KR0134845B1 - Apparatus and method for controlling p.m.v in airconditioner - Google Patents

Abstract

The present invention relates to a field learning control method and an apparatus of the air conditioner. In particular, in the absence of an external input, that is, an input such as a cold / hot key (5), a wind direction key (6), and an activity change key (7). By controlling the preset PMV value by the 3rd, 5th, 6th, and initial offline learning, if any one of the cold / hot key, the wind direction key, and the activity change key is generated, the fourth step of FIG. As the process, the value input by online learning is recalculated by the PMV equation, and the PMV table value is changed through the field learning algorithm, and the changed value is used as a control value which becomes a new PMV = 0. By learning PMV values, which are comfortable indicators, by controlling the living habits, it is possible to increase the satisfaction of the individual's comfort rather than simply learning the set temperature.

Description

Field learning control method of air conditioner and its device

1 is a block diagram of an apparatus of the present invention.

2 is a flowchart for explaining the method of the present invention.

3 is a diagram showing a field trip neural network.

* Description of the symbols for the main parts of the drawings *

1: Current temperature detector 2: Set temperature unit

3: elapsed time calculation unit 4: field learning unit

5: hot / cold key 6: wind direction key

7: Activity change key 8: Temperature difference setting unit

9: position setting unit 10: wind direction setting unit

11: air flow rate setting unit 12: PMV table

13: Compressor 14: Flap

15: Fan 16: New set temperature controller

17: wind direction control unit 18: air volume control unit

The present invention relates to a field learning control method and an apparatus of the air conditioner, in particular, to learn the predicted mean vote (PMV) value as a factor generated in the field to perform the learning control for individual air conditioning. It relates to a field learning control method and an apparatus of an air conditioner.

Here, the predicted comfort index which is PMV is a numerical value which shows the degree to which most people feel comfortable, and is a state which PMV = 0 feels comfortable.

In the conventional air conditioning, most of the learning methods in the air conditioner adopt a method of performing learning using an algorithm by neural networks as input and output elements in advance by offline field trips. -Line) Even when learning, there was a problem that the learning factor was limited to the set temperature.

In this case, online learning is not based on prior learning, but learning based on values generated in the field.

An object of the present invention is to control the reflection of the taste and lifestyle of the individual, in particular, learning the PMV value of the comfortable indicators, not the learning of a simple setting temperature and the like, the on-site learning control method of the air conditioning that can increase the satisfaction of the individual comfort and its To provide a device.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

As shown in FIG. 1, the field learning unit receives predetermined output signals from the present temperature detecting unit 1, the set temperature unit 2, and the elapsed time calculating unit 3, respectively, and performs field learning. 4) and; Output of the cover / cold key (5), wind direction key (6), active amount change key (7), temperature difference setting unit (8), position setting unit (9), wind direction setting unit (10), and wind direction setting unit (11) A PMV table 12 configured by PMV values calculated by receiving signals, respectively; The new set temperature control unit 16 and the wind direction control unit 17 which control the compressor 13, the flap 14 and the fan 15 by calculating the new set temperature, the wind direction and the air volume by the above factors. ) And the air volume adjusting unit 18 is characterized in that.

According to the method of the present invention, as shown in FIG. 2, when the cold / hot key 7, the wind direction key 6, and the activity change key 7 are all in the off state, a method for installing the PMV table configured by the offline learning is provided. 1 course; If the cold / hot key (5), the wind direction key (6), and the activity change key (7) are all on, the PMV table value obtained by the offline line is changed by the PMV equation. The second process of making it work; A third process of reading an output value corresponding to an input from a PMV table when the first process and the second process are completed; The control variable is characterized in that the fourth process of controlling the frequency of the compressor 13, the number of revolutions and the wind direction of the fan (15).

Referring to the effects of the present invention as follows.

First, briefly explaining the purpose of the field trip to help the understanding of the present invention, by learning the PMV data by the control method by the offline learning, that is, not the field trip, for example, what kind of environment (temperature, humidity, wind speed, This refers to controlling the comfort status of the occupants (radiation temperature, activity amount, and amount of clothing) to reach a comfortable condition such that PMV = 0 is achieved by using the PMV index.

In this process, to derive the PMV value for the input of the environmental conditions, a learning method called a BP (Back Propagation) algorithm, which is an algorithm that is trained to follow the reference value by feeding back deviation values of the learning information, is adopted.

However, such PMV control indicates that most people are satisfied with a certain environment, as most comfort indicators do.

As they may be dissatisfied with the current situation due to personal preferences or physiological differences, they have a habit of living with an air conditioner with individual dissatisfaction.

Therefore, in the present invention, it is possible to deviate from such a standard to make a comfortable state of each individual occupant, as well as to learn the PMV value itself in order to reduce the inconvenience caused by the user's own repetition of operation. If there is no input of the rudder key (5), the wind direction key (6), and the activity change key (7), the PMV value preset by the initial offline learning as shown in the third, fifth and sixth steps of FIG. If any one of the cold / hot keys (5), the wind direction key (6), and the activity change key (7) occurs, the value input by the online learning as shown in the fourth process of FIG. After recalculating by PMV equation, PMV table value is changed by field learning algorithm, and the changed value is used as the control value to make new PMV = 0.

At this time, the value of the PMV table cannot be changed to 1: 1 every time a key is input, so the input values are learned red.

In addition, the above-described online learning refers to a method of changing a relative value at which a PMV = 0 value is detected by detecting an environmental state at that time together with the control signal when a control signal is input.

In addition, PMV formula mentioned above is as follows.

S = M ± W ± R ± C ± K-E-Res (Kcal / h)

Where M and W are human activity and wind speed, R is radiation temperature, C is temperature, E is humidity, and Res is the amount of clothing.

3 shows Neural Nrtwork, whose algorithm is as follows.

F (Σ (wi-xi) ² -T

Where w is Weight, x is Input and T is Threshold.

The output obtained by the algorithm changes the actual PMV table value.

As described above, according to the present invention, by learning the PMV value, which is a comfort index, by controlling the preferences and lifestyles of the individual, the user's comfort can be improved rather than a simple set temperature learning.

Claims (2)

A first step of searching the prediction comfort index (PMV) table made by offline learning if the cold / hot key, the wind direction key, and the activity change key are all off; A second process of changing a PMV table value obtained by the offline by the PMV equation as an input element change for on-line learning when the cold / hot key, the wind direction key, and the activity change key are all on; A third process of reading an output value corresponding to an input from a PMV table when the first process and the second process are completed; Field learning control method of the air conditioning, characterized in that consisting of a fourth process of controlling the frequency of the compressor, the number of revolutions and the wind direction of the compressor as a control variable. A field learning unit for receiving a predetermined output signal from the current temperature detecting unit, the set temperature unit, and the elapsed time calculating unit, respectively, and performing field learning; A table composed of PMV calculation values calculated by receiving the output signals of the cold / hot key, the wind direction key, the activity change key, the temperature difference setting unit, the wind direction setting unit, and the air volume setting unit, respectively; On-site learning control device of the air conditioner comprising a new set temperature control unit, a wind direction control unit and a flow rate control unit for controlling the compressor, the flap and the fan by calculating a new set temperature, air volume and air volume by the above factors. .