RE: Rolling average
We had done a similar thing, but in the angle domain (if you will) and in our PI system. The hard part, in my opinion, was the poor graphical capabilities of the DM to produce a wind direction graphic. Long story that involved TAC and alot of Oh, so that's how our product works Kirk Kirk Carver ExxonMobil Chemical Beaumont Polyethylene Plant PO Box 2295 Beaumont, Texas 77704 Phone: 409-860-1314 [EMAIL PROTECTED] - Johnson,Alex [EMAIL PROTECTED] on 09/11/2000 08:57:27 AM Please respond to Foxboro DCS Mail List [EMAIL PROTECTED] To: Foxboro DCS Mail List [EMAIL PROTECTED] cc:(bcc: Kirk D Carver/Beaumont/Mobil-Notes) Subject: RE: Rolling average Neat and sweet. Regards, Alex Johnson The Foxboro Company 10707 Haddington Houston, TX 77043 713.722.2859 (v) 713.722.2700 (sb) 713.932.0222 (f) [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] -Original Message- From: Schouten, Frits JF [SMTP:[EMAIL PROTECTED]] Sent: Sunday, September 10, 2000 3:53 PM To: 'Foxboro DCS Mail List' Subject: RE: Rolling average Talking about rolling average, I use a AIN block with filter option. That just works a gem. Anyone ever thought about 0-360deg wind vane damping? See the problem around North? 1,360,1,360. would average 180 wouldn't it? Here is what I did. In a CALC block split out the SIN and COS and output it to say RO02/03 Those two outputs each go to an AIN block with filter option ON and filtertime say 1. The output of those AIN blocks is then fed back in the CALC block to be ASINed and ACOSed and with a bit of smarts put back together to form 0-360deg. There you have it. A very well stabilised wind vane signal which can now be used to drive an directional indicator (FoxView) without driving the operator mad. Cheers, Frits Schouten. BHP-NZSteel. EOM NOTICE - This message contains information intended only for the use of the addressee named above. It may also be confidential and/or privileged. If you are not the intended recipient of this message you are hereby notified that you must not disseminate, copy or take any action in reliance on it. If you have received this message in error please notify [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
RE: Rolling average
Talking about rolling average, I use a AIN block with filter option. That just works a gem. Anyone ever thought about 0-360deg wind vane damping? See the problem around North? 1,360,1,360. would average 180 wouldn't it? Here is what I did. In a CALC block split out the SIN and COS and output it to say RO02/03 Those two outputs each go to an AIN block with filter option ON and filtertime say 1. The output of those AIN blocks is then fed back in the CALC block to be ASINed and ACOSed and with a bit of smarts put back together to form 0-360deg. There you have it. A very well stabilised wind vane signal which can now be used to drive an directional indicator (FoxView) without driving the operator mad. Cheers, Frits Schouten. BHP-NZSteel. EOM NOTICE - This message contains information intended only for the use of the addressee named above. It may also be confidential and/or privileged. If you are not the intended recipient of this message you are hereby notified that you must not disseminate, copy or take any action in reliance on it. If you have received this message in error please notify [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
RE: Rolling average
Alex suggests averaging 8 buckets from the DEADTIME block and this is very practical. Some other suggestions are: First DTOPT is set to 0 (DEADTIME block) so that the input is averaged over the duration of bucket transfers (the so called rolling average feature). If the DEADTIME block period is at least as fast as the variable to be averaged, then every single available sample will be considered in the rolling average (this precision may not be required). If DT is set to 60 minutes then the 10 buckets will transfer every 6 minutes. If the average must be updated every 1 minute then 6 DEADTIME blocks could be ganged together, but now several CALC, SELECTOR, or SEQUENCE blocks (60 sec period) would be needed to average 60 buckets. As required precision is not specified, heck maybe a simple lag block would do. Rick Rys --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
Re: Rolling average
Steve, Try the following in a CALC block: RI01 = valve position RO01 = approx. one hour average valve position Block Period = whatever M01 = filter time = block period / 1 Time Constant If time average is 1 hour (approx. 4 time constants) then 1 time constant = 15 minutes or 900 seconds. Assume block period = 10 seconds, then M01 = 10 / 900 = .01 CALC BLOCK STEPS: (FIRST ORDER LAG FILTER) IN RI01 SUB RO01 MUL M01 ADD RO01 OUT RO01 END Ray Chouinard Process Control Mead, Escanaba, MI Steve Rigby [EMAIL PROTECTED]@lists.TheCassandraProject.org on 09/06/2000 09:01:10 AM Please respond to Foxboro DCS Mail List [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] To: 'Foxboro DCS Mail List' (E-mail) [EMAIL PROTECTED] cc: Subject: Rolling average Hi all, What I would like is a way to calculate the average position of a valve over the last hour, updated say every minute. and then available as an input for a block. Any ideas? Thanks in advance. Steve Rigby Air Liquide Industrie BV Merseyweg 10 3197 KG Botlek-Rotterdam The Netherlands Tel: +31(0)181-271131 Fax:+31(0)181-271100 Email: [EMAIL PROTECTED] [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
RE: Rolling average
So, Steve. What did you decide to do. Regards, Alex Johnson The Foxboro Company 10707 Haddington Houston, TX 77043 713.722.2859 (v) 713.722.2700 (sb) 713.932.0222 (f) [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] -Original Message- From: Steve Rigby [SMTP:[EMAIL PROTECTED]] Sent: Wednesday, September 06, 2000 8:01 AM To: 'Foxboro DCS Mail List' (E-mail) Subject:Rolling average Hi all, What I would like is a way to calculate the average position of a valve over the last hour, updated say every minute. and then available as an input for a block. Any ideas? Thanks in advance. Steve Rigby Air Liquide Industrie BV Merseyweg 10 3197 KG Botlek-Rotterdam The Netherlands Tel: +31(0)181-271131 Fax:+31(0)181-271100 Email: [EMAIL PROTECTED] [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
RE: Rolling average
Alex, Due to other work I've only had time to build 1 of them but I intend to try them all! Yours was the easiest and quickest so I chose that. Ever noticed how a valve doesn't open when you want it to? It's running using a temporary RAMP block for input to the DTIME and first indications are that it works great. The saw tooth has become a gentle sine wave. My pet Process Engineer will be very happy. Thanks to everybody for the tips. Regards Steve. Steve Rigby Air Liquide Industrie BV Merseyweg 10 3197 KG Botlek-Rotterdam The Netherlands Tel: +31(0)181-271131 Fax:+31(0)181-271100 Email: [EMAIL PROTECTED] [EMAIL PROTECTED] -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Johnson,Alex Sent: Thursday, September 07, 2000 3:25 PM To: Foxboro DCS Mail List Subject:RE: Rolling average So, Steve. What did you decide to do. Regards, Alex Johnson The Foxboro Company 10707 Haddington Houston, TX 77043 713.722.2859 (v) 713.722.2700 (sb) 713.932.0222 (f) [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] -Original Message- From: Steve Rigby [SMTP:[EMAIL PROTECTED]] Sent: Wednesday, September 06, 2000 8:01 AM To: 'Foxboro DCS Mail List' (E-mail) Subject:Rolling average Hi all, What I would like is a way to calculate the average position of a valve over the last hour, updated say every minute. and then available as an input for a block. Any ideas? Thanks in advance. Steve Rigby Air Liquide Industrie BV Merseyweg 10 3197 KG Botlek-Rotterdam The Netherlands Tel: +31(0)181-271131 Fax:+31(0)181-271100 Email: [EMAIL PROTECTED] [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
RE: Rolling average
Here's a way to do it using a TIMER and an IND block. You could do without
the timer and just use the WAIT for the period but then you might run into
problems if your CP goes into overrun.
This is set for a flow but will work on anything..
INDEPENDENT_SEQUENCE{***
** **
** INDEPENDENT SEQUENCE **
** CONTROL BLOCK**
** **
*}
CONSTANTS {*
* specify any Constants *
* in the following format: *
**
* 12_char_name = value ; *
*}
VARIABLES {*
* specify any Block Local Variables *
* in the following format: *
**
* 12_char_name [ , 12_char_name ] : type;*
**
*where type is one of B, I, R, S, S12, S6*
*}
AVG:R[360];
ICT:I;
USER_LABELS{
* specify any user labeled parameters*
* in one of the following formats: *
**
* 10_char_name : BI00nn; n = 01 -- 24 *
* 10_char_name : BO00nn; n = 01 -- 16 *
* 10_char_name : BA000n; n = 1 -- 4 *
* 10_char_name : RI00nn; n = 01 -- 15 *
* 10_char_name : RO00nn; n = 01 -- 15 *
* 10_char_name : RA000n; n = 1 -- 2 *
* 10_char_name : II000n; n = 1 -- 8 *
* 10_char_name : IO000n; n = 1 -- 5 *
* 10_char_name : IA0001; *
* 10_char_name : SN00nn; n = 01 -- 10 *
*}
FI3050: RI0001; {MEASURED FLOW}
FA3050: RO0001; {AVERAGE FLOW}
CNTIDX: IO0001; {AVERAGER INDEX}
SUM:RO0002; {TEMPORARY SUM}
TIM:RI0002; {10 SECOND TIMER INPUT}
REMOVE: RO0003; {VALUE TO REMOVE FROM SUM}
BTU:RI0003; {ENTERED VALUE FOR BTU}
BTU_VAL: RO0015; {CALCULATED BTU VALUE FOR CURRENT FLOW}
{*
* Specify any Subroutines*
**
**
* SUBROUTINE name ( formal arguments ) ; *
* VARIABLES subr. local variables ;*
* STATEMENTS*
* ; *
* ENDSUBROUTINE *
*}
{*
* Specify any*
*Standard Block Exception Handlers *
**
**
* BLOCK_EXCEPTION exc_name [ DISABLE ] *
* STATEMENTS*
* ; *
* ENDEXCEPTION *
*}
STATEMENTS{*
* Specify the statements here *
*}
CNTIDX := 1;
AVG := SET_ARRAY(FI3050);
SUM := FI3050 * 360.0;
START_TIMER(::10SECONDS.TIMR1,0.0);
START
REMOVE := AVG[CNTIDX];
AVG[CNTIDX] := FI3050;
SUM := SUM - REMOVE;{REMOVE THE OLD VALUE}
SUM := SUM + AVG[CNTIDX]; {ADD IN THE NEW VALUE}
CNTIDX := CNTIDX + 1;
IF CNTIDX 360 THEN
CNTIDX := 1;
ENDIF;
FA3050 := SUM / 360.0;
{CALCULATE BTU VALUE}
BTU_VAL := FA3050 * BTU;
::10SECONDS.TIMR1V := 0.0;
WAIT 1;
WAIT UNTIL TIM = 10.0;
GOTO START;
ENDSEQUENCE
-Original Message-
From: Steve Rigby [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, September 06, 2000 8:01 AM
To: 'Foxboro DCS Mail List' (E-mail)
Subject: Rolling average
Hi all,
What I would like is a way to calculate the average position of a valve over
the last hour, updated say every minute.
and then available as an input for a block. Any ideas?
Thanks in advance.
Steve Rigby
Air Liquide Industrie BV
Merseyweg 10
3197 KG Botlek-Rotterdam
The Netherlands
Tel: +31(0)181-271131
Fax:+31(0)181-271100
Email:
[EMAIL PROTECTED]
[EMAIL PROTECTED]
RE: Rolling average
I would use an independent sequence block. Run the block once per minute, dump the current value of the point to be averaged into a real array indexed by the minute in the current hour, sum up the values in the array, divide by 60 and you have a rolling hourly average. This will automatically overwrite a value which is over an hour old. You have some details to work out, such as how to handle a bad status on the input, getting the minute in the current hour, how to handle the first hour, initializations, restarts, etc., but it should be pretty straightforward. John Windle The Foxboro Company -Original Message- From: Steve Rigby [SMTP:[EMAIL PROTECTED]] Sent: Wednesday, September 06, 2000 9:01 AM To: 'Foxboro DCS Mail List' (E-mail) Subject: Rolling average Hi all, What I would like is a way to calculate the average position of a valve over the last hour, updated say every minute. and then available as an input for a block. Any ideas? Thanks in advance. Steve Rigby Air Liquide Industrie BV Merseyweg 10 3197 KG Botlek-Rotterdam The Netherlands Tel: +31(0)181-271131 Fax:+31(0)181-271100 Email: [EMAIL PROTECTED] [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
RE: Rolling average
This may be overly simplistic, but what about using a filter in an AIN block or an LLAG block? A low pass filter really isn't mathematically the same as a rolling average, but for a lot of the applications where the process engineers thought they needed a rolling average, we've found that it's close enough. You'd either need to do the math to calculate the filter time or experiment until you get a satisfactory value. You could set the PERIOD to obtain the desired sampling time. Valve position (or anything else where you don't need accuracy to the nth digit) would seem to be a good candidate for a filtered value. -Original Message- From: Steve Rigby [mailto:[EMAIL PROTECTED]] Sent: Wednesday, September 06, 2000 8:01 AM To: 'Foxboro DCS Mail List' (E-mail) Subject: Rolling average Hi all, What I would like is a way to calculate the average position of a valve over the last hour, updated say every minute. and then available as an input for a block. Any ideas? Thanks in advance. Steve Rigby Air Liquide Industrie BV Merseyweg 10 3197 KG Botlek-Rotterdam The Netherlands Tel: +31(0)181-271131 Fax:+31(0)181-271100 Email: [EMAIL PROTECTED] [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
RE: Rolling average
My personal favorite is: 1) Build a DTIME block and set the duration to the time period of interest * 1.2. 2) Take the 8 most recent bucket outputs of the DTIME block to the inputs of a SSEL block 3) Set the SSEL block to average its inputs. This approach is kind of cool because: 1) It requires no coding 2) It is pretty darn easy to build 3) It deals intelligently with bad inputs. Remember my Implementation Rules: 1) Do it in regulatory control blocks. (most efficient; least prone to problems) 2) If the regulatory control blocks can't do it, use a CALC block. (efficient, but must deal with BAD and coding errors; debuging is a pain) 3) If the CALC block can't do it, use a Sequence Block (pretty inefficient, but in the CP. Code/compilation relatively complex. Debugging is a pain.) 4) If you can't do it in a sequence block, reconsider if you need to do it. 5) If you must do it, see if you can do it in a shell (sh, ksh, csh, perl, etc.) script (easier to debug, easy to schedule, remember to use omgetimp and omsetimp to avoid broadcasts, remember to use show_params to check the IMPORT table) 6) If you can't do it in a script, reconsider doing it. 7) If you must do it, see if there is a 3rd party or Foxboro package that will do it. (Ask on the mailing list.) 8) If you must do it and can't find a package, write a program. Regards, Alex Johnson The Foxboro Company 10707 Haddington Houston, TX 77043 713.722.2859 (v) 713.722.2700 (sb) 713.932.0222 (f) [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] -Original Message- From: Steve Rigby [SMTP:[EMAIL PROTECTED]] Sent: Wednesday, September 06, 2000 8:01 AM To: 'Foxboro DCS Mail List' (E-mail) Subject:Rolling average Hi all, What I would like is a way to calculate the average position of a valve over the last hour, updated say every minute. and then available as an input for a block. Any ideas? Thanks in advance. Steve Rigby Air Liquide Industrie BV Merseyweg 10 3197 KG Botlek-Rotterdam The Netherlands Tel: +31(0)181-271131 Fax:+31(0)181-271100 Email: [EMAIL PROTECTED] [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED] --- This list is neither sponsored nor endorsed by the Foxboro Company. All postings from this list are the work of list subscribers and no warranty is made or implied as to the accuracy of any information disseminated through this medium. By subscribing to this list you agree to hold the list sponsor(s) blameless for any and all mishaps which might occur due to your application of information received from this mailing list. To be removed from this list, send mail to [EMAIL PROTECTED] with unsubscribe foxboro in the Subject. Or, send any mail to [EMAIL PROTECTED]
