R help - Jun 2017 - Please help(urgent) - How to simulate transactional data for reliability/survival analysis

Hi friends, 
I haven't done such a simulation before and any help would be greatly
appreciated. I need your guidance.

I need to simulate end to end data for Reliability/survival analysis of a Pump
,with correlation in place, that is at 'Transactional level' or at the
granularity of time-minutes, where each observation is a reading captured via
Pump's sensors each minute.
Once transactional data is prepared I Then need to summarise above data for
reliability/ survival analysis.

To begin with below is the transactional data format that i want prepare: 
Pump-id| Timestamp | temp | vibration | suction pressure| discharge pressure |
Flow

Above transactional data has to be prepared with below failure modes
Defects :
(1)    Cavitation ? very high in frequency but low impact
(2)    Bearing Damage ? very low in frequency but high impact
(3)    Worn Shaft ? medium frequency but medium impact

I have used survsim package but that's not what I need here. 
Please help and guide. 

Regards,
Sandeep

I think you need to find a local consultant. Someone here might have a
suggestion or two where to look (as I do below), but this list only
provides help on R programming code, not statistical issues (see
programming guide below for details).

You might wish to have a look at the CRAN survival analysis task view
to see if any packages might address your needs (but warning: it's
mostly about medical applications):

https://cran.r-project.org/web/views/Survival.html

Cheers,
Bert



Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )


On Tue, Jun 27, 2017 at 7:58 AM, sandeep Rana <sandykido at gmail.com>
wrote:
> Hi friends,
> I haven't done such a simulation before and any help would be greatly
appreciated. I need your guidance.
>
> I need to simulate end to end data for Reliability/survival analysis of a
Pump ,with correlation in place, that is at 'Transactional level' or at
the granularity of time-minutes, where each observation is a reading captured
via Pump's sensors each minute.
> Once transactional data is prepared I Then need to summarise above data for
reliability/ survival analysis.
>
> To begin with below is the transactional data format that i want prepare:
> Pump-id| Timestamp | temp | vibration | suction pressure| discharge
pressure | Flow
>
> Above transactional data has to be prepared with below failure modes
> Defects :
> (1)    Cavitation ? very high in frequency but low impact
> (2)    Bearing Damage ? very low in frequency but high impact
> (3)    Worn Shaft ? medium frequency but medium impact
>
> I have used survsim package but that's not what I need here.
> Please help and guide.
>
> Regards,
> Sandeep
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

I apologise as I had mistakenly posted this message via non- member mail. So
I'm reposting it with member id. I need help in this case.
> Hi friends, 
> I haven't done such a simulation before and any help would be greatly
appreciated. I need your guidance.
> 
> I need to simulate end to end data for Reliability/survival analysis of a
Pump ,with correlation in place, that is at 'Transactional level' or at
the granularity of time-minutes, where each observation is a reading captured
via Pump's sensors each minute.
> Once transactional data is prepared I Then need to summarise above data for
reliability/ survival analysis.
> 
> To begin with below is the transactional data format that i want prepare: 
> Pump-id| Timestamp | temp | vibration | suction pressure| discharge
pressure | Flow
> 
> Above transactional data has to be prepared with below failure modes
> Defects :
> (1)    Cavitation ? very high in frequency but low impact
> (2)    Bearing Damage ? very low in frequency but high impact
> (3)    Worn Shaft ? medium frequency but medium impact
> 
> I have used survsim package but that's not what I need here. 
> Please help and guide. 
> 
> Regards,
> Sandeep 
>

There is no member id. Anyone can sign up to post on this list. I
already responded to your prior post.

-- Bert


Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )


On Tue, Jun 27, 2017 at 8:20 PM, Sunny Singha
<sunnysingha.analytics at gmail.com> wrote:
> I apologise as I had mistakenly posted this message via non- member mail.
So I'm reposting it with member id. I need help in this case.
>
>> Hi friends,
>> I haven't done such a simulation before and any help would be
greatly appreciated. I need your guidance.
>>
>> I need to simulate end to end data for Reliability/survival analysis of
a Pump ,with correlation in place, that is at 'Transactional level' or
at the granularity of time-minutes, where each observation is a reading captured
via Pump's sensors each minute.
>> Once transactional data is prepared I Then need to summarise above data
for reliability/ survival analysis.
>>
>> To begin with below is the transactional data format that i want
prepare:
>> Pump-id| Timestamp | temp | vibration | suction pressure| discharge
pressure | Flow
>>
>> Above transactional data has to be prepared with below failure modes
>> Defects :
>> (1)    Cavitation ? very high in frequency but low impact
>> (2)    Bearing Damage ? very low in frequency but high impact
>> (3)    Worn Shaft ? medium frequency but medium impact
>>
>> I have used survsim package but that's not what I need here.
>> Please help and guide.
>>
>> Regards,
>> Sandeep
>>
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

In principle what you need to do is the following:

 - break down the time you wish to simulate into intervals.
 - for each interval, and each failure mode, determine the probability of an
event.
   Determining the probability is the fun part, where you make your domain
   knowledge explicit and include all the factors into your model: cumulative
load,
   failure history, pressure, temperature, phase of the moon ...
 - once you have a probability of failure, use the runif() function to give you
   a uniformly distributed random number in [0, 1]. If the number is smaller
than
   your failure probability, accept the failure event, and record it.
 - Repeat many times.

Hope this helps.
B.



> On Jun 27, 2017, at 10:58 AM, sandeep Rana <sandykido at gmail.com>
wrote:
> 
> Hi friends, 
> I haven't done such a simulation before and any help would be greatly
appreciated. I need your guidance.
> 
> I need to simulate end to end data for Reliability/survival analysis of a
Pump ,with correlation in place, that is at 'Transactional level' or at
the granularity of time-minutes, where each observation is a reading captured
via Pump's sensors each minute.
> Once transactional data is prepared I Then need to summarise above data for
reliability/ survival analysis.
> 
> To begin with below is the transactional data format that i want prepare: 
> Pump-id| Timestamp | temp | vibration | suction pressure| discharge
pressure | Flow
> 
> Above transactional data has to be prepared with below failure modes
> Defects :
> (1)    Cavitation ? very high in frequency but low impact
> (2)    Bearing Damage ? very low in frequency but high impact
> (3)    Worn Shaft ? medium frequency but medium impact
> 
> I have used survsim package but that's not what I need here. 
> Please help and guide. 
> 
> Regards,
> Sandeep 
> 
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

Thanks Boris and Bret,
I was successful in simulating granular/transactional data.
Now I need some guidance to transform the same data in format acceptable
for survival analysis i.e below format:

pump_id | event_episode_no. | event(0/1) | start | stop | time_to_dropout

The challenge I'm experience is to generate the 'start' and
'stop' in units
of minutes/days from single column of 'Timestamp' which is
the column from transactional/granular data based on condition tagged in
separate column, 'event 0/1, (i.e event ).

Please guide how to do such transformation in 'R'.

Regards,
Sandeep



On Wed, Jun 28, 2017 at 2:51 PM, Boris Steipe <boris.steipe at
utoronto.ca>
wrote:
> In principle what you need to do is the following:
>
>  - break down the time you wish to simulate into intervals.
>  - for each interval, and each failure mode, determine the probability of
> an event.
>    Determining the probability is the fun part, where you make your domain
>    knowledge explicit and include all the factors into your model:
> cumulative load,
>    failure history, pressure, temperature, phase of the moon ...
>  - once you have a probability of failure, use the runif() function to
> give you
>    a uniformly distributed random number in [0, 1]. If the number is
> smaller than
>    your failure probability, accept the failure event, and record it.
>  - Repeat many times.
>
> Hope this helps.
> B.
>
>
>
>
> > On Jun 27, 2017, at 10:58 AM, sandeep Rana <sandykido at
gmail.com> wrote:
> >
> > Hi friends,
> > I haven't done such a simulation before and any help would be
greatly
> appreciated. I need your guidance.
> >
> > I need to simulate end to end data for Reliability/survival analysis
of
> a Pump ,with correlation in place, that is at 'Transactional level'
or at
> the granularity of time-minutes, where each observation is a reading
> captured via Pump's sensors each minute.
> > Once transactional data is prepared I Then need to summarise above
data
> for reliability/ survival analysis.
> >
> > To begin with below is the transactional data format that i want
prepare:
> > Pump-id| Timestamp | temp | vibration | suction pressure| discharge
> pressure | Flow
> >
> > Above transactional data has to be prepared with below failure modes
> > Defects :
> > (1)    Cavitation ? very high in frequency but low impact
> > (2)    Bearing Damage ? very low in frequency but high impact
> > (3)    Worn Shaft ? medium frequency but medium impact
> >
> > I have used survsim package but that's not what I need here.
> > Please help and guide.
> >
> > Regards,
> > Sandeep
> >
> > ______________________________________________
> > R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide http://www.R-project.org/
> posting-guide.html
> > and provide commented, minimal, self-contained, reproducible code.
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/
> posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]