My father-in-law has recently added significant insulation to his basement windows and bulkhead. He created some great ways to plug the holes using bubble wrap and insulation in ways that also make it easy to install for the winter and store them away for the warm months.

He provided me with data from his oil bill, as well as the degree day information from his oil company for the years before and the year after this improvement.

One of the first problems I thought about in reviewing his data was that oil fill-ups don’t correspond well to the days that the oil was burned. I can see that there were 100 gallons put into the tank in March, for instance, but I can’t say how many of those gallons were used in the next week or the next month. The oil that goes into a tank in May may not get used until September.

The second problem to consider was that the oil company provided 9 months of degree data — from September to April. So, to try to match this up to the months when oil was added to tank required a little guesswork. I ended up throwing out some gallons of oil, hoping that I was making good guesses.

In the end, I did the calculations and spreadsheets three different ways, each a little more precise than the first. Below is the summary of the first (simplest) and third (most complex). The middle one provided data closer to the third, but basically in the middle.

**Simple Analysis:**

In the simple case, I used the oil company degree data AND the oil fill up data for only the months of September through April. After looking up a couple of estimates of kWatt-hours per gallon of heating oil, I used 1 gal=41KWh (which doesn’t take into account the efficiencies of the oil burner). I found an average number of degree days from the 5 years given by the oil company, 6245, and used that to ‘normalize’ the kWh and gallons of oil.

Year | Gal of Oil | kWh | Deg Days | kWh/DD | Norm kWh | Norm Gal | Savings |
---|---|---|---|---|---|---|---|

2004-2005 | 1169 | 47929 | 6487 | 7.4 | 46141 | 1125 | -$33 |

2005-2006 | 1036 | 42476 | 6020 | 7.1 | 44064 | 1075 | $93 |

2006-2007 | 1089 | 44649 | 6116 | 7.3 | 45591 | 1112 | $ 0 |

2007-2008 | 972 | 39852 | 6042 | 6.6 | 41191 | 1005 | $268 |

The first few columns are the numbers provided by the oil bills and the degree days provide by the oil company. The normalized kWh and normalized gallons are the amount of energy or gallons of gas your heating system would have used if there was the exact same number of degree days in each of these years. Now we can compare 2007-2008 against 2006-2007 and there is a savings of about 10%, which would be $268.28.

After going through these numbers I felt it would be much better if I could get a full year of data and more detailed degree days and see if the result was better. I used the degree day data from this website: DegreeDay.net, where I can set the base temperature and get monthly, weekly or even daily data. That way I can use the entire years worth of oil information as well. The only drawback is that for the zip code near this house, the degree data only goes back 2.5 years. So the monthly detailed analysis is only useful for the last two years.

**Complex Analysis: **

Year | Gal of Oil | kWh | Deg Days | kWh/DD | Norm kWh | Norm Gal | Savings |
---|---|---|---|---|---|---|---|

2006-2007 | 1225 | 50221 | 6075 | 8.3 | 51626 | 1259 | $ 0 |

2007-2008 | 1011 | 41467 | 5913 | 7.0 | 43796 | 1068 | $477 |

These numbers indicate a 15% increase in efficiency (savings in gallons of oil) between 2007 and 2008, which translates into a savings of $477 in a ‘typical’ year. That’s pretty nice.

There is a big difference between the first ‘simple’ estimate and the final one. The differences in the calculations that lead me to believe the final savings number is a better estimate of the savings than the simple one are: 1) I used more sophisticated degree day data, which allowed me to set the proper base temperatures and work with monthly data; and 2) I used the full year of data for both the oil fill ups and the degree data and set the calendar year to go from June to June assuming that the oil in the tank during the summer months was not used quickly.

I expect to return to this case study next year to see if the numbers are consistent. If you want the more detailed spreadsheet of numbers for this case study, please write to me at kim at energyemp.com.

This looks like a great analysis. Is it reasonable to do it on a monthly basis? My apartment has a leaky third floor bedroom, adjacent to an attic, and we haven’t attempted to insulate it any better yet. Is it reasonable to use our gas bill to determine whether we’re being successful at insulating?

Also, do you know how accurately degree-days correlate with actual energy used? Newton’s law of cooling suggests that rate of temperature change is proportional to the temperature difference, so it should take quadratically more energy when it’s colder out, but the degree day model seems to assume it is linear.

Let’s try a case study with a few months of data … But first, what is your plan for insulation of the 3rd floor? Can you plug holes with home made (or purchased) insulation products? How about plastic heat-shrunk to the windows, or heavy, well-fitting curtains. Maybe you would be willing to supply some pictures before and after.

You are right that this degree day model assumes a linear relationship with the amount of energy consumed. I would think that raising the temperature from 30F to 68F should take a lot more energy than 38 days of raising the temperature from 67F to 68F. Time to do some more digging into the data!