Re: [EnergyPlus_Support] Annual Peak cooling variable ?

On Tue, Sep 18, 2012 at 3:13 PM, Jeremiah Crossett <jcrossett@xxxxxxxxxxxxxxx> wrote:

Li- In response to Your statement---
No one, except you, would ask for the max peak value, to be displayed as a separate variable and cannot get it from the spread sheet yourself.

Check the old list by searching for requests to have the annual cooling loads reported to csv and you will notice that I am not the only one with this request.

Think about this: Say a user has 2 wall insulation thickness, 5 roof insulation thickness, 2 window U value, 2 solar heat gain coefficient, 4 PCM roof, 4 PCM wall options for a total of 1020 design iterations---This is a real building with a real engineer using the output to decide upon a the design. To decide such you could either guess (then why model) or you could model all iterations and choose the results that involve the tradeoffs that the building owner can afford- I opt to model, but am limited in ability to show the reduction of peak cooling because E+ does not show the peak value, rather shows the average.

I have even attempted work arounds such as running only the design day, but that produces the average value for the 24 hours- so is also not useful. This is something I was requested to produce by a firm who is wanting to use a valence unit HVAC for each space and the system size is limited, so the goal is to design the building with minimal peal cooling load.

This is a very useful output for any design, and why I publicly posted rather then posting to the help desk, as when I searched the old forum posts I noticed others with the same issue, and realized that this would be useful for others. (this is the most common question for HVAC designers- what can be done to reduce peak loads to minimize system size)

This is a useful request, that apparently you have not a real understanding of- possibly you have never actually explored such an approach (to report the peak value rather then the average for many models at once) and you might consider reading posts in there entirety before responding with such statements- it makes you seem presumptuous.

Jeremiah D. Crossett  | Senior Analyst  | Phase Change Energy Solutions
120 E. Pritchard St. | Asheboro, NC 27203 | Mobile 503-688-8951

www.phasechange.com



On Tue, Sep 18, 2012 at 11:55 AM, Jeremiah Crossett <jcrossett@xxxxxxxxxxxxxxx> wrote:

Linda, and Brent- I will be submitting this to the help desk for consideration for the wish list later on today- and note that in my request the same (or possibly more useful) result could be accomplished if there was a way to report the total component cost to output variable)

Total size of HVAC equipment
Total amount of items that reduce HVAC equipment size

With the use of objective functions users can examine tradoffs that allow the minimal HVAC equipment size and this will have a byproduct of minimizing energy consumption but at the lowest first cost.

Dr. Li,
Thanks for your comments but-----You never cease to amaze me with your presumptions about what might be useful outside of your non market driven single design iteration ddy only runs. I have read the documentation and no the average annual does not correlate to the smallest calculated system size. As Linda said there is not currently a way to do this, possibly you did not read the entire post.

If you want to see what I am dealing with to possibly understand why this is needed for parametric studies or optimization routines you can download the 1020 simulation run from here:

ftp://ftp.nrgsim.com/
user: nrgsim
passcode: globalnrgmodel

And can view the combined results here: https://docs.google.com/spreadsheet/ccc?key=0Ag7cD5ZV9QwRdFgyWjJXN3IyTVduc2VrcXRlLU95ZlE

No one, except you, would ask for the max peak value, to be displayed as a separate variable and cannot get it from the spread sheet yourself.
Check the old list by searching for requests to have the annual cooling loads reported to csv and you will notice that I am not the only one with this request.

Think about this: Say a user has 2 wall insulation thickness, 5 roof insulation thickness, 2 window U value, 2 solar heat gain coefficient, 2

A simple program would extract the time and values from the ESO file, and give you an output, variable name, value min at time min, value max at time max..May be not simple, if you only code using Microsoft Visual series SDK. I believe that the consol mode is still it there..
This is not possible for thousands of parametric options or millions optimization options because of the file structure of directories that are populated with the results.

Annual max and annual min are EPW dependent data, and are already given in the .stat file. Environment, annual and runPeriod has nothing to do with the max single peak value..
The value I am referring to would be to have a setting in the output varrible reporting frequency that provides an option of the model results to produce minimum or maximum values from the .eso file, the epw file is non-relevent to this requirement.

The ReadVarsESO.exe file is in the PostProcessor folder. There is also a batch file example there.
Obviously

There is a number of exercise for you to do in the AdvancedOutput folder of the exampleFiles folder. The instructin is in ExerciseOutput1 Instructions.pdf.

Obviously, but this does not address or solve the issue.

After you have used the .rvi file to select the required variable to the .csv file, click on the cloumn and ask the spread sheet to show the max and min value. It will be a waste of resources to create these two values as separate variables..

This is incorrect- currently E+ averages the annual values then prints those to CSV, so when sorting with the spreadsheet the values that are sorted put the highest and lowest AVERAGED values. This correlated to number 288 down the list for for the run that had the lowest MAXIMUM system size as reported to HTML/ESO.

This is not the first time that "how to use .rvi file" was asked in the forum.
You provided relevant and helpful feedback to a user on July 13th about a similar subject of RVI, but that was an entirely different request than this.

The variable names in the .rvi file are those in the header of the .eso file. If the variable is not in the Outpupt variable list, it will not be there. If the variable name is not in the .rdd file, it will not be generated during simulation.   This is why you cannot generate your own name with a "max" in it..
Obviously, this is what I am requsting to be added to E+ as a new feature to solve the problem that E+ is producing annual averages for a value that should not be averaged, and would be useful to have MIN/MAX reported to csv via the readvarseso, and is only possible if it is added to the IDD.

As you use 60 as timestep for simulation, all the variable will reported in one minute interval. If you asked for hourly result, of cource, the value must be avaraged. I have mentioned to you many times, that I do not use averaged daily, monthly and annual result.

Incorrect- I use 60 timesteps for all of my runs becuse I model PCM's (required to be no less than 20) and when not modeling PCM's for increased accuracy. The hourly results can not be used for large scale parametric or optimisation studies because the requirement is to have single annual numbers for each iteration- it is impossible to run millions of models and compare hourly values.

If you add the names listed below in the IDF, all of them will be in the .ESO file as hourly averaged. Only when you have more than 250 columns, then you need to change the 250 to no limit and then use .rvi file to customize the .csv file.

Obviously,(this is in the documentation becuse it is the limit of excel, but could be overriden for viewing results from .eso or .sql) and why would needing one peak annual output of sensible cooling rate need more than 250 colums?

Add two more rows at the bottom of the spread sheed and fill them with formula to extract max. and min. from all the data above.

Incorrect, this would produce the MIN/.MAX of the annual averaged number produced in the .eso file. If it was that simple I would not have posted this. For that to work the max values would have to be the starting numbers in the csv file but are not.

End of your task.

Not possible without max values, as I said the max value comes in at # 288 of 1020 runs because the current logic is to report to csv the annual average. O
I do not think that your 1000 odd parameter simulation data mean anything, if you to not have a linked optimization rules.

Respectfully I do not think you have ever done such a parametric simulation, the linked optimisation rules are the objective functions of first cost vs total consumption.If you can reduce the first cost and minimize total consumption than you have an optimal design. This can be done via parametric studies if you wish to explore all design alternatives or by optimisation if you wish to solve the equation with a single optimal design. Again respectfully I do not think you have any idea what you are talking about when it comes to parametric or optimisation runs as you have said that you have not done one before.

The peak cooling load cannot be reduced by simulation. The value itself does not mean anything. You need to determine what is the source causing the maximum, it it is not the weather condition.
Obviously the weather can not change and yes the goal is to determine the cause of the maximum, hence the need for parametric studies or optimisation runs.
Yes the peak cooling load can be reduced by changing the any of the following examples (and surly more possibilities) window size, window type (U/SHGC) Insulation thickness, PCM type/position, orientation, lighting type.

The only way to use the output to determine what is the cause of the maximum is to have the maximum reported to csv as a single annual output- hence the entire premise of this Post (I guess you missed the point?)

Thanks for your response- I enjoy the way you cause me to explain in detail the problem space, and do learn form our conversations (as I hope you do as well)

Best-

Jeremiah D. Crossett  | Senior Analyst  | Phase Change Energy Solutions
120 E. Pritchard St. | Asheboro, NC 27203 | Mobile 503-688-8951

www.phasechange.com



On Tue, Sep 18, 2012 at 10:58 AM, YuanLu Li <yli006@xxxxxxxxxxx> wrote:

See my comments below.

No need to reply to my comments, as there are enough said on this topic..

Dr. Li

To: EnergyPlus_Support@xxxxxxxxxxxxxxx
From: jcrossett@xxxxxxxxxxxxxxx
Date: Mon, 17 Sep 2012 16:49:17 -0700
Subject: Re: [EnergyPlus_Support] Annual Peak cooling variable ?

Bent & Linda Thanks- Could I ask if this feature could be considered for any version of E+, and if you can understand why it would be left out of version 1?

No one, except you, would ask for the max peak value, to be displayed as a separate variable and cannot get it from the spread sheet yourself.

A simple program would extract the time and values from the ESO file, and give you an output, variable name, value min at time min, value max at time max..May be not simple, if you only code using Microsoft Visual series SDK. I believe that the consol mode is still it there..

I was thinking it would make a good replacement for the redundant "environment", "annual" and "Runperiod" options- and could be a simple change to have one "Annual" and add an option for "Annual Max" and an option for "Annual Min"..

Annual max and annual min are EPW dependent data, and are already given in the .stat file. Environment, annual and runPeriod has nothing to do with the max single peak value..

Prior to posting this searched documentation for readvarseso, read every page of documentation (1832 to 1834 input output manual and pages 34 to 36 of "tips and tricks" and page 206 of "axillary programs")  I do not see any documentation as to how one would attain the information from readvarseso that is not contained in the Input Data Dictionary. If there is a way to use .RVI files, or to set the IDF to report the maximal value, or to use get the maximum data out of the .ESO files that would be great.

The csvproc program runs statistics on hourly values, where I am looking to use .RVI file for annual values so this does not appear to be of any use.

I spent three days on this before I posted, and I have read all documentation I could think of, but will go-ahead and look some more considering you are saying that this is possible and documented. (any advise where to look?)

The ReadVarsESO.exe file is in the PostProcessor folder. There is also a batch file example there.

There is a number of exercise for you to do in the AdvancedOutput folder of the exampleFiles folder. The instructin is in ExerciseOutput1 Instructions.pdf.

After you have used the .rvi file to select the required variable to the .csv file, click on the cloumn and ask the spread sheet to show the max and min value. It will be a waste of resources to create these two values as separate variables..

This is not the first time that "how to use .rvi file" was asked in the forum. The variable names in the .rvi file are those in the header of the .eso file. If the variable is not in the Outpupt variable list, it will not be there. If the variable name is not in the .rdd file, it will not be generated during simulation.   This is why you cannot generate your own name with a "max" in it..

Jeremiah D. Crossett  | Senior Analyst  | Phase Change Energy Solutions
120 E. Pritchard St. | Asheboro, NC 27203 | Mobile 503-688-8951

www.phasechange.com



On Mon, Sep 17, 2012 at 3:49 PM, Linda Lawrie <linda@xxxxxxxxxxxxxx> wrote:

At 04:25 PM 9/17/2012, Jeremiah Crossett wrote:

The need for an object that can output the annual min/max loads to a simple single number that can be compared with other design iterations is common and should be considered for inclusion in Energy Plus V7.2.

Will not happen for V7.2.

Are you saying that ReadVarsEso PostProsser is capable of producing CSV files with the peak value of sensible cooling rate by some undocumented method?

The ReadVars is a very simple method of reading the .eso file. Brent is saying that the .eso file contains the information you want. (And, it is documented).

Have you looked at the csvproc program? I do not remember if that might help you or not.

Linda

Niraj Poudel, Architectural Engineer.

PhD student, PDBE Program.

Clemson University, Clemson, SC.