Building energy usage reduction by automation of optimized plant operation times and sub-hourly building energy forecasting to determine plant faults

摘要

The invention provides a method for improved building energy usage reduction by computer automation of optimized plant operation times and sub-hourly building energy forecasting to determine plant faults. The invention provides a computer system to derive the NTL, mechanical heat-up (MHL) and mechanical cool-down (MCL) lags and in conjunction with a readily available interval weather forecast, the system can output various signals to indicate optimized start and stop times for heating and cooling equipment. The algorithm to calculate the 15-minute energy forecast is used to indicate out-of-control conditions in the operation of the plant.

主权项

1. An improved computer implemented method to forecast energy usage in buildings in 15-minute intervals, wherein the computer performs the following steps:
a) determination of the unique building's natural thermal lag (NTL) by linear regression based on energy usage and external temperature; b) using the NTL to generate the best fit regression model to forecast energy usage on a daily total basis; c) determine the days from the model training data which exhibit a similar external temperature profile over the working part of the day in question to yield a set of ‘matching days’ by use of the following equation:SS=∑i=1p(TAFi-TCi)2Eqn1 where SS represents the summation of the squares over the period i=1 to p, where p matches the optimal window size as determined in the previous section p is the number of 15 minute observations we wish to examine (e.g. 40 for a typical 10-hour working day window) TAFi is the actual outside temperature at time period i, on the day the energy usage has been forecast TC1 is the actual external temperature at time period i on any day to which the forecast day is being compared d) using the energy profile exhibited during these ‘matching days’, proportionately assign the forecast of total daily energy for the day in question into 15-minute usage figures.