Ultimately, how do we define what the typical load profile is for each industry?
A few initial ideas:
Industrial Assessment Center: reports annual production hours. Advantages of source is its many data points, detailed industry coverage, and additional info about each reporting facility (e.g., production, size). Disadvantages is its limitation to small-to-medium sized facilities and annual average. Approach to using data would be to estimate hourly operation (based on assumed number of 8-hour shifts/day). Then develop an approach to capture the distribution of operating hours within each NAICS code, potentially accounting for facility production and/or size (or location?).
Quarterly Plant Capacity (QPC) Survey (Census Bureau): reports quarterly average weekly hours of operation. Advantages are its reliability, provision of standard error estimates, weekly average data, and quarterly reporting (potentially capturing seasonality). Disadvantages are its relatively small number of NAICS codes (~ 90, with a mix of aggregation). Approach to using data would be to estimate hourly operation from weekly average data, accounting for quarter. Then develop an approach to capture the distribution of operating hours within each NAICS code (develop 99% confidence intervals with standard errors?). As indicated by its name, the QPC survey also includes capacity data for each quarter. These data may be useful somehow.
Note about QPC survey standard error reporting:
Sampling error is the difference between estimates obtained from the sample and results theoretically obtainable from a comparable complete enumeration of the sampling frame. This error results because only a subset of the sampling frame is measured in a sample survey. Standard errors of the estimates shown in this publication, which are estimated measures of the sampling variability, are displayed along with the corresponding estimates in Tables 1 and 2 of this report. These standard errors may be used to define confidence intervals about the corresponding estimates with a desired level of confidence. Only one of many possible samples was selected. If confidence intervals were constructed for each of these possible samples, then it would be expected that the percentage of confidence intervals containing the complete coverage result would equal the percent of the level of confidence. For example, the interval defined by a margin of error of one standard error, in which the interval is constructed one standard error below the estimate to one standard error above the estimate, approximately yields a 68-percent confidence interval; the interval defined by a margin of error of two standard errors approximately yields a 95-percent confidence interval; and the interval defined by a margin of error of two and a half standard errors approximately yields a 99-percent confidence interval. In the Introduction for the 2017 fourth-quarter release, the margin of error given in parentheses for each of the three full production utilization rates (current quarter, prior quarter, and same quarter one year ago) is two standard errors, so that approximate 95-percent confidence intervals may be easily constructed using the sample for the 2017 fourth-quarter release.
Ultimately, how do we define what the typical load profile is for each industry?
A few initial ideas:
Note about QPC survey standard error reporting: