This commit is contained in:
Robin Clark 2010-12-02 13:50:14 +00:00
parent b6ae7ffeff
commit ea108b30a6

View File

@ -112,8 +112,8 @@ two different functional groups.
For the sake of example, let our temperature environment
for the SYSTEM be ${{0}\oc}$ to ${{125}\oc}$, but let the component
type `K' have a de-graded performance
\footnote{ A real world example of
degraded performace with temperature is the isolating opto coupler.
\footnote{A real world example of
degraded performance with temperature is the isolating opto coupler.
These can typically only cope with lower baud rate ranges
at high temperatures \cite{tlp181}.}
failure mode between
@ -149,7 +149,7 @@ The UML model shows the relationships between data types (or classes) that
are used in the FMMD process.
The purpose of failure mode analysis, is to tie SYSTEM level failures
to their possible causes in the base components.
By doing this accurate statistics can be obtained for SYSTEM level
By doing this, accurate statistics can be obtained for SYSTEM level
failures, and an insight into how we can make the system safer
can be determined.
In order to do this, we need to be able to trace the component
@ -157,7 +157,7 @@ failure modes from the functional groups, to the symptoms
they cause, and to the failure modes in the {\dcs}.
We can use graph theory to represent this.
As it would make no sense for a derived component to
derive failure modes form itsself, we can apply an acyclic constraint
derive failure modes from itself, we can apply an acyclic constraint
to the graph. This means the graph must be a Directed Acylic
Graph (DAG).
@ -174,7 +174,7 @@ Graph (DAG).
Consider the SYSTEM environment with its temperature range of ${{0}\oc}$ to ${{125}\oc}$.
We must check this against all components used.
For our example, we component `K' which has an extra
For our example, component `K' which has an extra
failure mode for degraded performance `d'. Thus applying the function $fm$
to component type `K' under these temperature range conditions
gives the following failure modes, $fm(K) =\{ K^0_a, K^0_b, K^0_d \}$.
@ -691,7 +691,7 @@ This is shown in the DAG in figure \ref{fig:dag3}.
\node[annot,right of=s](dcl) {Derived Component};
\end{tikzpicture}
% End of code
\caption{DAG representing failure modes and symptoms $FG^0_1 \rightarrow C^1_1$ and $FG^0_2 \rightarrow C^1_2$}
\caption{DAG representing failure modes and symptoms $FG^0_1 \rightarrow C^1_1$, $FG^0_2 \rightarrow C^1_2$ and $FG^0_3 \rightarrow C^1_3$}
\label{fig:dag3}
\end{figure}
@ -734,6 +734,7 @@ TO RACE BACK DOWN THE DAG
\def\layersep{2.0cm}
\begin{figure}
@ -863,13 +864,55 @@ TO RACE BACK DOWN THE DAG
\path (s8) edge (DC-3);
\node[failure, right of=DC-1] (as1) {$a_{s1}$};
\node[failure, below of=as1] (as2) {$a_{s2}$}; % will this overwrite ?
\node[failure] (as1) at (\layersep*4,-2) {$a_{s1}$};
\node[failure] (bs2) at (\layersep*4,-3) {$b_{s2}$};
\path (DC-1) edge (as1);
\path (DC-1) edge (as2);
\path (DC-1) edge (bs2);
\node[failure, above of=DC-2] (as3) {$a_{s3}$}; % will this overwrite ?
\node[failure] (as3) at (\layersep*4,-5) {$a_{s3}$};
\node[failure] (bs4) at (\layersep*4,-6) {$b_{s3}$};
\node[failure] (cs5) at (\layersep*4,-7) {$c_{s3}$};
\path (DC-2) edge (as3);
\path (DC-2) edge (bs4);
\path (DC-2) edge (cs5);
\node[failure] (as6) at (\layersep*4,-12) {$a_{s6}$};
\node[failure] (bs7) at (\layersep*4,-13) {$b_{s7}$};
\node[failure] (cs8) at (\layersep*4,-14) {$c_{s8}$};
\path (DC-3) edge (as6);
\path (DC-3) edge (bs7);
\path (DC-3) edge (cs8);
\node[symptom] (s9) at (\layersep*5,-5) {s9};
\node[symptom] (s10) at (\layersep*5,-10) {s10};
\node[symptom] (s11) at (\layersep*5,-13) {s11};
\path (bs2) edge (s9);
\path (as1) edge (s10);
\path (as3) edge (s10);
\path (bs4) edge (s10);
\path (as6) edge (s10);
\path (bs7) edge (s10);
% Single component failures causing same error in tree
\path (cs5) edge (s11);
\path (cs8) edge (s11);
\node[component,right of=s10] (DC2-1) {$C^2_1$};
\path (s9) edge (DC2-1);
\path (s10) edge (DC2-1);
\path (s11) edge (DC2-1);
\node[failure] (as9) at (\layersep*7,-9) {$a_{s9}$};
\node[failure] (as10) at (\layersep*7,-10) {$b_{s12}$};
\node[failure] (as11) at (\layersep*7,-11) {$c_{s11}$};
\path (DC2-1) edge (as9);
\path (DC2-1) edge (as10);
\path (DC2-1) edge (as11);
% Connect every node in the hidden layer with the output layer
%\foreach \source in {1,...,5}
@ -879,10 +922,15 @@ TO RACE BACK DOWN THE DAG
\node[annot,above of=C-1a, node distance=1cm] (hl) {Failure modes};
\node[annot,left of=hl] {Base Components};
\node[annot,right of=hl](s) {Symptoms};
\node[annot,right of=s](dcl) {Derived Component};
\node[annot,right of=s](dcl) {Derived Components};
\node[annot,right of=dcl](dcf) {Derived Component Failure Modes};
\node[annot,right of=dcf](S2s) {Symptoms};
\node[annot,right of=S2s](DC2) {Derived Components};
\node[annot,right of=DC2](dc2f) {Derived Component Failure Modes};
\end{tikzpicture}
% End of code
\caption{DAG representing failure modes and symptoms $FG^0_1 \rightarrow C^1_1$ and $FG^0_2 \rightarrow C^1_2$}
\caption{DAG representing failure modes and symptoms $FG^1_1 \rightarrow C^2_1$}
\label{fig:dag3}
\end{figure}
@ -893,9 +941,26 @@ Show how the hierarchy can be represented as a DAG
draw a dag
\subsection{Inhibit Conditions represented in the DAG}
Inhibit node type. Octagon (to follow example from FTA).
a -> OCT
inhibitcond--
\subsection{Failure Mode Conjuction Conditions represented in the DAG}
White filled node with an \& in it.
\subsection{Traversing the datamodel}
Show how we can find multiple causes for a SYSTEM level error
Show how we can find multiple causes for a SYSTEM level error.
Constrast this to the bottom-up approaches of FMEA, FMECA and FMEDA where
without necessarily knowing complex interactions between
functionally adjacent components, we must take each component failure
mode and tie to to a SYSTEM level failure.
\subsubsection{Common mode failure detection}