TWI746087B - Air conditioning system control method - Google Patents

Abstract

An air conditioning system control method is used to control the air conditioning set temperature of the air conditioning system. It is implemented by an air conditioning control device and includes the following steps: (A) According to the possible set temperature of multiple air conditioners and at least one indoor temperature sensing value, use one The indoor temperature prediction model used to predict the indoor temperature after a preset time, to obtain a plurality of predicted indoor temperature groups corresponding to the possible set temperatures of the air conditioners and a predicted time; (B) according to a set of expected indoor temperatures Obtain a target expected indoor temperature from the expected indoor temperature data and the predicted time; and (C) according to a plurality of error values related to the predicted indoor temperature groups and the target expected indoor temperature , Obtain a target air conditioner set temperature from the possible set temperatures of the air conditioners.

Description

Air conditioning system control method

The present invention relates to a control method, in particular to a control method of an air conditioning system.

With the rapid development of science and technology in recent years, the pursuit of a comfortable living environment has also been gradually valued by people. After all, most people spend most of their time indoors. Thanks to the advancement and development of air-conditioning systems, the adjustment and control of the indoor environment are more convenient. While enjoying warm winter and cool summer and excellent air quality, it can greatly improve work efficiency and maintain healthy physiological functions.

Traditionally, the control of the air-conditioning system is based on indoor comfort indicators such as Predicted Mean Vote (PMV), and the control goals are set in conjunction with energy-saving strategies. The control personnel then set the control goals based on the accumulation of experience and inheritance. control. However, the control effect varies from person to person, resulting in the control of the existing air-conditioning system, unable to reach the target indoor temperature all the time, resulting in the inability to instantly respond to the feelings of the people moving in the space.

Therefore, the object of the present invention is to provide an air conditioning system control method that can automatically control the air conditioning system.

Therefore, the air-conditioning system control method of the present invention is used to control the air-conditioning set temperature of the air-conditioning system, which is implemented by an air-conditioning control device that is electrically connected to the air-conditioning system and at least one indoor temperature sensor set in an indoor space The at least one indoor temperature sensor is used to sense the indoor temperature to generate at least one pair of indoor temperature sensing values corresponding to the at least one indoor temperature sensor. The indoor temperature data includes a plurality of first time periods and a plurality of desired indoor temperatures corresponding to the first time periods, including a step (A), a step (B), and a step (C).

In the step (A), the air-conditioning control device uses an indoor temperature prediction model for predicting the indoor temperature after a preset time according to the possible set temperatures of the air-conditioners and the at least one indoor temperature sensing value to obtain a plurality of The predicted indoor temperature groups corresponding to the possible set temperatures of the air conditioners and a predicted time, wherein the predicted time is a current time plus the preset time, and each predicted indoor temperature group includes at least one corresponding to at least one indoor temperature The predicted indoor temperature of the sensed value.

In the step (B), the air-conditioning control device obtains a target expected indoor temperature from the expected indoor temperatures according to the expected indoor temperature data and the predicted time, wherein the predicted time is located at the second corresponding to the target expected indoor temperature. A period of time

In the step (C), the air-conditioning control device obtains a target air-conditioning set temperature from the possible set temperatures of the air-conditioners according to a plurality of error values related to the predicted indoor temperature groups and the target desired indoor temperature.

The effect of the present invention is that the air-conditioning control device predicts that the air conditioners may set the temperature to cause the predicted indoor temperature groups after the preset time, and based on the error between the predicted indoor temperature groups and the target expected indoor temperature Value, the target air-conditioning setting temperature is obtained from the possible setting temperatures of the air-conditioners, so as to reach the target desired indoor temperature in each time period, and to reflect the feelings of the active personnel in the space in real time.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are denoted by the same numbers.

Referring to FIG. 1, an air-conditioning system 11 and an air-conditioning control device 12 electrically connected to the air-conditioning system 11 are described to implement an embodiment of the air-conditioning system control method of the present invention.

The air conditioning system 11 includes an Air Handling Unit (AHU) 111, a plurality of Variable Air Volume System (VAV) 112, a plurality of small air coolers (Fan Coil Unit, FCU) 113, and an external Make-Up Air Unit (MAU) 114. It is worth noting that in this embodiment, the air conditioning system 11 includes, for example, 5 variable air volume systems 112 and 5 small air coolers 113, and these are arranged in the window area. In other embodiments, the variable air volume system 112 The number of the air cooler 113 and the small air cooler 113 can also be 1, but is not limited to this.

The air-conditioning control device 12 is also electrically connected to a plurality of indoor temperature sensors 13 arranged in an indoor space, a plurality of air-conditioning sensors 14 arranged in the air-conditioning system 11, a plurality of socket power consumption sensors 15, and A sunny rain sensor 16. It is worth noting that, in this embodiment, for example, there are 5 indoor temperature sensors 13, 28 air-conditioning sensors 14, and 2 socket power consumption sensors 15. In other embodiments, indoor temperature sensors The number of 13 and the socket power consumption sensor 15 can also be 1, but is not limited to this.

The indoor temperature sensors 13 are used for sensing indoor temperature to generate a plurality of indoor temperature sensing values corresponding to the indoor temperature sensors 13. It is worth noting that, in this embodiment, the indoor temperature sensors 13 are respectively disposed in the variable air volume systems 112 to sense the return air temperature of the variable air volume systems 112 as the indoor temperature Indoor temperature at 5 different locations in the space, but not limited to this.

The air-conditioning sensors 14 are used to sense the state of the air-conditioning system 11 to generate multiple pieces of air-conditioning system state sensing data. It is worth noting that in this embodiment, there are 28 air-conditioning system state sensing data corresponding to 28 air-conditioning sensors 14, and one of the 28 air-conditioning sensors 14 is set in the air-conditioning box 111. It is used to sense the pressure in the sealed pipe of the air conditioning box 111; 5 are respectively set in the variable air volume systems 112 to respectively sense whether the variable air volume systems 112 enter the energy-saving mode; 5 are respectively set in the variable air volume systems 112 The constant variable air volume system 112 is used to respectively sense the opening of the air volume damping of the variable air volume systems 112; five are respectively arranged on the small air coolers 113 to respectively sense whether the small air coolers 113 are activated; 5 One is respectively installed on the small air coolers 113 to sense the opening of the ice water valve of the small air coolers 113; 5 are respectively installed on the small air coolers 113 to sense the small cold air The return air temperature of the machine 113; one is installed in the outside air conditioning box 114 to sense the temperature of the output gas of the outside air conditioning box 114; one is installed in the outside air conditioning box 114 to sense the The temperature of the inhaled outside air of the outside air conditioning box 114, but not limited to this.

The socket power consumption sensors 15 are used for sensing the power supply of the sockets in the indoor space to generate a plurality of socket power consumption data corresponding to the socket power consumption sensors 15 respectively.

The sunny and rainy sensor 16 is used to sense whether it is raining or not, so as to generate sunny and rainy sensing data indicating whether it is raining.

The air conditioning control device 12 stores multiple training data, a desired indoor temperature data, multiple synthetic air conditioning setting temperatures, a first controllable coefficient, a second controllable coefficient, a threshold value, a first adjustable parameter, and a The second adjustable parameter, the possible setting temperature of multiple air conditioners, and an expert recommends the setting temperature data of the air conditioner.

Each training data includes a historical air-conditioning set temperature, a plurality of first historical indoor temperatures, a plurality of historical air-conditioning system state sensing data related to the air-conditioning system 11, and a plurality of historical air-conditioning system state sensing data related to the indoor space. Socket power consumption data, a historical sunny and rainy sensor data, and a plurality of second historical indoor temperatures corresponding to the first historical indoor temperatures, and the second historical indoor temperatures have passed through a period after the first historical indoor temperatures are generated The preset time is generated. It is worth noting that, in this embodiment, the preset time is, for example, 18 minutes. In other embodiments, the number of the first historical indoor temperature, historical socket power consumption data, and the second historical indoor temperature may be 1 respectively. In addition, in other embodiments, each training data may only include the historical air conditioner setting temperature, the first historical indoor temperature, and the second historical indoor temperature, but it is not limited to this.

The expected indoor temperature data includes a plurality of first time periods and a plurality of expected indoor temperatures respectively corresponding to the first time periods.

The first controllable coefficient and the second controllable coefficient are greater than 0 and less than or equal to 1, and the first controllable coefficient is less than the second controllable coefficient.

The first adjustable parameter is greater than zero and less than 1, and the second adjustable parameter is any positive number.

The air conditioner may be set to a temperature of 20 to 27 degrees Celsius, for example, but it is not limited to this.

The expert recommends that the air conditioner setting temperature data includes a plurality of second time periods and a plurality of recommended air conditioner setting temperatures respectively corresponding to the second time periods.

Referring to FIGS. 1 and 2, the embodiment of the control method of the air conditioning system of the present invention includes a modeling program and a control program. The modeling program is described below.

In step 21, for each training data, the air-conditioning control device 12 performs data synthesis based on the training data, the synthetic air-conditioning setting temperatures, the first controllable coefficient, and the second controllable coefficient to obtain multiple The pens respectively correspond to the synthetic training data of the synthetic air conditioner setting temperature.

With reference to FIG. 3, step 21 includes sub-steps 211 to 213, and the sub-steps included in step 21 are described below.

In step 211, for each training data, and for each synthetic air-conditioning setting temperature of which the historical air-conditioning setting temperature minus the training data is greater than zero, the air-conditioning control device 12 is based on the first controllable coefficient and the synthetic air-conditioning setting The temperature, the historical air-conditioning setting temperature and the second historical indoor temperatures of the training data are used to obtain a plurality of synthetic second historical indoor temperatures corresponding to the second historical indoor temperatures.

In step 212, for each training data, and for each synthetic air-conditioning setting temperature of which the historical air-conditioning setting temperature minus the training data is less than zero, the air-conditioning control device 12 according to the second controllable coefficient and the synthetic air-conditioning setting The temperature, the historical air-conditioning setting temperature and the second historical indoor temperatures of the training data are used to obtain a plurality of synthetic second historical indoor temperatures corresponding to the second historical indoor temperatures.

以下式表示：

， 其中，

為該等第二歷史室內溫度，

為該第一可控係數，

，

為該合成空調設定溫度，

為該歷史空調設定溫度，

為該第二可控係數，

。 It is worth noting that in this embodiment, the synthesis of the second historical indoor temperature in step 211 or 212

The following formula represents:

, in,

Is the second historical indoor temperature,

Is the first controllable coefficient,

,

Set the temperature for the synthetic air conditioner,

Set the temperature for this historical air conditioner,

Is the second controllable coefficient,

.

In step 213, for each training data and for each synthetic air conditioner setting temperature, the air conditioner control device 12 generates a synthetic training data. The synthetic training data includes the synthetic air conditioner setting temperature and the first training data. The historical indoor temperature, the air conditioning system state sensing data of the training data, the historical socket power consumption data of the training data, the historical sunny and rainy sensing data of the training data, and the synthesized second data obtained in step 211 or 212 Historical indoor temperature.

It should be particularly noted that, in this embodiment, the first controllable coefficient is, for example, 0.6, to simulate the characteristics of faster cooling of active cooling, and the second controllable coefficient is, for example, 0.4 to simulate the characteristics of passive heating. Slow characteristics, but not limited to this.

In step 22, for each training data and synthetic training data, the air-conditioning control device 12 uses the training data or the synthetic training data to sense the state of the air-conditioning system, the historical socket power consumption data, and the historical sunny or rainy weather. Principal Components Analysis (PCA) is performed on the sensing data to obtain multiple training analysis data. It is worth noting that, in this embodiment, the air-conditioning control device 12 performs principal component analysis on 28 air-conditioning system state sensing data, 2 historical socket power consumption data, and 1 historical sunny and rainy sensing data, and re-debugging Save the Jing, eliminate redundant parts to obtain 13 analysis data, but not limited to this.

In step 23, the air-conditioning control device 12 uses the historical air-conditioning setting temperature, the first historical indoor temperature, and the second historical indoor temperature of the training data, and the corresponding training analysis data, and the synthetic air conditioning of the synthetic training data Set the temperature, the first historical indoor temperature, synthesize the second historical indoor temperature, and the corresponding training analysis data, and use a Neural Network (NN) algorithm to create a predictive indoor temperature after the preset time Indoor temperature prediction model.

It should be noted again that in this embodiment, the number of training data samples is insufficient, which causes the training data to be biased, and the trained model is not accurate. Therefore, step 21 is required to solve the problem by data synthesis. The problem of data deviation, but in other implementations where the number of training data samples is sufficient, steps 22 and 23 can be directly performed without step 21. In addition, in other embodiments, the principal component analysis of step 22 can also be omitted, and in step 23 , It is not limited to establish the indoor temperature prediction model directly based on the training data and the synthetic training data, or only based on the training data.

Referring to Figures 1 and 4, the control procedure is described below.

In step 31, the air-conditioning control device 12 performs principal component analysis on the air-conditioning system state sensing data, the socket power consumption data, and the sunny and rainy sensing data to obtain multiple pieces of analysis data.

In step 32, for each possible setting temperature of the air conditioner, the air conditioning control device 12 uses the indoor temperature prediction model to obtain a corresponding response based on the possible setting temperature of the air conditioner, the indoor temperature sensing values, and the analysis data. The air conditioner may set the temperature and predict the indoor temperature group at a predicted time. The predicted time is a current time plus the preset time, and each predicted indoor temperature group includes a plurality of predicted indoor temperatures corresponding to the indoor temperature sensing values.

In other embodiments, step 32 can obtain the predicted indoor temperature group only according to the possible set temperature of the air conditioner and the indoor temperature sensing values, and step 31 is not required, and step 32 is performed directly, but it is not limited to this. .

In step 33, the air-conditioning control device 12 obtains a target expected indoor temperature from the expected indoor temperatures according to the expected indoor temperature data and the predicted time, where the predicted time is located at the first corresponding to the target expected indoor temperature. Time period. For example, the first time period is a 24-hour time period in a day. If the predicted time is 09:15, then the target expected indoor temperature corresponds to the first time period when the target expected indoor temperature is 09:00~09:59. temperature.

In step 34, the air-conditioning control device 12 obtains a target air-conditioning setting temperature from the possible setting temperatures of the air-conditioning based on a plurality of error values related to the predicted indoor temperature groups and the target expected indoor temperature.

Referring to FIG. 5 in conjunction, step 34 includes sub-steps 341 to 346. The sub-steps included in step 34 are described below.

以下式表示：

， 其中，

為該空調可能設定溫度對應的預測室內溫度組，

為該室內溫度預測模型，

為該等室內溫度感測值，

為該等分析資料，

為該等空調系統狀態感測資料、該等插座功耗資料，及該晴雨感測資料，

為該空調可能設定溫度，

為該目標期望室內溫度，

為該預設時間，

為該預測時間，

為範數函數。要特別注意是，在可僅根據該空調可能設定溫度及該等室內溫度感測值，獲得該預測室內溫度組的實施方式中，該空調可能設定溫度對應的預測室內溫度組以

表示。 In step 341, for each possible set temperature of the air conditioner, the air conditioner control device 12 obtains an error value according to the predicted indoor temperature group corresponding to the possible set temperature of the air conditioner and the target expected indoor temperature, the error value

The following formula represents:

, in,

The predicted indoor temperature group corresponding to the possible set temperature of the air conditioner,

Is the indoor temperature prediction model,

Are the indoor temperature sensing values,

For such analysis data,

For the air-conditioning system state sensing data, the socket power consumption data, and the sunny or rainy sensing data,

May set the temperature for this air conditioner,

The desired room temperature for this goal,

For the preset time,

For the forecast time,

Is the norm function. It should be particularly noted that in the embodiment in which the predicted indoor temperature group can be obtained only based on the possible set temperature of the air conditioner and the indoor temperature sensing values, the predicted indoor temperature set corresponding to the possible set temperature of the air conditioner is

Express.

In step 342, the air-conditioning control device 12 obtains a possible set temperature of the first candidate air-conditioner that should have the smallest error value among the equal error values.

In step 343, the air-conditioning control device 12 obtains a pair of second candidate air-conditioning possible set temperatures corresponding to the second smallest error value of one of the equal error values.

In step 344, the air-conditioning control device 12 determines whether the possible set temperature of the second candidate air conditioner is greater than the possible set temperature of the first candidate air conditioner, and the second small error value is less than the threshold value. When it is determined that the possible set temperature of the second candidate air conditioner is greater than the possible set temperature of the first candidate air conditioner, and the second small error value is less than the threshold value, the process proceeds to step 345; and when it is determined that the second candidate air conditioner may be set When the temperature is not greater than the possible set temperature of the first candidate air conditioner, or the second small error value is not less than the threshold value, the process proceeds to step 346.

In step 345, the air-conditioning control device 12 determines that the second candidate possible air-conditioner set temperature is the target air-conditioner set temperature.

In step 346, the air conditioner control device 12 determines that the first candidate air conditioner possible set temperature is the target air conditioner set temperature.

It is worth noting that, in this embodiment, if the possible set temperature of the second candidate air conditioner is greater than the possible set temperature of the first candidate air conditioner, and the second small error value is still within an acceptable range (that is, less than the threshold value), The second candidate air conditioner may be determined to be the target air conditioner set temperature to save energy. However, in other embodiments, only steps 341 and 342 may be performed, and the first candidate air conditioner may be directly determined to be the target air conditioner. The setting temperature of the air conditioner is not limited to this.

以下式表示：

， 其中，

為該第一可調參數，

為該第二可調參數，

為該室內溫度預測模型，

為該等前一室內溫度感測值，

為該等前一分析資料，

為該前一空調設定溫度，

為該至少一室內溫度感測值，

為前一時間，

為範數函數，

為該前一信心度，且

。 In step 35, the air-conditioning control device 12 is based on a previous air-conditioning setting temperature set at a previous time, a plurality of previous analysis data obtained at the previous time, and a plurality of indoor temperature sensors. 13 The previous indoor temperature sensing value generated at the previous time, the indoor temperature sensing values, the first adjustable parameter, the second adjustable parameter, and a previous confidence level at the previous time , Obtain a current confidence level, the previous time is the current time minus the preset time, the previous confidence level and the current confidence level are greater than zero and less than 1, the current confidence level

The following formula represents:

, in,

Is the first adjustable parameter,

Is the second adjustable parameter,

Is the indoor temperature prediction model,

Are the previous indoor temperature sensing values,

For the previous analysis data,

Set the temperature for the previous air conditioner,

Is the at least one indoor temperature sensing value,

For the previous time,

Is the norm function,

Is the previous confidence level, and

.

In step 36, the air-conditioning control device 12 obtains a target recommended air-conditioning setting temperature from the recommended air-conditioning setting temperature according to the expert-recommended air-conditioning setting temperature data and the previous time, wherein the previous time is located in the target recommended air-conditioning setting temperature. The second time period corresponding to the set temperature of the air conditioner.

以下式表示：

， 其中，

為該目標空調設定溫度，

為該目標建議空調設定溫度。 In step 37, the air-conditioning control device 12 obtains an optimal air-conditioning setting temperature based on the current confidence level, the target recommended air-conditioning setting temperature, and the target air-conditioning setting temperature.

The following formula represents:

, in,

Set the temperature for the target air conditioner,

The air conditioner setting temperature is recommended for this goal.

In step 37, the air-conditioning control device 12 sets the air-conditioning system 11 according to the optimal air-conditioning set temperature.

It should be particularly noted that in this embodiment, because the initial stage simulation model uses synthetic data training, the accuracy is not perfect. Therefore, proceed to step 35 to calculate the confidence level of the indoor temperature prediction model, and follow the Current confidence, the target air-conditioning setting temperature and the target recommended air-conditioning setting temperature to adjust the optimal air-conditioning setting temperature, and when the accuracy of the indoor temperature prediction model is not credible, the target recommended air-conditioning setting temperature can be increased The proportion at least provides the user's basic comfort level. In other embodiments, steps 35 to 37 may not be performed, and the target air conditioner setting temperature may be directly used for control, which is not limited to this.

In summary, the air-conditioning system control method of the present invention uses the air-conditioning control device 12 to perform principal component analysis on the air-conditioning system state sensing data, the socket power consumption data, and the sunny and rainy sensing data to obtain the According to the possible set temperature of the air conditioners, the analysis data, and the indoor temperature sensing values, the indoor temperature prediction model is used to obtain the predicted indoor temperature groups, and then based on the predicted indoor temperature groups. The error value of the target expected indoor temperature is obtained from the possible setting temperatures of the air conditioners, and the current confidence level is calculated based on the current confidence level, the target air conditioner setting temperature, and the target recommendation The air-conditioning set temperature is obtained, and the air-conditioning system 11 is finally set according to the optimal air-conditioning set temperature to achieve the target desired indoor temperature in each time period, and it can reflect the feelings of the active personnel in the space in real time. To achieve the purpose of the invention.

However, the above are only examples of the present invention. When the scope of implementation of the present invention cannot be limited by this, all simple equivalent changes and modifications made in accordance with the scope of the patent application of the present invention and the content of the patent specification still belong to Within the scope covered by the patent of the present invention.

11: Air conditioning system

111: Air conditioning box

112: Variable air volume system

113: small air cooler

114: outside air conditioning box

12: Air conditioning control device

13: Indoor temperature sensor

14: Air conditioner sensor

15: Socket power consumption sensor

16: Sunny rain sensor

21~23: Steps

211~213: Steps

31~37: Steps

341~346: Steps

Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, in which: 1 is a block diagram illustrating the connection relationship of an air conditioning control device used to implement an embodiment of the air conditioning system control method of the present invention; Figure 2 is a flowchart illustrating a modeling procedure of this embodiment; FIG. 3 is a flowchart to assist in explaining the sub-steps of step 22 in FIG. 2; Figure 4 is a flowchart illustrating a control procedure of this embodiment; and FIG. 5 is a flowchart to assist in explaining the sub-steps of a step 34 in FIG. 4.

31~37: Steps

Claims (8)

An air conditioning system control method for controlling the air conditioning set temperature of the air conditioning system is implemented by an air conditioning control device, the air conditioning control device is electrically connected to the air conditioning system, at least one indoor temperature sensor arranged in an indoor space, and more An air-conditioning sensor, at least one socket power consumption sensor, and a rain or rain sensor are provided in the air-conditioning system. The at least one indoor temperature sensor is used to sense the indoor temperature to generate at least one pair of at least The indoor temperature sensing value of an indoor temperature sensor, the air-conditioning control device stores a plurality of possible setting temperatures of the air conditioner and a desired indoor temperature data, and the desired indoor temperature data includes a plurality of first time periods and a plurality of corresponding first A period of expected indoor temperature, the air-conditioning sensors are used to sense the state of the air-conditioning system to generate multiple pieces of air-conditioning system state sensing data, and the at least one socket power consumption sensor is used to sense the indoor space The power supply of the socket to generate at least one socket power consumption data corresponding to the at least one socket power consumption sensor respectively, and the sunny/rain sensor is used to sense whether it is raining to generate a sunny/rainy sensor indicating whether it is raining Data, the method includes the following steps: (A) according to the possible set temperature of the air conditioners, the at least one indoor temperature sensing value, and the air conditioning system state sensing data, the at least one socket power consumption data, and the sunny/rainy sensor Obtain a plurality of predicted indoor temperature groups corresponding to the possible set temperatures of the air conditioners and a predicted time by using an indoor temperature prediction model for predicting the indoor temperature after a preset time, wherein the predicted time is one The current time plus the preset time, each predicted indoor temperature group includes at least one predicted indoor temperature corresponding to at least one indoor temperature sensing value, step (A) includes the following sub-steps: (A-1) Perform principal component analysis on the air conditioning system state sensing data, the at least one socket power consumption data, and the sunny or rainy sensing data to obtain multiple pieces of analysis data, and (A-2) for each The possible setting temperature of the air conditioner, according to the possible setting temperature of the air conditioner, the at least one indoor temperature sensing value, and the analysis data, the indoor temperature prediction model is used to obtain a prediction room corresponding to the possible setting temperature of the air conditioner and at the predicted time Temperature group; (B) according to the expected indoor temperature data and the predicted time, obtain a target expected indoor temperature from the expected indoor temperatures, wherein the predicted time is in the first time period corresponding to the target expected indoor temperature; and ( C) According to a plurality of error values related to the predicted indoor temperature groups and the target expected indoor temperature, a target air-conditioning set temperature is obtained from the possible set temperatures of the air-conditioning. According to the air-conditioning system control method of claim 1, the air-conditioning control device further stores a plurality of training data, each training data includes a historical air-conditioning set temperature, at least one first historical indoor temperature, and at least one corresponding to at least one The second historical indoor temperature of the first historical indoor temperature, the at least one second historical indoor temperature is generated after the at least one first historical indoor temperature is generated after the preset time, before step (A), the following steps are further included: ( D) Based on the training data, a neural network algorithm is used to establish the indoor temperature prediction model. According to the air-conditioning system control method of claim 2, the air-conditioning control device further stores a plurality of synthetic air-conditioning set temperatures, a first controllable coefficient, and a second controllable coefficient, the first controllable coefficient and the second controllable coefficient Controllable coefficient is greater than 0 and small Is equal to 1, and the first controllable coefficient is less than the second controllable coefficient, wherein step (D) includes the following sub-steps: (D-1) for each training data, according to the training data, the synthetic air conditioners Set the temperature, the first controllable coefficient, and the second controllable coefficient, and perform data synthesis to obtain a plurality of synthetic training data corresponding to the synthetic air conditioner setting temperature; and (D-2) according to the training data And the synthetic training data, using the neural network algorithm to establish the indoor temperature prediction model. The air conditioning system control method according to claim 3, wherein step (D-1) includes the following sub-steps: (D-1-1) for each training data, and for each historical air conditioning minus the training data Set a synthetic air conditioner set temperature with a set temperature greater than zero, and obtain at least one corresponding respectively according to the first controllable coefficient, the synthetic air conditioner set temperature, and the historical air conditioner set temperature and the at least one second historical indoor temperature of the training data A synthetic second historical indoor temperature of at least one second historical indoor temperature; (D-1-2) For each training data, and for each synthetic air-conditioning setting temperature of which the historical air-conditioning setting temperature minus the training data is less than zero, According to the second controllable coefficient, the synthetic air conditioner setting temperature, and the historical air conditioner setting temperature of the training data and the at least one second historical indoor temperature, at least one synthetic first corresponding to at least one second historical indoor temperature is obtained. 2. Historical indoor temperature; and (D-1-3) For each training data and for each synthetic air-conditioning set temperature, generate a set temperature that includes the synthetic air-conditioning set temperature and the training data The at least one first historical indoor temperature, and the synthetic training data obtained in step (D-1-1) or (D-1-2) to synthesize the second historical indoor temperature. According to the air-conditioning system control method of claim 1, the air-conditioning control device also stores a threshold value, wherein step (C) includes the following sub-steps: (C-1) The temperature may be set for each air-conditioner, according to the air-conditioning possibility Obtain an error value between the predicted indoor temperature group corresponding to the set temperature and the target expected indoor temperature; (C-2) Obtain the possible set temperature of the first candidate air conditioner that should have the smallest error value among the equal error values; (C-3) ) Obtain a pair of possible second candidate air conditioner set temperatures that should be equal to one of the second smallest error values; (C-4) determine whether the second candidate air conditioner may set temperature is greater than the first candidate air conditioner's possible set temperature, And the second small error value is less than the threshold value; (C-5) When it is determined that the possible setting temperature of the second candidate air conditioner is greater than the possible setting temperature of the first candidate air conditioner, and the second small error value is less than the threshold value , Determine the possible setting temperature of the second candidate air conditioner as the target air conditioner setting temperature; and (C-6) When it is determined that the possible setting temperature of the second candidate air conditioner is not greater than the possible setting temperature of the first candidate air conditioner, or the second lower When the error value is not less than the threshold value, it is determined that the possible set temperature of the first candidate air conditioner is the target air conditioner set temperature. The air conditioning system control method according to claim 5, wherein, in step (C-1), the error value

The following formula represents:

_{Wherein, F (s t, a t} ) may be set as the air temperature corresponding to the indoor temperature prediction group, F (.) For the indoor temperature prediction model, s _t room temperature for at least a sensing value, a _t for The air conditioner may set the temperature,

Is the desired indoor temperature for the target, t + △ T is the predicted time, and ∥.∥ is the norm function. According to the air-conditioning system control method of claim 1, the air-conditioning control device further stores a first adjustable parameter, a second adjustable parameter, and an expert recommended air-conditioning setting temperature data, the first adjustable parameter is greater than zero and Less than 1, the second adjustable parameter is any positive number. The expert recommends that the air-conditioning setting temperature data includes a plurality of second time periods and a plurality of recommended air-conditioning setting temperatures corresponding to the second time periods. After step (C), it also includes The following steps: (E) According to a previous air conditioning set temperature set at a previous time, at least one previous indoor temperature sensing value generated by the at least one indoor temperature sensor at the previous time, the at least An indoor temperature sensing value, the first adjustable parameter, the second adjustable parameter, and a previous confidence level at the previous time to obtain a current confidence level, where the previous time is the current time minus The preset time, the previous confidence level and the current confidence level are greater than zero and less than 1; (F) According to the expert’s recommended air conditioning setting temperature data and the previous time, obtain a target from the recommended air conditioning setting temperatures Recommended air-conditioning setting temperature, where the previous time is in the second time period corresponding to the target recommended air-conditioning setting temperature; and (G) According to the current confidence, the target recommended air-conditioning setting temperature, and the target air-conditioning setting temperature, obtain a maximum Optimal air conditioning setting temperature. The air conditioning system control method according to claim 7, wherein, in step (E), the current confidence level α _t is expressed by the following formula:

Where β is the first adjustable parameter, σ is the second adjustable parameter, F (.) is the indoor temperature prediction model, s _{t -Δ T is} the at least one previous indoor temperature sensing value, a _{t -△ T is} the previous air conditioner setting temperature, s _{t is} the at least one indoor temperature sensing value, t -△ T is the previous time, ∥.∥ is the norm function, α _{t -△ T is} the previous confidence Degree, and α ₀ =0.5. In step (G), the optimal air conditioner set temperature

The following formula represents:

in,

Set the temperature for the target air conditioner,

The air conditioner setting temperature is recommended for this goal.

Images