JPH08247522A - Heat load forecast device - Google Patents

Abstract

PURPOSE: To improve the accuracy of heat load forecast and to automate the handling of a computing device by receiving weather forecast data delivered from a weather forecast data-providing source and computing heat load forecast values needed, in the future, for a heat source device based on the data. CONSTITUTION: Actual operation data and actual weather data outputted from a detector 3 of a heat source device 2 are inputted into a personal computer 1 to correlate the data with time.day of the week data, which are successively stored in storage devices 5, 6. A heat load forecast computing device 12 computes a weight coefficient based on these actual stored data. On the other hand, weather forecast data published from a weather forecast providing source 7 are inputted into a personal computer 1 through a circuit L to store the data in a storage device 9. A forecast heat load is computed every one hour by a neural network heat load forecast means 14 based on the weather forecast data and actual data so that the results of computation are outputted to a heat source operation control computing device 17 to control the operation of the heat source.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a heat load predicting device for predicting a heat load utilizing weather information in a heat source device system for operating and controlling a heat source device for an air conditioning facility of a building.

[0002]

2. Description of the Related Art In an air conditioner for a building, a heat load required for the next day is predicted from the viewpoint of energy saving, and correspondingly, a cheap power rate at night is used to store heat. In addition, it is important to predict the heat load required for setting pool water and facility temperature in advance even at events such as indoor pools and various leisure facilities.

One of the factors that greatly affect the predicted value
There are meteorological data such as temperature, humidity and rainfall rate, which are given as meteorological forecast data. An example of using this for heat load prediction is disclosed in Japanese Patent Laid-Open No. 4-365101. Here, in addition to weather forecast data, past operation record data and day of the week are input as factor data for heat load prediction, and heat load prediction is performed using a neuro network to improve accuracy.

[0004]

However, in the above-mentioned conventional method, since the weather forecast data is manually input, an operator is required and the input is very troublesome. In addition, the weather forecast changes from moment to moment, and in order to accurately predict the load, it is necessary to input the latest weather forecast data announced by the weather forecast provider at regular intervals into the arithmetic unit. . In the conventional method, it is possible to input the latest weather forecast data in real time, but since it is a manual operation by the operator, the weather forecast data is input once a day for practical reasons due to the inconvenience. Therefore, there is a problem that the accuracy of the heat load prediction decreases. In particular, work outside the working hours is required for inputting data that is normally provided outside the working hours, which is not preferable because it increases the load on the worker. There is also a case where an operator makes an input error or forgets to make an input, which causes a fatal defect in accuracy in the load prediction process due to erroneous data, abnormal values, or omissions.

Therefore, in the present invention, in order to eliminate the above-mentioned problems of the conventional method, the accuracy of heat load prediction is improved by automatically inputting the latest meteorological data which is regularly announced from the weather forecast data supplier. The objective is to improve and unmanned the handling of arithmetic units.

[0006]

In order to solve the above problems, the invention of claim 1 stores an interface for fetching weather forecast data sent from a weather forecast data provider and storing the fetched data. A storage device for storing the data and a calculation device for calculating a predicted heat load value of the heat source device based on the taken-in data are provided.
According to the invention of claim 2, the heat source device is used for an air conditioning facility of a building. The invention of claim 3 is
Further, according to the heat load prediction value output by the arithmetic unit, while controlling the heat storage amount to the heat medium of the heat source device in advance and comparing the heat load prediction value and the amount of residual ice, the heat radiation amount from the heat medium is calculated. It is designed to be controlled.

[0007]

According to the invention of claim 1 as described above, the weather forecast data is sent from the provider, taken in by the interface, and stored in the storage device.
The calculation unit calculates the expected heat load value for the heat source device based on the captured data in the future (usually the next day). Therefore, the weather forecast data is on-line and in real time without human intervention. An accurate heat load prediction value that reflects the latest weather forecast data is input to the computing device and calculated.

According to the second aspect of the present invention, the heat source device is used for an air conditioning facility of a building, and the weather forecast data is sent from a provider to a heat load forecasting device belonging to each building, where the building is constructed. The heat load prediction value is calculated by adding the individual conditions of the product, and necessary measures for the heat source device are taken based on the calculated value. In the invention of claim 3, the heat storage amount to the heat medium is controlled in advance based on the heat load predicted value output from the arithmetic unit, and the heat load predicted value and the residual ice amount are compared and the heat medium is transferred from the heat medium. The heat radiation amount of is controlled.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIG. The heat load predicting device A efficiently operates the heat source device, but the heat source device to be targeted is an air conditioner for cooling and heating a building, an air conditioner for a relatively large vehicle, A warm water pool is possible.

In the following embodiments, a commercially available small general-purpose computer, a so-called personal computer, is used as the data storage device and the arithmetic unit among the elements constituting the present invention. Therefore, the data storage device is appropriately set in the memory of the personal computer body, the built-in hard disk, or an external storage medium such as a floppy disk. Further, the heat load prediction calculation device is stored as a program in a storage medium such as a hard disk, and is read and operated by the main body CPU when the heat load prediction device is operated.

The heat load predicting apparatus A of this embodiment comprises a personal computer main body 1 and a number of interfaces associated with the personal computer main body 1 for exchanging data with the outside. The first interface is an input interface 4 that is connected to the heat source device 2 and various detectors 3 installed in the environment around the heat source device 2 and captures actual data from these detectors 3. The data taken in here is stored in the performance data storage device that stores each of them according to each characteristic. In this embodiment, they are respectively stored in the operation record data storage device 5 in the actual operation of the heat source device 2 and the meteorological record data storage device 6 that stores the outside air temperature and the outside air humidity.

The second interface is an input device for automatically inputting the weather forecast data provided from the weather forecast data providing source 7 using the predetermined communication line L to the personal computer 1, and here is a binary value. A modem interface 8 for converting information into sound signals is adopted. The weather forecast data fetched by the modem 8 is stored in the weather forecast data storage device 9 in the storage medium in the personal computer 1. It should be noted that time / day of the week data is output from the calendar 10 incorporated in the personal computer 1, and this is associated with the above-mentioned operation record data, meteorological record data, forecast data, etc. Remembered.

A thermal load predicting / calculating device 12 is provided in the CPU of the personal computer. This computing device uses a neuro network, and the neural network learning means 13 for calculating the weighting coefficient of the neural network using the above-mentioned driving record data, meteorological record data and time / day data as teacher data, and Based on the weighting coefficient set as above, the heat load prediction is performed in two steps, including a neural network heat load prediction means 14 for calculating the heat load prediction using the operation result data, the meteorological result data, the time / day of the week data and the weather forecast data as factor data. It is designed to calculate values.

The personal computer 1 is also provided with a heat load prediction storage device 15 for storing the heat load prediction and an output device (interface) 16 for outputting the heat load prediction result in various formats. This output device 16
Is connected to a heat source device operation control arithmetic unit 17 that controls the operation of the heat source device through a predetermined line.

Next, the operation of the heat load predictive computing device A configured as described above will be described. The operation record data and the weather record data output from the detector 3 of the heat source device 2 are input to the personal computer 1 at regular intervals, for example, every one hour, and are stored in the respective storage devices 5 and 6. These actual result data are organized in association with the time / day of the week data from the calendar 10 built in the personal computer, and are sequentially accumulated. In the heat load prediction calculation device 12, the neural network learning means calculates the weighting coefficient by using these record accumulation data as teacher data.

On the other hand, the weather forecast data announced at a fixed time from the weather forecast provider 7 is automatically input to the personal computer 1 online via the line L and the modem interface 8 and stored in the weather forecast data storage device 9. Accumulated. Neural network heat load predicting means 14 using the weather forecast data and each of the above-mentioned performance data as factor data.
Thus, the expected heat load is calculated every hour, which is temporarily stored in the heat load prediction storage device, and is output to the heat source device control calculation device 17 by the output device 16.

Since the weather forecast data is automatically updated every time it is taken into the personal computer 1, the heat load prediction is always performed based on the latest information, and the prediction is highly accurate. In addition, since the weather forecast data is input through the line L and the modem 8 without human intervention, the forecast issued at a time outside the working hours can be captured without requiring overtime work, and input errors can be made. Instead, it always calculates quickly based on accurate data.

What is predicted in this way is usually the heat load of the next day in the air conditioning equipment of buildings, etc., and the heat source device 2 establishes the operation schedule of the next day based on such predicted values. , It can take the necessary advance measures quickly. These measures include, for example, heat storage using inexpensive electric power at night, securing fuel and a heat medium necessary for operation, switching valves of a cooling and heating device, and inspecting predetermined places.

In the above-mentioned embodiment, the neural network is adopted in the arithmetic unit, and the data inputted from each interface is constantly accumulated, and the weighting coefficient is corrected based on this, so that the learning effect is effective. A high predicted value of is obtained. Also, in the above example, the main component is a personal computer, and if you install a predetermined program on it, install a modem and connect it with a weather data provider, you can configure the forecasting device, so what condition user But it can be installed easily.

However, according to the present invention, a larger-scale engineering workstation (EWS) may be used instead of the above personal computer depending on the use and facility conditions.
Alternatively, an IC chip for storage and calculation may be incorporated on the board to form a more specialized and smaller device, which may be mounted on the controller of the heat source device. In addition, as a heat load prediction calculation method, methods other than neural networks (time series analysis model [ARIM
A EWMA model], fuzzy theory) is also possible.

[0021]

As described above, according to the present invention,
The latest weather forecast data announced by a weather forecast provider as an input element of the heat load forecast is automatically captured online, for example, using a communication line, so that the latest weather data can be quickly reflected to predict the heat load. In addition to improving the accuracy of the above, unmanned handling of the predictive calculation device can be achieved, further improving the accuracy by preventing input mistakes, and eliminating the need for overtime work.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a heat load prediction calculation device according to an embodiment of the present invention.

[Explanation of symbols]

 2 heat source device 5, 6 result data storage device 7 weather forecast provider 8 modem interface 9 weather forecast data storage device 12 heat load prediction calculation device 16 output device 17 heat source operation control calculation device

Front page continuation (72) Inventor Tsuneo Uekusa 1-34, Roppongi, Minato-ku, Tokyo Inside NTT Ft. Inside the EBARA CORPORATION (72) Inventor Mitsuteru Furuya 11-11 Haneda-Asahi-cho, Ota-ku, Tokyo Inside the EBARA CORPORATION

Claims (3)

[Claims] 1. An interface for fetching weather forecast data sent from a weather forecast data provider, a storage device for storing the fetched data, and a future heat source device based on the fetched data. A heat load prediction device comprising: a calculation device that calculates a heat load prediction value. 2. The heat load prediction device according to claim 1, wherein the heat source device is a heat source device used for an air conditioning facility of a building. 3. Further, the amount of heat stored in the heat medium of the heat source device is controlled in advance according to the predicted heat load value output from the arithmetic unit, and the predicted heat load value is compared with the residual ice amount to determine the heat load. The heat load predicting apparatus according to claim 1, wherein the heat radiation amount from the medium is controlled.