Reconciling Manual J with Real World

beenthere · Mar 11, 2007

Those are the published ratings.

And thats why I suggested he look to lower his gain.

pstu · Mar 11, 2007

>>Why do you mention latent when it is NOT the critical factor?

Simply because I though you were supposed to look at the latent load as well as the sensible load. It may well be that the latent capacity will always be in excess but I did not know that.

Regards -- Pstu

.

jkish · Mar 11, 2007

pstu said:
In the world of engineering, to get within 10% of the real number is a very good result. Be thankful you are that close.

>>Manual J says 37K sensible load which leads to 5 ton unit after derating.

Does anyone else see this as a peculiar statement, perhaps a typo? My own house says 48.2K sensible load plus 7.4K latent load, which amounts to 4.5 tons nominal. I have not done any derating yet. Why do you mention sensible load but not latent? Do you really derate 5 tons down to almost 3 tons?

My numbers however are using Hvac-Calc's "loose" house assumption which I know is worse than the truth for my house. A blower door test tells me tightness is 5.3 ACH50, Manual J rate 0.24/hr summer. What are you using for infiltration figures and is it purely an estimate?

Hvac-Calc only allows you one digit of precision for infiltration, which I find a little awkward. For me to round down to 0.2 from 0.24 sounds presumptions and perhaps risky, while to round up to 0.3 just doesn't seem right.

Best wishes -- Pstu

.

From what I've seen, the sensible rating is the gating item on most of the sizing issues. Total load is much more easily met. Almost all 4 ton units I've seen are running in the 35-36K range for sensible load. Currently, I'm sitting right above that.

In HVAC-Calc, I'm using .4 ach winter and .3 summer for infiltration. I had a blower test run, but the numbers don't make too much sense to me. I see the CFM50 number at 2233. The volume of the first two floors (not including basement) is probably around 50000 cu ft. The formula I've seen is ACHnat = CFM50/20 * (60/volume). If I use that formula, I get an ACHnat of under .2.

If I go back and put .2 in for infiltration in HVAC-Calc, my real world numbers almost match exactly with the Manual J, but I may be doing something wrong. The tester (who was brain dead by the way) listed ACHnat at .59 but I don't know how he got that value.

jkish · Mar 11, 2007

dan sw fl said:
Lots of windows means > 400 total square feet of glass

Did you account for internal shading in at least the bedrooms?
Is it reasonable to add film to the west facing windows?

TRANE 2-stage model 4TTX6048 with 4TEE3F049 at 1,600 CFM
Sensible is rated at 35,000 BTUh
for 76'F /63 'F DB/WB, 49% R.H. return air condition, 95'F outside.

First floor has 476 sq ft of windows.
Second floor has 238 sq ft of windows

I marked windows with blinds or curtains in HVAC-Calc. I also gave some windows a percentage of external shading due to large overhangs.

Not too many directly facing west windows, but house faces SW. A lot of window area faces NE.

faith · Mar 11, 2007

Validation of Manual J results

Please allow me to preface the following reflections by honoring Don Sleeth with a deep felt appreciation for the wonderful mission he has accomplished by automating ACCA Manual J Seventh Edition into a format that facilitates its implementation by a broad and diverse audience. Good Job!

That said, to attempt to meet jkish’s expectations I offer the subsequent (perceived) realities.

ACCA Manual J Seventh edition was authored by Hank Rutkowski, P.E. (Professional Engineer) and is at best an approximation of procedures found in the ASHRAE Handbook of Fundamentals (and other related documents). Any serious effort to better understand the methodologies of non-empirical approaches to heat loss/gain in structures should be in the ASHRAE literature, not ACCA publications.

I had to smile at the use of monthly averages to attempt to justify, validate or otherwise corroborate observed data with theoretical calculated results.

Important Note: If you do not have a working knowledge of calculus read no further.

What is necessary is to bring the structure in question to it’s designed temperature on a challenging day (approaching design minimums or maximums) and simply discount the HVAC system and record the actual temperature drift toward outdoor ambient. Typically within six hours you should have sufficient data to interpolate between data points and extrapolate beyond data points.

The delta increment necessary to capture meaningful results to accurately calculate a legitimate measure of the thermodynamics of the structure in question is in the magnitude of minutes, not months, understood?

Might I suggest that you place temperature sensors in all critical rooms of the structure in question (ideally recording devices capable of readouts to a time increment of significance to allow serious analysis).

Armed with this empirical data, you can use integral calculus one can determine the amount of energy available in the structure at any given point of time. Only then will you begin to approach the realities of equalibrium(s) involved.

Sound strange? It’s not, and in fact this methodology is used quite widely in a vast number of sciences to achieve the ends you desire. Validation of a theoretical model.

Once again, Don Sleeth is a fantastically gifted gentleman and a scholar. HEAT-CALC is one of the finest tools ever developed for residential heating/cooling loads.

God Bless

jkish · Mar 11, 2007

faith said:
Might I suggest that you place temperature sensors in all critical rooms of the structure in question (ideally recording devices capable of readouts to a time increment of significance to allow serious analysis).

Armed with this empirical data, you can use integral calculus one can determine the amount of energy available in the structure at any given point of time. Only then will you begin to approach the realities of equalibrium(s) involved.

Thanks for the interesting thoughts.

I suppose that if you had actual heat loss vs. time data at the design temperature, then you wouldn't really need a Manual J calc.

What you describe seems like a validation of the Manual J method itself. Emperical evidence (results from usage) seems to say that Manual J does a pretty good job of estimating heat loss and gain at single point temperature inputs. If so, then it would also seem valid to compare single point average heat loss / gain with the Manual J calculations. Based on some testing of results, Manual J seems to produce linear results from different temperature inputs which makes using an average design temp (somewhat) feasible.

I took a little time to try and compare the cooling load predicted by HVAC-Calc to my actual cooling bills. The results seem to indicate that Manual J overstates cooling load by some amount. In my case, that was consistently about 33%. This matches some published work that also states that Manual J overstates cooling load to some degree.

suemarkp · Mar 11, 2007

I believe another factor ignored by Manual J (and perhaps HVAC Calc, haven't used that), is the heat generated by electrical appliances in your house. They do have a factor for that in cooling, but not heating. And if they did, you'd have to look at your electric bill to obtain a value to use and hope it covers only items in the conditioned space (no outside compressors, pool pumps, etc).

I'd be real pleased with a 10% accuracy value. I didn't do a blower door test, nor do I know how well the insulation in my house was installed. I do know the window values since I recently replaced them.

beenthere · Mar 11, 2007

Can't allow for appliances that may not be on for 10 to 12 hours.
Have to be able to heat the kitchen in the middle of the night when appliances aren't on.

Shophound · Mar 11, 2007

suemarkp said:
I believe another factor ignored by Manual J (and perhaps HVAC Calc, haven't used that), is the heat generated by electrical appliances in your house. They do have a factor for that in cooling, but not heating. And if they did, you'd have to look at your electric bill to obtain a value to use and hope it covers only items in the conditioned space (no outside compressors, pool pumps, etc).

I'd be real pleased with a 10% accuracy value. I didn't do a blower door test, nor do I know how well the insulation in my house was installed. I do know the window values since I recently replaced them.

Why would heat generated by appliances be a concern for doing a heat loss calculation?

Appliance heat is a heat gain...in winter that works in one's favor. Manual J heat loss is sized for coldest part of night with lowest activity level in house. As beenthere said, gotta keep the kitchen warm so the pipes don't freeze.

suemarkp · Mar 13, 2007

It is not a concern, but it is an ignored factor which would affect the calculated result versus the actual result. I think it is right to ignore it for heating. But the OP asked why his actuals and calculated didn't match. On the days when you're using a lot of power (lets say you're cooking a turkey in the oven all day), your furnace won't have to work as hard so it will appear that the furnace is oversized.

If manual J truely sizes for night and not an overall happy average, then you'd have to monitor gas usage only at night and compare only that along with the night time average temperature instead of looking at a months worth of gas usage.

Reconciling Manual J with Real World

beenthere

pstu

jkish

jkish

faith

jkish

suemarkp

beenthere

Shophound

suemarkp

Top Contributors this Month

Recommended Communities