Air-conditioning load prediction method and system based on big data

Abstract

The invention discloses an air-conditioning load prediction method based on big data. The method includes the steps: 1) acquiring outdoor temperature t in DEG C of a site with an air conditioner at a simulated prediction moment; 2) acquiring a temperature interval [t-s and t+s] based on a predetermined prediction step size s in DEG C and the outdoor temperature t; 3) screening data in the temperature interval from historically accumulated big data of cold or heat load measured values, total power of lighting and devices, the number of indoor people, open water surface area, indoor and outdoor air enthalpy differences and indoor and outdoor temperature differences of the air conditioner running under stable conditions; 4) calculating final load value of the prediction moment based on screened date. Based on a lot of gradually accumulated historical data of practical cooling capacity and heating capacity of the air conditioner running under the stable conditions, models for calculating cold or heat loads at the simulated prediction moment are continually optimized, and the built models are more accurate with the passage of time. The invention further discloses an air-conditioning load prediction system based on the big data.

Description

technical field [0001] The invention relates to an air-conditioning load forecasting technology, in particular to an air-conditioning load forecasting method and system based on big data. Background technique [0002] With the rapid development of urban construction, building energy consumption is also increasing rapidly. The energy consumption of air-conditioning operation accounts for a large proportion of the energy consumption of buildings, especially large public buildings. Therefore, reducing the energy consumption of air-conditioning systems has always been the focus of building energy conservation. Adopting a reasonable operation adjustment method is one of the main technical ways to improve the energy utilization efficiency of the air-conditioning system, and the realization of this way depends on whether the air-conditioning load can be accurately predicted. It can be seen that the method of predicting the air-conditioning load is particularly important. [0003] ...

AI Technical Summary

Problems solved by technology

The characteristic of the time series forecasting model of air-conditioning load is that the autoregressive moving average time-series model with external input is used to forecast the hourly cooling load. It does not meet the requirements of stationarity, so the prediction accuracy is not high; the neural network method is to construct a suitable neural network structure, and then select the load of the past period of time as a training sample, and use a certain training algorithm to train the network to meet the accuracy requirements , using ANN for air-conditioning load forecasting, its shortcomings are that with the deepening of neural network in the research of air-conditioning hourly load forecasting, neural network is easy to fall into local minimum point, the network generalization ability is poor, and the prediction accuracy is not enough in practical applications Ideal and other new issues

The most commonly used air-conditioning system is the temperature difference and pressure difference control mode in the actual operation adjustment. These two control methods respectively predict the air-conditioning load change by collecting the temperature difference and pressure difference signal of the air-conditioning chilled water / hot water supply and return water. The disadvantages of the two control methods are: the position of the sampling point is a certain distance from the position of the real-time change of the load, which will cause a large delay in the signal transmission of the temperature / pressure change. In addition, the position of the signal detection point is difficult to locate accurately. It is easy to cause insufficient cooling / heating supply at some ends

Images