Weather pattern based electrical demand forecasting for a building

• US 10,068,298 B2
• Filed: 10/14/2015
• Issued: 09/04/2018
• Est. Priority Date: 08/22/2014
• Status: Active Grant

• Current Assignee: Siemens Corp. (Siemens AG)
• Original Assignee: Siemens Corp. (Siemens AG)
• Inventors: Srivastava, Sanjeev
• Primary Examiner(s): Wang, Philip
• Application Number:
  US14/882,480
• Publication Number:
  US 20160246272A1
• Time in Patent Office: 1,056 Days
• Field of Search: None
• CPC Class Codes: G05B 15/02 electric
• CPC Class Codes: G06Q 50/06 Energy or water supply

Abstract

Embodiments include methods and systems for forecasting electrical demand of a building. Aspects include receiving a weather forecast for an upcoming day and receiving a historical data file comprising historical weather data and historical electrical load data for the building for a plurality of previous days. Aspects also include comparing the historical weather data for each of the plurality of previous days to the forecast and calculating a baseline electrical consumption of the building for the upcoming day based on the comparison.

Claims

1. A computer based method for forecasting electrical demand of a building for short response to an upcoming demand response event, the method comprising:

receiving a weather forecast for a demand response event of an upcoming day;

receiving a historical data file comprising historical weather data and historical electrical load consumption data for the building for a plurality of previous days;

ranking N days of historical weather data as best matching the weather forecast, comprising;

comparing, by a processor, the historical weather data for each of the plurality of previous days to the weather forecast, wherein the comparing includes calculating an average difference between at least one weather metric for one day of the historical weather data and the at least one weather metric for the weather forecast for a plurality of n time intervals;

calculating correlation coefficients for historical load consumption data and the at least one weather metric;

calculating a weighted error for each of the plurality of previous days using the average difference and the correlation coefficients; and

determining N days having the least calculated error; and

calculating, by the processor, a baseline electrical load consumption forecast of the building for the upcoming day based on the historical electrical load consumption data of the N days comparison;

calculating, by the processor, a ratio of an average electrical load consumption of the building for a past number n of time intervals and an average baseline electrical load consumption for the past number n of time intervals; and

determining the electrical load demand forecast by adjusting the baseline electrical load consumption by the ratio.

2. The method of claim 1, wherein the comparing includes calculating, for each of n time intervals, an absolute difference of a first weather metric between a first forecasted weather metric of the weather forecast and a first historical weather metric of the historical weather data for one of the plurality of previous days.

3. The method of claim 1, wherein calculating the weighted error comprises calculating a first influence factor between the historical electrical load and a first weather metric and a second influence factor between the historical electrical load and a second weather metric.

4. The method of claim 3, wherein calculating the consumption weighted error comprises calculating weighted error for each of the plurality of previous days based on the first influence factor and the second influence factor.

5. The method of claim 1, wherein the determining electrical load demand is performed in response to receiving a demand response signal prior to the demand response event.

6. The method of claim 5, wherein the demand response signal is received within 15 minutes before an occurrence of the demand response event.

7. An electrical demand forecasting system comprising:

a processor in communication with one or more types of memory, the processor configured to;

receive a weather forecast for a demand response event of an upcoming day;

receive a historical data file comprising historical weather data and historical electrical load consumption data for the building a plurality of previous days;

rank N days of historical weather data as best matching the weather forecast, comprising;

compare the historical weather data for each of the plurality of previous days to weather the forecast, wherein the comparing includes calculating an average difference between at least one weather metric for one day of the historical weather data and the at least one weather metric for the weather forecast for a plurality of n time intervals;

calculate correlation coefficients for historical load consumption data and the at least one weather metric;

calculate a weighted error for each of the plurality of previous days using the average difference and the correlation coefficients; and

determine N days having the least calculated error; and

calculate a baseline electrical load consumption forecast of the building for the upcoming day based on the historical electrical load consumption data of the N days comparison

calculate a ratio of an average electrical load consumption or the building for a past number n of time intervals and an average baseline electrical load consumption for the past number n of time intervals; and

determine the electrical load demand forecast by adjusting the baseline electrical load consumption by the ratio.

8. The electrical demand forecasting system of claim 7, wherein the comparing includes calculating, for each of n time intervals, an absolute difference of a first weather metric intervals between a first forecasted weather metric of the weather forecast and a first historical weather metric of the historical weather data for one of the plurality of previous days.

9. The electrical demand forecasting system of claim 7, wherein calculating the baseline electrical weighted error comprises calculating a first influence factor between the historical electrical load and a first weather metric and a second influence factor between the historical electrical load and a second weather metric.

10. The electrical demand forecasting system of claim 9, wherein calculating the electrical weighted error further comprises calculating weighted error for each of the plurality of previous days based on the first influence factor and the second influence factor.