/CH5_Examples/context_diagram2_PID.png}
% context_diagram_PID.png: 818x324 pixel, 72dpi, 28.86x11.43 cm, bb=0 0 818 324
\caption{Yourdon data flow diagram for PID Temperature Controller identifying initial processing nodes.}
\label{fig:context_diagram2_PID}
\end{figure}
%
\clearpage
%
The Yourdon methodology provides model refinement, by zooming into data transform bubbles, analysing them in more
depth and creating more paths and transform bubbles which further define the data flow and processing. % required.
%
The Yourdon diagram is refined, by adding detail to both the afferent data flow coming through the MUX and ADC on the micro-controller and the efferent
channelled through a PWM module. %again built into the micro-controller,
%
This next stage of model refinement is shown in figure~\ref{fig:context_diagram2_PID}.
%
The controlling software is then further refined, by looking at or zooming into transform bubbles
and adding more detail i.e. following the data streams through the process, additional transform bubbles are created as required.
%
The lines connecting the `transform~bubbles' define the data passed between them.
%
When the data flow analysis is finished, each transform bubble represents a software function.
%
Because the connecting lines define the data passed between transform bubbles,
the inputs and outputs of the associated software functions are also defined.
%
The Yourdon methodology thus allows the refinement and modelling
of a process from a data~flow perspective
defining software functions in its final stage (see figure~\ref{fig:contextsoftware}).
%, and
%this in terms of software functions.
%
In all `bare~metal'\footnote{`Bare~metal' is a term used to indicate a micro-processor
controlled system that does not use a traditional operating system. These are generally
coded in 'C' or assembly language and run immediately from power-up.}
software architectures, a rudimentary operating system is required, often referred to as the `monitor'.
%
PID, because the algorithm depends heavily on integral calculus~\cite{dcods}[Ch.3.3] is time sensitive
and it is necessary to execute it at precise intervals determined by its proportional, integral and differential (PID) coefficients.
%
Most micro-controllers feature several general purpose timers~\cite{pic18f2523}.
%
An internal timer can be used in conjunction with the monitor function
to call the PID algorithm at a regular and precise time interval. % specified interval.
%
\paragraph{Data flow model to programmatic call tree.}
The Yourdon methodology also gives guidance as to which software
functions should be called to control the process, or in `C' terms be the main function.
%
\begin{figure}[h]
\centering
\includegraphics[width=400pt]{../../submission_thesis/CH5_Examples/context_software.png}
% context_software.png: 1023x500 pixel, 72dpi, 36.09x17.64 cm, bb=0 0 1023 500
\caption{Final Yourdon data flow diagram which has defined the software functions for the PID temperature controller}
\label{fig:contextsoftware}
\end{figure}
%
Using figure~\ref{fig:contextsoftware} the transform bubble
to represent the `main' or controlling function in the software must be chosen.
%
This can be thought of as picking one bubble and holding it up.
%
The other bubbles hang underneath
forming the software call tree hierarchy, see figure~\ref{fig:context_calltree}.
%
From examining the diagram, and in common with established embedded programming practise,
this is clearly going to be the monitor function.
%
\begin{figure}[h]+
\centering
\includegraphics[width=300pt]{../../submission_thesis/CH5_Examples/context_calltree.png}
% context_calltree.png: 800x783 pixel, 72dpi, 28.22x27.62 cm, bb=0 0 800 783
\caption{Software: Yourdon data flow diagram converted to programatic call tree.}
\label{fig:context_calltree}
\end{figure}
%
%
\paragraph{Software Algorithm.}
%
The monitor function will orchestrate the control process.
%
Firstly it will examine the timer value, and when appropriate, call the \cf{PID} function.
%
The \cf{PID} function calls \cf{determine\_set\_point\_error} which calls \cf{convert\_ADC\_to\_T}
which in turn calls \cf{Read\_ADC} (the function developed in the earlier example)
which reads from hardware.
%
With the set point error value the \cf{PID} function will return an output control value to its calling
function (i.e. the PID demand which will be returned to the monitor function).
%
%On returning to the monitor function, it will return the PID demand value.
The PID demand value will be applied via the pulse width modulation (PWM) module.
%
A rudimentary closed loop control system incorporating both hardware and software has been defined.
%
By using the Yourdon methodology a programmatic design frame-work i.e. a call tree structure was obtained.
%
All the components, i.e. hardware elements and software functions
that will be used in the temperature controller are now defined.
%
These are listed, and from the bottom-up, FMMD analysis is begun.
%
\clearpage
\subsection{FMMD Analysis of PID temperature Controller}
%
To summarise from the design stage,
the electronic components identified thus far:
\begin{itemize}
\item ADCMUX --- Electronics, analysed in previous example,
\item TIMER --- Internal micro controller timer,
\item HEATER --- Heating element, essentially a resistor,
\item Pt100 --- Pt100 Temperature sensor, as analysed in section~\ref{sec:Pt100},
\item PWM --- Internal micro controller pulse width modulation module,
\item General Purpose I/O (GPIO) --- I/O used to drive LEDS, %. %source LED current
\item LEDs --- Indication LEDs via GPIO,
\item micro-controller --- the medium for running the software.
\end{itemize}
%
\subsection{Temperature Controller Hardware Elements FMMD.}
%
\paragraph{ADCMUX and Read\_ADC.}
We re-use the {\dc} from section~\ref{readADC}.
$$ fm(RADC) = \{ VV\_ERR, HIGH, LOW \} .$$
%
%
\paragraph{TIMER.}
%
The internal timer, from a programmer's perspective is a register, which when read
returns an incremented time value.
%
Essentially its a free running integer counter with an interfacing register.
%
Using two's complement mathematics, by subtracting
the time last read value, we can calculate the interval
between readings (assuming the timer has not wrapped around more than once).
%
A timer can fail by
incrementing its value at an incorrect rate, or can stop incrementing.
%
The failure modes of $TIMER$ are defined thus:
$$ fm(TIMER) = \{ STOPPED, INCORRECT\_INTERVAL \}.$$
%
\paragraph{HEATER.}
A heating element is typically some configuration of resistive wire.
It therefore has the same failure modes as a resistor:
$$fm(HEATER) = \{ OPEN, SHORT \} .$$
%
\paragraph{Pt100 Platinum Temperature Sensor.}
The Pt100 four wire configuration was analysed in section~\ref{sec:Pt100}, the {\dc} is re-used here:
$$ fm(Pt100) = \{ OUT\_OF\_RANGE \} . $$
%
%
\paragraph{PWM.}
%The PWM, in use, is a hardware register written to with an integer value~\cite{pic182523}[Ch.15].
From a programmatic perspective a PWM output is a register to which software writes
an unsigned magnitude value~\cite{pic18f2523}[Ch.15].
%
The PWM hardware module
applies this using a mark space ratio proportional to that value, providing
a means of varying the amount of power supplied.
%
When the PWM action is halted, or fails, the digital output pin associated with it
will typically be held in a high or low state.
%
The PWM has the following failure modes:
$$ fm(PWM) = \{ HIGH, LOW \}.$$
\paragraph{Micro-Controller.}
The Micro controller is a complex piece of highly integrated electronics.
%
At a minimum it would include a micro-processor with PROM and RAM
general I/O and external interrupt lines.
%
Typically there are many other I/O modules incorporated (e.g. TIMERS, UARTS, PWM, ADC, ADCMUX, CAN).
%
In this project the ADCMUX, TIMER, PWM and general purpose computing facilities are used.
%
Consider the general~computing, CLOCK, PROM and RAM failure modes:
$$fm (micro-controller) =\{ PROM\_FAULT, RAM\_FAULT, CPU\_FAULT, ALU\_FAULT, CLOCK\_STOPPED \}.$$
%
\subsection{Temperature Controller Software Elements FMMD}
Identified Software Components:
\begin{itemize}
\item --- \cf{Monitor} (which calls \cf{PID},\cf{output\_control} and \cf{setLEDS}),
\item --- \cf{PID} (which calls \cf{determine\_set\_point\_error} ),
\item --- \cf{determine\_set\_point\_error} (which calls \cf{convert\_ADC\_to\_T}),
\item --- \cf{convert\_ADC\_to\_T} (which calls \cf{read\_ADC}), % which has been analysed as the {\dc} read\_ADC which can be re-used.} % from the last example),
\item --- \cf{read\_ADC} (analysed in the previous section~\ref{sec:readadc}),
\item --- \cf{output\_control} (which sets the PWM hardware according to the PID demand value).
\end{itemize}
%
%
With the call tree structure defined (see figure~\ref{fig:context_calltree}),
a hierarchy compatible with FMMD for analysis has been obtained.
%
However, it is only the top, i.e. the software, part of the hierarchy.
%
FMMD is a bottom-up process, thus it starts with the lowest level, i.e. the electronics.
%
The Yourdon context diagram (see figure~\ref{fig:context_diagram_PID}) is useful here as its data sources and sinks are
by definition the lowest levels in the system.
%
The input, or origin of the afferent data flow can be followed to find system inputs,
and the output, or efferent flow to find the bottom level for outputs/actuators etc.
%
Starting with the afferent flow, the reading of the temperature and its conversion
to a PID calculated heater output demand is examined.
%
\subsubsection{Afferent flow FMMD analysis, Pt100, temperature, set point error, PID output demand.}
%
Staring with the afferent data flow for the temperature readings, the lowest
level in the hierarchy is found, the Pt100 sensor.
%with the software, and consider the hardware elements
%used (if any) by each software function.
%Starting
Beginning at the bottom, a {\fg} is formed with
the function \cf{read\_ADC} and the Pt100.
This gives a {\dc}, %which we call
`Read\_Pt100' (see appendix~\ref{sec:readPt100}).
%
%
%
%The {\dc} Read\_Pt100 is obtained from analysis of a {\fg} comprising of the
%\cf{Read\_ADC} software function and the Pt100 hardware.
%The {\dc} Read\_Pt100 is a failure mode model of the \cf{Read\_ADC} software function and the Pt100
%hardware, this
The {\dc} Read\_Pt100 has the following failure modes:
%
$$ fm (Read\_Pt100) = \{ VOLTAGE\_HIGH, VAL\_ERR, VOLTAGE\_LOW \}. $$
%
%
Moving along the afferent flow, the \cf{convert\_ADC\_to\_T} function is next up the hierarchy.
%
This will call \cf{Read\_ADC} twice, once for the high Pt100 value, again for the lower. % and once for to read a current sense.
%
The resistance of the Pt100 element is then calculated, and with this---using a
polynomial or a lookup table~\cite{eurothermtables}---the temperature determined.
%
%\fmmdglossCONTRACTPROG
%
The pre-conditions for the function are that:
\begin{itemize}
% \item The current calculated is within pre-defined bounds i.e. Pt100\_current,
\item The lower Pt100 value is within an acceptable voltage range i.e. Pt100\_lower\_voltage,
\item The higher Pt100 value is within an acceptable voltage range i.e. Pt100\_higher\_voltage,
\item The lower and higher values agree to within a given tolerance i.e. Pt100\_high\_low\_mismatch.
\end{itemize}
%
Any violation of these pre-conditions is equivalent to a failure mode\footnote{An actual measured temperature outside the
pre-defined range would be detected as an unacceptable voltage range failure.}.
%
The post-condition is that it returns a temperature within a given tolerance to the temperature at the sensor.
%
A failure of this post-condition can be termed `temp\_incorrect'.
%
\clearpage
Applying FMMD to the {\fg} formed by \cf{Read\_Pt100} and the function \cf{convert\_ADC\_to\_T}.
gives the {\dc} {Get\_Temperature}.
%
This analysis is presented in table~\ref{tbl:gettemperature}.
%
%The analysis for the Pt100 circuit is presented in table~\ref{tbl:readPt100}.
%
%
Failure symptoms are collected and the {\dc} created with the following failure modes:
%
%
$$fm(Get\_Temperature) = \{ Pt100\_out\_of\_range, temp\_incorrect \} . $$
%\clearpage
%
%
Following the afferent flow further, the function to determine the control error value is examined.
%
This is simply the target temperature subtracted from that measured by the sensor.
%
A {\fg} is formed with the newly {\dc} Get\_Temperature
and the function \cf{determine\_set\_point\_error}.
%
The pre-condition for \cf{determine\_set\_point\_error} is that the temperature read by it
is accurate, and its post-condition is to return the correct control error value.
%
%All single failure modes from a four wire Pt100 sensor are detectable (see section~\ref{sec:singlePt100FMEA}).
%
%For most practical purposes this would suffice, but for the purpose of example
%a particular double failure scenario, potentially giving an undefined value is
%considered (see section~\ref{sec:Pt100floating}).
%
The post-condition can fail, or the temperature read could be incorrect.
%
This could be detectable (i.e. we detect a failure from the Pt100 $ Pt100\_out\_of\_range $)
or undetectable (i.e. the post condition for this function simply fails
or the failure mode $temp\_incorrect$ occurs).
%and so an incorrect value that is detected, KnownIncorrectErrorValue
%where we can detect the Pt100 value is suspect,
%and IncorrectErrorValue where there is simply
%an incorrect value but this cannot be determined (i.e. its an undetectable failure). % this.
%
This analysis is presented in table~\ref{tbl:geterror}.
%
%
%
%
Failure mode symptoms are collected and a new {\dc} GetError created
where:
$$fm(GetError) = \{ KnownIncorrectErrorValue, IncorrectErrorValue \}.$$
%
Following the afferent path the PID algorithm is next in the software call tree.
%
%Here we assume that the PID constants are fixed (i.e. are not parameters).
%
The $GetError$ {\dc} and the \cf{PID} function form a {\fg}.
%
The pre-condition for the \cf{PID} function is that
it receives the correct error value.
%
The post-condition is that it outputs correct control values.
% RESP FOR TIMEING IS ON CALLING FUNCTION AND IS A SEPARATE ERROR- TGHINK ABOUT JITTER.....
% and controll values..... Jitter might not matter, wrong int times would
% controlling function provdes context of use.
%Those familiar with the PID algorithm may realise that digital signal processing algorithms are sensitive to calling frequency.
All digital signal processing algorithms are sensitive to calling frequency, and thus should be time invariant~\cite{fpodsadsp}[p.58].
Were this function to be called at an incorrect rate, its output
could be erroneous (the differential and integral parameters would effectively have been changed).
%
However this problem is a failure mode for the consideration of the function calling it i.e. the context of use. %(see section~\ref{sec:subjectiveobjective}).
%
That is, the \cf{PID} function is called, but its calling function is responsible for the timing,
or in more general terms,
it is the calling function that sets the context for the \cf{PID} function (i.e. what it is used for).
%If this PID were to be used, say as some form of low pass filter, we could consider jitter
%for instance.
%
%In a control environment with PID, jitter would not be a significant factor.
%
%HARK THE HERALD ANGELS SING... HARK????
%
%
%
The {\dc} PID is created, see table~\ref{tbl:pidfunction}, with the following failure modes:
%
$$ fm(PID) = \{ KnownControlValueErrorV, IncorrectControlErrorV \} .$$
%
%
\begin{figure}[h]
\centering
\includegraphics[width=400pt]{../../submission_thesis/CH5_Examples/euler_afferent_PID.png}
% euler_afferent_PID.png: 1002x342 pixel, 72dpi, 35.35x12.06 cm, bb=0 0 1002 342
\caption{Euler diagram representing the hierarchy of FMMD analysis applied to the afferent branch of call tree for the PID temperature controller example.}
\label{fig:euler_afferent_PID}
\end{figure}
%
%
%
The software call tree for the afferent flow has now been modelled using FMMD;
this is represented as an Euler diagram in figure~\ref{fig:euler_afferent_PID}.
Two call tree branches remain. The LED indication branch and the
PWM/heater output.
%
\subsubsection{Efferent flow, PID demand value to PWM output}
%
The monitor function calls the \cf{output\_control} function with the PID demand.
%
The \cf{output\_control} function then sets the PWM hardware register, which causes the mark space output of the PWM module to
apply the demanded power.
%
A {\fg} with the Heating element, a PWM module and the \cf{output\_control} function is formed to model this branch
of the efferent flow.
%
This {\fg} is a hardware/software hybrid.
%
FMMD analysis is applied to this {\fg} in table~\ref{tbl:heateroutput}.
%
For the \cf{output\_control} function, there is a pre-condition that the PWM module is
configured and working, and has the correct clock frequency.
%
A second pre-condition is that the heating element is connected and working.
%
The post-condition is that it sets the correct value into the PWM register
to implement the power output demand.
%
%
%
A {\dc} is created called HeaterOutput, see table~\ref{tbl:heateroutput},
with the following failure modes:
$$fm(HeaterOutput) = \{ HeaterOnFull, HeaterOff, HeaterOutputIncorrect \} .$$
%
As an aside: the $HeaterOnFull$ failure should raise alarm bells for designers and
upon its discovery, measures may be recommended to inhibit this (such as perhaps
adding a safety relay to cut the power to the heater).
%
%
\begin{figure}[h]
\centering
\includegraphics[width=300pt]{../../submission_thesis/CH5_Examples/euler_heater_output.png}
% euler_heater_output.png: 392x141 pixel, 72dpi, 13.83x4.97 cm, bb=0 0 392 141
\caption{Euler diagram showing HeaterOutput with its two hardware components, PWM and HEATER, and its software component \cf{output\_control}.}
\label{fig:eulerheateroutput}
\end{figure}
%
%
%
%
\subsubsection{Efferent flow: LED status LEDs}
%
The status LEDS will be controlled by general purpose (GPIO) I/O pins.
%
Three LEDS could be used, one flashing with a human readable mark
space ratio representing the heater output, one flashing at a regular interval to
indicate the processor is alive and another flashing at an interval related to the temperature,
(to indicate if the temperature readings are within expected ranges).
%
Each LED should flash in normal operation, and any LED being permanently on or off
would indicate to the operator that an error had occurred.
%
The pre-condition for this function is that the GPIO
is connected to working LEDS.
%
The post-condition is that the function \cf{setLEDS} will supply correct indication by flashing the LEDs.
%
A {\fg} is formed from the GPIO, the LEDs and the software function \cf{setLEDs}.
%
FMMD analysis is applied to this {\fg} in table~\ref{tbl:ledoutput}.
%
%
%
%
%
\begin{figure}[h]
\centering
\includegraphics[width=300pt]{../../submission_thesis/CH5_Examples/euler_led_output.png}
% euler_heater_output.png: 392x141 pixel, 72dpi, 13.83x4.97 cm, bb=0 0 392 141
\caption{Euler diagram showing LEDOutput with its three LEDs and GPIO hardware elements,
and its software component setLEDS.}
\label{fig:eulerheateroutput}
\end{figure}
%
%
The {\dc} for the setLED function, GPIO and LEDs has the following failure modes:
$$ fm(LEDoutput) = \{FailureIndicated, IndicationError \} $$
%
%
\subsubsection{Final Analysis Stage: PID Temperature Controller}
%
The possibility of each software function failing its post-condition without a direct
underlying cause from one of its components has been included in each analysis stage
involving software.
%
This is because software introduces the possibility of
anything going wrong!
%
The common causes for software failing are:
\begin{itemize}
\item Value/RAM corruption typically from interrupt contention problems~\cite{concurrency_c_tool} or accidental over writing~\cite{swseatbelt},
but can be from external sources such as radiation changing bits/values at runtime~\cite{5963919, 5488118};
\item Address bus errors leading to program errors (program sequence);
\item ROM memory failures;
\item Unintended behaviour of software.
\item Electro Magnetic Compatibility (EMC) interference.
\end{itemize}
Because the software is running on a medium, that of the processor or micro-controller,
the FMMD analysis at the final or highest level (see table~\ref{tbl:pid}), must include all possible failure modes of this medium i.e.
$$fm (micro-controller) =\{ PROM\_FAULT, RAM\_FAULT, CPU\_FAULT, ALU\_FAULT, CLOCK\_STOPPED \}.$$
The final FMMD stage forms a {\fg} with the {\dcs}
determined previously:
%
\begin{itemize}
\item the micro-controller,
\item PID,
\item HeaterOutput,
\item LEDoutput,
\item the function \cf{monitor}.
\end{itemize}
%
The post-condition for the monitor function is that it implements the PID control task correctly.
%\fmmdglossCONTRACTPROG
A {\dc} for the standalone temperature controller is now created, and given the name TempController.
It will have the following failure modes:
%
\begin{eqnarray*}
fm ( TempController ) = \{ ControlFailureIndicated, \\ ControlFailure, \\ KnownIndicationError, \\ UnknownIndicationError \}.
\end{eqnarray*}
%
%
The failure mode analysis of the complete PID controller is represented
as an Euler diagram in figure~\ref{fig:euler_temp_controller}.
%
%
\begin{figure}[h]
\centering
\includegraphics[width=400pt]{../../submission_thesis/CH5_Examples/euler_temp_controller.png}
% euler_temp_controller.png: 714x251 pixel, 72dpi, 25.19x8.85 cm, bb=0 0 714 251
\caption{Euler diagram of the temperature controller final analysis stage, showing the hybrid software/hardware {\dcs} and the function at the head of the call tree \cf{monitor}.}
\label{fig:euler_temp_controller}
\end{figure}
%
\subsection{Conclusion: Standalone system, PID Temperature Controller}
%
The PID temperature control example above, shows that complete hybrid software/electronic systems can be
modelled using FMMD.
%
This analysis has revealed system level failure modes that are un-handled and some that are undetectable.
The FMMD model can be traversed from undesirable top level failures to the {\bc} {\fms} that are the causes.
%\fmmdglossOBS
%
This means that by using FMMD, the sub-systems which require
re-design to eliminate or reduce the likelihood of undetectable failure modes can be identified.
%
The demands of EN61508~\cite{en61508} for minimum safe failure fraction thresholds~\cite{scsh}[p.52] associated with
SIL levels, make this a desirable feature of any FMEA based methodology.
%
For the failure modes caused
by electronics, reliability statistics can be applied, and the possibilities of using higher rated
components instead of potentially expensive re-design can be simulated/modelled.
%
For software errors, it may be necessary to provide extra functions to provide self checking.
%
EN61508 high reliability software measures such as
duplication of functions with checking functions arbitrating them (diverse programming~\cite{en61508}[C.3.5]) could be applied.
%
For instance, measures may included to validate the processor clocking with an external watchdog and a simple
communications protocol. For PROM and RAM faults measures such as run-time checksums
and ram complement checking can be applied.
%
%Using FMMD in conjunction with extra safety measures it can be ensured that no single hardware failure could lead to a
%system failure, something difficult to prove with current FMEA techniques.
\subsection{Hardware: Sensors, actuators and indication}
\subsection{Simple Software Example}
\subsection{Software FMEA - The software/hardware interface}
\section{Conclusion}
%
% This paper has picked a very simple example %(the industry standard {\ft}
% %input circuit and software)
% to demonstrate
% SFMEA and HFMEA methodologies used to describe a failure mode model.
% %Even a modest system would be far too large to analyse in conference paper
% %and this
% %
% %The {\dc} representing the {\ft} reader
% %shows that by taking a
% %modular approach for FMEA, i.e. FMMD, we can integrate
% Our model is described by four FMEA reports; and these % we can model the failure mode behaviour from
% model the system from several failure mode perspectives.
% %
% With traditional FMEA methods the reasoning~distance is large, because
% it stretches from the component failure mode to the top---or---system level failure.
% %
% With these four analysis reports
% we do not have stages along the `reasoning~path' linking the failure modes from the
% electronics to those in the software.
% %Software is often written `defensively' but t
% %Each {\fg} to {\dc} transition represents a
% %reasoning stage.
% %
% %
% %For this reason applying traditional FMEA to software stretches
% %the reasoning distance even further.
% %
% In fact many these reasoning paths overlap---or even by-pass one another---
% it is very difficult to gauge cause and effect.
% For instance, hardware failures are not analysed in the context of how they will
% be handled (or missed) by the software.
% %
% System outputs commanded from software may not take into account particular
% hardware limitations etc.
%
% The interface FMEA does serve to provide a useful
% check-list to ensure data and synchronisation conventions used by the hardware
% and software are not mismatched. However, the fact it is perceived as required
% highlights the the miss-matches possible between the two types of analysis
% which could run deeper than the mere interface level.
%
%
% However, while these techniques ensure that the software and hardware is
% viewed and analysed from several perspectives, it cannot be termed a homogeneous
% failure mode model.
% % For instance
% % were the ADC to have a small value error, say adding
% % a small percentage onto the value, we would be unable to
% % detect this under the analysis conditions for this model, or
% % be able to pinpoint it.
% %
%
% Need wishlist ticks and solved problems here.
{
\footnotesize
\bibliographystyle{plain}
\bibliography{../../vmgbibliography,../../mybib}
}
\today
%\today
\end{document}
Reference in New Issue View Git Blame Copy Permalink