PDA

View Full Version : infoSWMM Time Pattern for Weekends. Changed?



Stewart Sargent NZ
November 10, 2014, 08:56 PM
Can someone help me clarify how Time Patterns for DWFs interact.-- especially the Weekend one

ver= 13 Update3

In the Past the patterns in DWF were, I believe, used as follows
Avg x Patt1 x Patt2 x Patt3 x Patt4
& IFF (if and only if) a Pattern had a type =WEEKEND it was applied during (and only during) weekends -- [and others were always applied]

eg if Patt1=monthly, Patt=Daily, Patt3=Hrly, & Patt4=Weekend
then Patt3 was effectively the variation during the day (by hr) for ALL days
BUT on the Weekends it was modified by Patt4.
[ From HELP>Dry Weather Inflow
A pattern multiplier can be specified for each of the following:

Month of the year
Day of the week
Hour of the day
Weekend (a weekend pattern is only applied to Saturday and Sunday, otherwise the multiplier value is 1.0 for weekdays)

If multiplier patterns are specified then the flow into the node at any given timestep is (Average DWF * Monthly Multiplier * Daily Multiplier * Hourly Multiplier * Weekend Multiplier)
]

BTW. I also presume that the Pattern TYPE controls the way they are used NOT the postion they are specified (ie, it is the TYPE=Weekend, rather than a pattern positioned at Pattern4, that controls whether it is used during the Weekends).
???Can anyone cofirm this????

????NOW it seems as if:
The Hrly patterns are ONLY used on weekdays and
The Weekend Patterns are ONLY used on weekends
ie an either/Or situation, Thus, for the eg above;
Avg x Patt1 x Patt2 x ( Patt3 | Patt4 )


Can anyone confirm or deny this for me ... or correct me if I have the basic methodolgy wrong.

Robert Dickinson
November 11, 2014, 10:39 PM
Hello Stewart,
That is a good observation on your part. We are now using the SWMM 5.1.007 engine and the way the DWF is calculated was changed in the new SWMM 5 code. SWMM 5 code changes can be seen here

http://www2.epa.gov/sites/production/files/2014-10/epaswmm5_updates_0.txt

and the change that applies to your comment is number 52. I mentioned this was a problem in the past to the EPA and they changed it for the new engine, so you can blame me for this development.

52. Weekday and weekend hourly time patterns for Dry Weather inflows are now correctly applied in a mutually exclusive manner.From your comment, it used to be as you say:If multiplier patterns are specified then the flow into the node at any given timestep is (Average DWF * Monthly Multiplier * Daily Multiplier * Hourly Multiplier * Weekend Multiplier)"but now it is If multiplier patterns are specified then the flow into the node at any given time step is (Average DWF * Monthly Multiplier * Daily Multiplier * [Hourly Multiplier or Weekend Multiplier])It uses the weekend pattern if it is the weekend and the hourly pattern during the week or there is no weekly pattern.I think this is a better way of doing the multiplication and sorry for the confusion.Regards,Robert DickinsonInnovyze Inc.....the code for doing this

p1 = inflow->patterns[MONTHLY_PATTERN];
if ( p1 >= 0 ) f *= inflow_getPatternFactor(p1, month, day, hour);
p1 = inflow->patterns[DAILY_PATTERN];
if ( p1 >= 0 ) f *= inflow_getPatternFactor(p1, month, day, hour);
p1 = inflow->patterns[HOURLY_PATTERN];
p2 = inflow->patterns[WEEKEND_PATTERN];
if ( p2 >= 0 )
{
if ( day == 0 || day == 6 ) // weekend pattern if it is the weekend
f *= inflow_getPatternFactor(p2, month, day, hour);
else if ( p1 >= 0 ) // hourly pattern if if is NOT the weekend
f *= inflow_getPatternFactor(p1, month, day, hour);
}
else if ( p1 >= 0 ) f *= inflow_getPatternFactor(p1, month, day, hour);
return f * inflow->avgValue;

Stewart Sargent NZ
November 12, 2014, 12:33 AM
Thanks Robert.
Yes the new method is more sensible and will be easier to manage. The old way could be a real trap.

HOWEVER (don't jut 'luv' that word) -- ALL old models will need to be modified before they will give the same answers when rerun. It could be a big pit-fall into which may could plunge - me-thinks (speaking from experience and torn fingernails and reputation from having to climb out of the hole >grin< )

SO.... Users need warned and the HELP needs to be very clear and have a WARNING (at least for a year or so). Maybe even a warning in the dialog boxs

Anyway -- new method will be good... and making the transition is not too hard (provide one doesn't fall ... before making the change).
Tanks again.

Robert Dickinson
November 12, 2014, 12:39 AM
Thanks Robert.
Yes the new method is more sensible and will be easier to manage. The old way could be a real trap.

HOWEVER (don't jut 'luv' that word) -- ALL old models will need to be modified before they will give the same answers when rerun. It could be a big pit-fall into which may could plunge - me-thinks (speaking from experience and torn fingernails and reputation from having to climb out of the hole >grin< )

SO.... Users need warned and the HELP needs to be very clear and have a WARNING (at least for a year or so). Maybe even a warning in the dialog boxs

Anyway -- new method will be good... and making the transition is not too hard (provide one doesn't fall ... before making the change).
Tanks again.

Another astute comment, Stewart. I had an Excel Macro in the past to untie the Weekend and Daily pattern connection. I think to get the same answer you need to multiply the old hourly * the weekend hours and call it the new weekend pattern. As i said in some past blog about this double multiplying:

"Dry Weather Flow in InfoSWMM and H2OMap SWMM

Dry weather flow can be added to any node in H2OMAP SWMM. The dry weather flow is computed as the average flow * the monthly pattern * the daily pattern * hourly pattern * the weekend daily pattern to give the Dry Weather Flow at any time step (Figure 1). Since the four types of patterns (Figure 2) are all multiplied together then for Saturday and Sunday the hourly pattern and the weekend hourly pattern will both be used. This will have the effect of overestimating the flow if the multipliers are greater than 1 and underestimating theflow if the multipliers are less than one. You should enter the Pattern X for the Weekend Hourly Pattern in H2OMAP SWMM where
X = Weekend Hourly Pattern / Hourly Pattern
So that when the pattern X is multiplied by the Hourly Pattern the program will use the intended Weekend Pattern."

Stewart Sargent NZ
November 12, 2014, 11:36 AM
Thanks Robert.
Yes that is basically what I did to move from old to new method. Not difficult - just a bit laborious to go thru all models and update them. But doesn't take long (as long as one doesn't have scrabble out of a Trap hole first <grin>.... like I did).