WO2020098405A1 - 空调器的控制方法、空调器及存储介质 - Google Patents

空调器的控制方法、空调器及存储介质 Download PDF

Info

Publication number
WO2020098405A1
WO2020098405A1 PCT/CN2019/109080 CN2019109080W WO2020098405A1 WO 2020098405 A1 WO2020098405 A1 WO 2020098405A1 CN 2019109080 W CN2019109080 W CN 2019109080W WO 2020098405 A1 WO2020098405 A1 WO 2020098405A1
Authority
WO
WIPO (PCT)
Prior art keywords
air conditioner
parameters
parameter
neural network
network model
Prior art date
Application number
PCT/CN2019/109080
Other languages
English (en)
French (fr)
Inventor
黄志刚
冯静娅
黎顺全
陶骙
朱合华
Original Assignee
广东美的制冷设备有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from CN201811364391.XA external-priority patent/CN109323425B/zh
Priority claimed from CN201910878640.5A external-priority patent/CN110553353B/zh
Application filed by 广东美的制冷设备有限公司 filed Critical 广东美的制冷设备有限公司
Publication of WO2020098405A1 publication Critical patent/WO2020098405A1/zh

Links

Images

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F11/00—Control or safety arrangements
    • F24F11/30—Control or safety arrangements for purposes related to the operation of the system, e.g. for safety or monitoring
    • F24F11/46—Improving electric energy efficiency or saving
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F11/00—Control or safety arrangements
    • F24F11/62—Control or safety arrangements characterised by the type of control or by internal processing, e.g. using fuzzy logic, adaptive control or estimation of values
    • F24F11/63—Electronic processing
    • F24F11/64—Electronic processing using pre-stored data

Definitions

  • the present application relates to the technical field of air conditioners, and in particular, to a control method of an air conditioner, an air conditioner, and a computer-readable storage medium.
  • control table is mainly searched according to the setting parameters set by the user to obtain the operating parameters of the air conditioner.
  • the control table is often made by engineers based on work experience, and the energy efficiency of the air conditioner under different operating conditions is also different, so the idea of the control table is often only enough to ensure the normal operation of the air conditioner and the air conditioning capacity reaches the target , And ignore the consideration of energy efficiency when the air conditioner is running. This makes the air conditioner often unable to operate with better energy efficiency.
  • the main purpose of the present application is to provide a control method of an air conditioner, an air conditioner and a computer-readable storage medium, which solve the problem of how to optimize the energy efficiency ratio of the air conditioner.
  • the present application provides a method for controlling an air conditioner.
  • the method for controlling an air conditioner includes the following steps:
  • the current operating condition parameters include the current environmental parameters and the current setting parameters of the air conditioner
  • the steps of obtaining a plurality of historical operating parameters corresponding to the current operating condition parameters and obtaining a target operating parameter according to the first energy efficiency ratio corresponding to each of the historical operating parameters include:
  • the historical operating parameter with the largest first energy efficiency ratio is used as the target operating parameter.
  • the method further includes:
  • the multiple historical operating parameters corresponding to the current operating condition parameters are executed, and the target operation is obtained according to the first energy efficiency ratio corresponding to each of the historical operating parameters Parameter steps;
  • the number of the historical operating parameters is less than the preset number, and the target operating parameters are obtained according to the correspondence between the current operating condition parameters and the preset operating condition parameters and the operating parameters.
  • the steps of obtaining a plurality of historical operating parameters corresponding to the current operating condition parameters and obtaining a target operating parameter according to the first energy efficiency ratio corresponding to each of the historical operating parameters include:
  • the neural network model is trained or updated according to multiple training samples, and each training sample includes a historical operating condition parameter, the historical operating parameter corresponding to the historical operating condition parameter, and the first energy efficiency ratio.
  • the neural network model is configured to output the historical operating parameter with the largest energy efficiency ratio as the target operating parameter among the multiple training samples corresponding to the historical operating condition parameter and the current operating condition parameter.
  • the method further includes:
  • the sample is added to the neural network model as the training sample to update the neural network model.
  • the method further includes:
  • cooling capacity is within a preset cooling capacity interval, or the heating capacity is within a preset heating capacity interval, performing the generation according to the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio Sample steps.
  • the method further includes:
  • the compressor operating condition parameter is less than a preset threshold, the step of generating a sample according to the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio is performed.
  • the method further includes:
  • the step of adding the sample as the training sample to the neural network model to update the neural network model is performed.
  • the preset condition includes any one of the following:
  • the ratio between the total number of the samples and the first preset value is greater than the reference value
  • the ratio between the total number of the samples and the total number of the training samples and the second preset value is greater than the reference value
  • the reference value is cumulatively increased by one.
  • the method before the step of inputting the current operating condition parameter into the neural network model to obtain the target operating parameter, the method further includes:
  • the method further includes:
  • the step of inputting the current working condition parameters into the neural network model to obtain the target operating parameters is performed.
  • the neural network model has not been trained, continue to perform the step of network training the neural network model according to the first training sample.
  • control method of the air conditioner further includes:
  • the neural network model is trained.
  • the method before the step of dividing the training samples into a first training sample and a second training sample according to a preset ratio, the method further includes:
  • Initialize the neural network model before training set the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes in the neural network model, and the initial connection weight of the input layer and hidden layer, the hidden layer and The initial connection weight of the output layer, the initial hidden layer threshold and the initial output layer threshold.
  • the method further includes:
  • the method further includes:
  • the step of controlling the operation of the air conditioner according to the target operating parameter is performed.
  • the present application also provides an air conditioner, the air conditioner includes:
  • the air conditioner includes a memory, a processor, and a control program of the air conditioner stored on the memory and operable on the processor.
  • the control program of the air conditioner is executed by the processor, the air conditioner is implemented as described above Of the control method of the device.
  • the present application also provides a computer-readable storage medium on which a control program for an air conditioner is stored.
  • a control program for an air conditioner is stored.
  • the control program of the air conditioner is executed by a processor, the above air conditioner Steps of the control method.
  • An air conditioner control method, an air conditioner, and a computer-readable storage medium provided by the present application to obtain current operating condition parameters of the air conditioner, where the current operating condition parameters include current environmental parameters and current setting parameters of the air conditioner; A plurality of historical operating parameters corresponding to the current operating condition parameters, and obtaining a target operating parameter according to a first energy efficiency ratio corresponding to each of the historical operating parameters; controlling the air conditioner to operate with the target operating parameter.
  • the problem of how to optimize the energy efficiency ratio of the air conditioner is solved by finding the operating parameter of the air conditioner according to the energy efficiency ratio corresponding to the operating parameter and controlling the operation of the air conditioner.
  • FIG. 1 is a schematic diagram of a hardware operating environment of a terminal according to an embodiment of the present application
  • FIG. 2 is a schematic flowchart of a first embodiment of a control method for an air conditioner of this application
  • FIG. 3 is a schematic flowchart of a second embodiment of a control method for an air conditioner of this application.
  • FIG. 4 is a schematic flowchart of a third embodiment of a control method for an air conditioner of this application.
  • FIG. 5 is a schematic flowchart of a fourth embodiment of a control method for an air conditioner of this application.
  • FIG. 6 is a schematic flowchart of a fifth embodiment of a control method for an air conditioner of this application.
  • FIG. 7 is a schematic flowchart of a sixth embodiment of a control method for an air conditioner of this application.
  • FIG. 8 is a schematic flowchart of a seventh embodiment of a control method for an air conditioner of this application.
  • FIG. 9 is a frame diagram of a neural network model according to an embodiment of a control method of an air conditioner of this application.
  • FIG. 10 is a schematic flowchart of an eighth embodiment of a control method for an air conditioner of this application.
  • FIG. 11 is a schematic flowchart of a ninth embodiment of a control method for an air conditioner of this application.
  • FIG. 12 is a schematic flowchart of a tenth embodiment of a control method for an air conditioner of this application.
  • FIG. 13 is a schematic flowchart of an eleventh embodiment of a control method for an air conditioner of this application.
  • 15 is a schematic flowchart of a thirteenth embodiment of a control method for an air conditioner of this application.
  • 16 is a schematic flowchart of a fourteenth embodiment of a control method for an air conditioner of this application.
  • This application provides a control method for an air conditioner, which solves the problem of how to optimize the energy efficiency ratio of the air conditioner.
  • FIG. 1 is a schematic diagram of a hardware operating environment of a terminal according to an embodiment of the present application
  • the terminal may be an air conditioner, or may be a control terminal or a server that controls the air conditioner.
  • the terminal may include: a processor 1001, such as a CPU (central processing unit), a memory 1002, and a communication bus 1003.
  • the communication bus 1003 is used to realize the connection communication between the various components in the terminal.
  • the memory 1002 may be a high-speed RAM (random-access memory) or a stable memory (non-volatile memory), such as a disk memory.
  • the memory 1002 may optionally be a storage device independent of the foregoing processor 1001.
  • FIG. 1 does not constitute a limitation on the terminal in the embodiments of the present application, and may include more or fewer components than those illustrated, or combine certain components, or different components Layout.
  • the memory 1002 as a computer storage medium may include a control program of the air conditioner.
  • the processor 1001 may be used to call the control program of the air conditioner stored in the memory 1002, and perform the following operations:
  • the current operating condition parameters include the current environmental parameters and the current setting parameters of the air conditioner
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the historical operating parameter with the largest first energy efficiency ratio is used as the target operating parameter.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the multiple historical operating parameters corresponding to the current operating condition parameters are executed, and the target operation is obtained according to the first energy efficiency ratio corresponding to each of the historical operating parameters Parameter steps;
  • the number of the historical operating parameters is less than the preset number, and the target operating parameters are obtained according to the correspondence between the current operating condition parameters and the preset operating condition parameters and the operating parameters.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the neural network model is trained or updated according to a plurality of training samples, each training sample includes a historical operating condition parameter, the historical operating parameter corresponding to the historical operating condition parameter, and the first energy efficiency ratio, the neural network model It is configured to output the historical operating parameter with the largest energy efficiency ratio as the target operating parameter among the multiple training samples corresponding to the historical operating condition parameter and the current operating condition parameter.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the sample is added to the neural network model as the training sample to update the neural network model.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • cooling capacity is within a preset cooling capacity interval, or the heating capacity is within a preset heating capacity interval, performing the generation according to the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio Sample steps.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the compressor operating condition parameter is less than a preset threshold, the step of generating a sample according to the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio is performed.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the step of adding the sample as the training sample to the neural network model to update the neural network model is performed.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the ratio between the total number of the samples and the first preset value is greater than the reference value
  • the ratio between the total number of the samples and the total number of the training samples and the second preset value is greater than the reference value
  • the reference value is cumulatively increased by one.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the step of inputting the current working condition parameters into the neural network model to obtain the target operating parameters is performed.
  • the neural network model has not been trained, continue to perform the step of network training the neural network model according to the first training sample.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the neural network model is trained.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • Initialize the neural network model before training set the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes in the neural network model, and the initial connection weight of the input layer and hidden layer, the hidden layer and The initial connection weight of the output layer, the initial hidden layer threshold and the initial output layer threshold.
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • processor 1001 can call the control program of the air conditioner stored in the memory 1002, and also perform the following operations:
  • the step of controlling the operation of the air conditioner according to the target operating parameter is performed.
  • control method of the air conditioner includes:
  • Step S10 Obtain the current operating condition parameter of the air conditioner, where the current operating condition parameter includes the current environmental parameter and the current setting parameter of the air conditioner.
  • the terminal in the embodiment may be an air conditioner, or a control terminal or server that controls the air conditioner.
  • the terminal of the embodiment is an air conditioner.
  • the current environmental parameter includes at least one of a current outdoor temperature, a current indoor temperature, a current outdoor humidity, and a current indoor humidity;
  • the current set parameter includes at least one of a current set temperature and a current set humidity .
  • the terminal is provided or connected with a data collection module, the data collection module has an outdoor temperature sensor, an indoor temperature sensor, an indoor humidity sensor, and an outdoor humidity sensor, the outdoor temperature sensor is used to detect the current outdoor temperature, the indoor temperature sensor It is used to detect the current indoor temperature, the outdoor humidity sensor is used to detect the current outdoor humidity, and the indoor humidity sensor is used to detect the current indoor humidity.
  • the terminal may obtain the setting parameters of the last startup operation of the air conditioner when the air conditioner is turned off as the current setting parameters when the air conditioner is turned on; Air conditioner setting parameters), obtain the changed setting parameters as the current setting parameters.
  • the terminal may obtain the current working condition parameters of the air conditioner regularly or in real time.
  • Step S20 Acquire a target operating parameter according to a plurality of historical operating parameters corresponding to the current operating condition parameter and a first energy efficiency ratio corresponding to each of the historical operating parameters.
  • the operating parameter corresponding to each historical operating parameter is the same as the current operating parameter, or the operating parameter corresponding to each historical operating parameter is the same as the current operating parameter
  • the parameters are in the same value range.
  • the first energy efficiency ratio is the energy efficiency ratio achieved by the air conditioner when the air conditioner is operated with the historical operating parameter based on the operating condition parameter corresponding to the historical operating parameter or the current operating condition parameter.
  • the terminal may use the historical operating parameter with the largest energy efficiency ratio among the multiple historical operating parameters corresponding to the current operating condition parameter as the target operating parameter.
  • the terminal is set or connected with a neural network model.
  • NN Neural Networks
  • the neural network model reflects many basic functions of the human brain Feature is a highly complex nonlinear dynamic learning system.
  • the neural network model has large-scale parallelism, distributed storage and processing, self-organization, self-adaptive and self-learning capabilities, and is particularly suitable for processing inaccurate and fuzzy information processing problems that need to consider many factors and conditions at the same time.
  • the terminal may use the current operating condition parameter as an input parameter, and input it into a neural network model previously trained based on multiple historical operating parameters corresponding to the current operating condition parameter and the first energy efficiency ratio corresponding to the historical operating parameter,
  • the neural network model will output a set of operating parameters as output values as the target operating parameters corresponding to the current working condition parameters.
  • the current environmental parameters include the current outdoor temperature and the current indoor temperature
  • the current set parameters include the current set temperature
  • the target operating parameters corresponding to the output of the neural network model may include compressor frequency, electronic expansion Valve opening, internal fan speed and external fan speed
  • the current environmental parameters include the current outdoor temperature, the current indoor temperature, the current outdoor humidity and the current indoor humidity
  • the current set parameters include the current set temperature and the current
  • the target operating parameters corresponding to the output of the neural network model may include compressor frequency, electronic expansion valve opening, internal fan speed, external fan speed, and humidification frequency.
  • the historical operating parameters in the neural network model can be the historical operating parameters entered by the engineer into the training sample in the neural network model, or it can be the new training obtained by the neural network model based on the training sample self-learning The historical operating parameters in the sample.
  • the automatic optimization of the energy efficiency ratio of the air conditioner can be achieved by using the pre-built neural network model, and further optimization of the energy efficiency ratio of the air conditioner can be achieved through continuous iterative training of the neural network model.
  • Step S30 Acquire a target operating parameter according to a plurality of historical operating parameters corresponding to the current operating condition parameter and a first energy efficiency ratio corresponding to each of the historical operating parameters.
  • the terminal can control the air conditioner to operate with the target operating parameter.
  • the terminal may perform steps S10 to S30 when it detects a change in the setting parameters of the air conditioner; the terminal may also detect that the amount of change in any of the current environmental parameters exceeds the preset
  • steps S10 to S30 are executed, where the preset change amount range corresponding to the temperature value may be 0-1 ° C, and the preset change amount corresponding to the humidity value may be 0-5%.
  • the current operating condition parameters of the air conditioner are obtained, and the current operating condition parameters include the current environmental parameters and the current setting parameters of the air conditioner; according to the multiple historical operating parameters corresponding to the current operating condition parameters And obtaining a target operating parameter according to the first energy efficiency ratio corresponding to each of the historical operating parameters; controlling the air conditioner to operate with the target operating parameter.
  • the problem of how to optimize the energy efficiency ratio of the air conditioner is solved by finding the operating parameter of the air conditioner according to the energy efficiency ratio corresponding to the operating parameter and controlling the operation of the air conditioner.
  • the multiple historical operating parameters corresponding to the current operating condition parameters, and according to each of the historical operating parameters include:
  • Step S21 Input the current operating condition parameter into a neural network model to obtain the target operating parameter, wherein the neural network model is based on multiple historical operating parameters corresponding to the current operating condition parameter, and according to each Train or update the first energy efficiency ratio corresponding to the historical operating parameters.
  • a neural network model is set or connected to the terminal.
  • the neural network model can output a set of operating parameters as output values as target operating parameters corresponding to the current operating condition parameters.
  • the neural network model is trained or updated according to multiple training samples, and each training sample includes a historical operating condition parameter, the historical operating parameter corresponding to the historical operating condition parameter, and the first energy efficiency ratio.
  • the pre-built neural network model uses a five-layer fully connected neural network, the hidden layer activation function is relu, and the output layer uses a linear activation function.
  • Use multiple data sets that traverse the numerical combination of all operating condition parameters (the operating condition parameters at this time are historical operating condition parameters), and the historical operating parameters corresponding to the data sets and the first energy efficiency ratio corresponding to the historical operating parameters as the neural network model
  • Training samples (the historical operating parameters in each training sample have corresponding historical operating parameters, and each historical operating parameter has its corresponding first energy efficiency ratio), train the neural network to stop iteration after the loss function no longer drops .
  • the first energy efficiency ratio corresponding to the historical operating parameter is that the air conditioner runs under the historical operating parameter corresponding to the historical operating parameter, and runs on historical When the parameter is running, the energy efficiency ratio output by the air conditioner.
  • these 90 data sets are added as training samples to the neural network model for iterative training.
  • another 21 sets of training samples containing the full range of operating conditions are used to verify the performance of the neural network model. If the deviation between the predicted value and the actual value on the 21 training samples is found to be within ⁇ 5%, it means that the neural network model It is a feasible trained neural network model.
  • the historical operating condition parameters include historical environmental parameters and historical setting parameters, where the historical environmental parameters include at least one of historical outdoor temperature, historical indoor temperature, historical outdoor humidity, and historical indoor humidity; the historical setting The parameters include at least one of historically set temperature and historically set humidity.
  • the historical operating parameters include at least one of compressor frequency, electronic expansion valve opening, internal fan speed, external fan speed, and humidification frequency.
  • each training sample generated by self-learning also has historical operating parameters and historical operating parameters And the first energy efficiency ratio.
  • the terminal after acquiring the current working condition parameters, the terminal inputs the current working condition parameters as input values into the neural network model, and the neural network model may query all historical working condition parameters corresponding to the current working condition parameters in the training sample (These historical operating condition parameters are the same as the current operating condition parameters or within the same value range as the current operating condition parameters), and the historical operating parameters with the largest first energy efficiency ratio corresponding to these historical operating condition parameters are used as output values The output is used as the target operating parameter corresponding to the current operating condition parameter.
  • the training sample corresponding to the same historical working condition parameter has at least one, and the training sample corresponding to the historical working condition parameter and the current working condition parameter has At least two.
  • the historical operating parameters in the training sample with the same historical operating parameter may be different, for example, the first training sample has historical operating parameter A and historical operating parameter B, while the second training sample has the historical operating parameter A and the historical operating parameter C, compared with the first training sample and the second training sample, the historical operating parameter is the same, and the historical operating parameter is different.
  • the historical operating parameter B corresponds to The energy efficiency ratio is also different from the historical operating parameter C.
  • the current operating condition parameter is input to a neural network model to obtain the target operating parameter, wherein the neural network model is based on multiple historical operating parameters corresponding to the current operating condition parameter, and according to The first energy efficiency ratio corresponding to each of the historical operating parameters is trained or updated.
  • the neural network model is based on multiple historical operating parameters corresponding to the current operating condition parameter, and according to The first energy efficiency ratio corresponding to each of the historical operating parameters is trained or updated.
  • Step S40 Detect the number of historical operating parameters corresponding to the current operating condition parameters.
  • Step S50 The number of the historical operating parameters is less than the preset number, and a target operating parameter is obtained according to the current operating condition parameter and the correspondence between the preset operating condition parameter and the operating parameter.
  • the terminal may query the number of historical operating parameters corresponding to the current operating condition parameters, or query the existing training samples in the neural network model, the historical operating condition parameters and the current operating condition parameters The number of training samples of the condition parameter is taken as the number of historical operating parameters corresponding to the current working condition parameter.
  • the terminal detects whether the number of historical operating parameters corresponding to the current operating condition parameter is greater than or equal to a preset number, and detects When the number of historical operating parameters corresponding to the operating parameters is less than the preset number, the corresponding relationship between the preset operating parameters and the operating parameters is obtained, and then obtained according to the corresponding relationship between the current operating parameters and the preset operating parameters and the operating parameters
  • the target operating parameter corresponding to the current operating condition parameter in addition, when the terminal detects that the number of historical operating parameters corresponding to the current operating condition parameter is greater than or equal to a preset number, it executes the corresponding to the current operating condition parameter Multiple historical operating parameters, and the step of obtaining the target operating parameter according to the first energy efficiency ratio corresponding to each of the historical operating parameters (ie step 20).
  • the preset number may be set according to actual needs, such as 10-50, optionally 30.
  • the target operating parameter needs to be obtained according to the corresponding relationship between the current operating condition parameter and the preset operating condition parameter and the operating parameter.
  • the preset correspondence relationship between the operating condition parameter and the operating parameter may be obtained by an air conditioner control table having correspondence relationships between multiple operating condition parameters and the operating parameter developed by the engineer according to work experience.
  • the terminal can query the air conditioner control table according to the current working condition parameters for the working condition parameters whose values correspond to the current working condition parameters (the value is the same as the current working condition parameter, or the value and the current working condition parameter are in the same value range), and then obtain The operating parameters corresponding to the obtained working condition parameters are used as the target operating parameters.
  • the number of historical operating parameters corresponding to the current operating condition parameters is detected; the number of historical operating parameters is less than the preset number, based on the current operating condition parameters and the preset operating condition parameters and operation The corresponding relationship of the parameters obtains the target operating parameters. In this way, the stability when acquiring the target operating parameters of the air conditioner is improved.
  • the method further includes:
  • Step S60 Determine a second energy efficiency ratio corresponding to the target operating parameter.
  • Step S70 Generate a sample according to the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio.
  • Step S80 Add the sample as the training sample to the neural network model to update the neural network model.
  • the terminal may input the current operating condition parameter into a neural network model to obtain the target operating parameter; or may be based on the corresponding relationship between the current operating condition parameter and the preset operating condition parameter and the operating parameter Get the target operating parameters. After acquiring the target operating parameters, the terminal controls the air conditioner to operate with the target operating parameters.
  • the terminal may detect the energy efficiency parameter of the air conditioner after the indoor environmental parameter reaches the target operating parameter, and the energy efficiency parameter includes the air-out dry bulb temperature, the air-out wet bulb temperature, indoor Dry bulb temperature, indoor wet bulb temperature and current power of the air conditioner.
  • the terminal further determines the second energy efficiency ratio corresponding to the target operating parameter according to the target operating parameter and the energy efficiency parameter.
  • the terminal may first determine the current cooling capacity or heating capacity of the air conditioner according to the target operating parameter and the energy efficiency parameter. Then, the second energy efficiency ratio corresponding to the target operating parameter is determined according to the cooling capacity or the heating capacity, and according to the current power of the air conditioner. It should be noted that when the air conditioner is operating in the cooling mode, the cooling capacity is determined; when the air conditioner is operating in the heating mode, the heating capacity is determined.
  • the calculation formula of the second energy efficiency ratio is as follows:
  • E is the second energy efficiency ratio
  • Q is the cooling capacity or heating capacity
  • W is the current power of the air conditioner.
  • cooling capacity or heating capacity Q is as follows:
  • the enthalpy value ha and the air-enthalpy value hb can be obtained by looking up the enthalpy humidity chart;
  • the sphere and the rheumatic sphere are determined;
  • V1 is the air volume of the air conditioner;
  • V2 is the specific volume of the wind, and V2 can be determined by looking at the enthalpy chart.
  • the calculation formula of the air volume of the air conditioner is as follows:
  • n is the speed of the internal fan
  • N is the speed of the external fan
  • V3 is the rated air volume
  • the terminal may generate a sample based on the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio, where the current operating condition parameter in the sample becomes the historical operating condition parameter, and the target operation
  • the parameters are changed to historical operating parameters, and the second energy efficiency ratio is changed to the first energy efficiency ratio.
  • the terminal may add the sample as a training sample to the neural network model to update the neural network model.
  • the target operating parameter is obtained according to the corresponding relationship between the current operating condition parameter and the preset operating condition parameter and the operating parameter, by generating a sample corresponding to the target operating parameter, as the training samples of the neural network model are continuously updated Iteration can realize that the number of training samples corresponding to the current working condition parameters is greater than or equal to the preset number.
  • the neural network model training is completed so that the trained neural network model can automatically optimize the energy efficiency ratio of the air conditioner Requirements; when the target operating parameters are obtained from multiple historical operating parameters corresponding to the current operating condition parameters and the first energy efficiency ratio corresponding to each of the historical operating parameters, by generating a sample corresponding to the target operating parameters, When added as a training sample to the neural network model for training, it can also achieve continuous optimization of the neural network model.
  • the method further includes:
  • Step S61 Obtain the cooling capacity or heating capacity corresponding to the second energy efficiency ratio.
  • Step S71 The cooling capacity is within a preset cooling capacity interval, or the heating capacity is within a preset heating capacity interval, then the execution according to the current operating condition parameter, the target operating parameter and the second The steps of generating samples for energy efficiency ratio.
  • the terminal may determine the current cooling capacity or heating capacity of the air conditioner according to the target operating parameter and the energy efficiency parameter. It should be noted that when the air conditioner is currently operating in the cooling mode, the current cooling capacity of the air conditioner is determined; when the air conditioner is currently operating in the heating mode, the current heating capacity of the air conditioner is determined.
  • the obtained cooling capacity or heating capacity determined according to the target operating parameter and the energy efficiency parameter is the cooling capacity or heating capacity corresponding to the second energy efficiency ratio.
  • the terminal may query the air conditioner control table for the working condition parameters corresponding to the current working condition parameters (the value is the same as the current working condition parameter, or the value is at the same value as the current working condition parameter Range), and then obtain the operating parameter corresponding to the obtained operating condition parameter, and then determine the cooling capacity or heating capacity corresponding to the operating condition parameter according to the operating condition parameter and the operating parameter. Adding and subtracting the preset value before and after the cooling capacity to obtain the two end values of the preset cooling capacity interval, so that the preset cooling capacity interval can be obtained; adding and subtracting the preset value before and after the heating capacity to obtain the preset heating capacity Two end values of the interval, so that the preset heating interval can be obtained.
  • the value range of the preset value may be 50W-200W, optionally 100W.
  • Step 70 when the air conditioner is currently operating in the cooling mode, when the cooling capacity for the second energy efficiency ratio is within a preset cooling capacity interval, it is determined that the sample generation condition is currently satisfied, and the process according to the current operating condition parameter is performed. Step of generating the target operating parameter and the second energy efficiency ratio (ie step 70). When the cooling capacity for the second energy efficiency ratio is not in the preset cooling capacity interval, this set of data can be discarded, and samples of the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio can no longer be generated.
  • the air conditioner when the air conditioner is currently operating in the heating mode, when the heating capacity for the second energy efficiency ratio is within a preset heating range, it is determined that the current generation condition of the sample is satisfied, and the operation according to the current operating condition is performed.
  • the parameter, the target operating parameter and the second energy efficiency ratio generate a sample step (ie step 70).
  • this set of data can be discarded, and samples of the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio can no longer be generated.
  • the training process of the neural network model may also be: when a set of historical operating condition parameters are input to the neural network model, all possible control parameters of the air conditioner are The parameter value range and step size traverse all possible values to obtain all possible control parameter combinations as the historical operating parameters corresponding to the historical operating condition parameters.
  • the neural network model For each combination of possible control parameters of this historical operating parameter, input it to the neural network model in turn, and calculate the cooling capacity or heating capacity obtained after running according to this group of parameters as Q, when the cooling capacity is within the preset cooling capacity interval, or When the heating capacity is within the preset heating capacity range, the historical operating condition parameter and the cooling capacity corresponding to the historical operating condition parameter are within the preset cooling capacity range, or the heating capacity is within the historical operating parameter within the preset heating capacity range , And the first energy efficiency value corresponding to the historical operation parameter constitute a qualified training sample, and the training sample together with the output value of the neural network is recorded in the list as a piece of data.
  • the cooling capacity or heating capacity corresponding to the second energy efficiency ratio is acquired; if the cooling capacity is within a preset cooling capacity interval, or the heating capacity is within a preset heating capacity interval, then executing the The step of generating a sample according to the current operating condition parameter, the target operating parameter and the second energy efficiency ratio. In this way, it can be ensured that the cooling capacity or heating capacity that can be achieved by the target operating parameters output by the neural network model can meet the normal cooling or heating requirements of the air conditioner.
  • the step of controlling the air conditioner to operate with the target operating parameter further includes:
  • Step S62 Obtain the compressor operating condition parameter of the air conditioner.
  • Step S72 The compressor operating condition parameter is less than a preset threshold, and then the step of generating a sample according to the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio is performed.
  • the terminal may detect and obtain the compressor operating condition parameter of the air conditioner after the indoor environmental parameter reaches the target operating parameter; the terminal may also obtain the prediction by the neural network model It is obtained that when the air conditioner is operated with the target operating parameter under the environment corresponding to the current operating condition parameter, the compressor operating condition parameter of the compressor of the air conditioner.
  • the compressor operating condition parameter includes at least one of compressor discharge temperature, compressor current, and compressor refrigerant pressure. It should be noted that different types of compressor operating parameters correspond to different preset thresholds. When the terminal detects that the compressor operating parameters are less than the preset thresholds corresponding to the compressor operating parameters, the terminal executes the The step of generating a sample by the operating condition parameter, the target operating parameter and the second energy efficiency ratio. The following uses the compressor operating condition parameter as the compressor discharge temperature as an example for description.
  • the terminal when the terminal detects that the acquired compressor exhaust temperature is less than a preset threshold (ie, preset temperature) corresponding to the compressor exhaust temperature, it determines that the air conditioner ’s The load is within a reasonable operating range, that is, the compressor is in a stable operation state.
  • the step of generating a sample according to the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio (that is, Step 70); when the terminal detects that the acquired compressor exhaust temperature is greater than or equal to the preset threshold corresponding to the compressor exhaust temperature, it is determined that when the air conditioner is operating at the current target operating temperature, the load of the air conditioner is not It is within a reasonable operating range, that is, the compressor is in an unstable operating state.
  • this set of data can be discarded, and samples of the current operating condition parameter, the target operating parameter, and the second energy efficiency ratio can no longer be generated.
  • the value range of the preset threshold corresponding to the compressor exhaust temperature may be 90 ° C-100 ° C, optionally 95 ° C.
  • the training process of the neural network model may also be: when a set of historical operating condition parameters are input to the neural network After the network model, all possible control parameters of the air conditioner are traversed according to the value range and step size of the parameters to obtain all possible control parameter combinations as historical operating parameters corresponding to the historical operating condition parameters.
  • the neural network model For each combination of possible control parameters of this historical operating parameter, input it to the neural network model in turn, and calculate the cooling capacity or heating capacity obtained after running according to this group of parameters as Q, when the cooling capacity is within the preset cooling capacity interval, or When the heating capacity is within the preset heating capacity range, it is further determined whether the compressor operating condition parameter corresponding to the historical operating parameter in this set of data is less than the preset threshold, and when the compressor operating condition parameter is less than the preset threshold, the The historical operating parameter and the historical operating parameter corresponding to the historical operating parameter in the preset cooling capacity interval or the heating capacity in the preset heating capacity interval, and the first energy efficiency value corresponding to the historical operating parameter constitute a Qualified training samples, record the training samples together with the output value of the neural network as a piece of data in the list.
  • the compressor operating condition parameter of the air conditioner is acquired; if the compressor operating condition parameter is less than a preset threshold, then performing the operation according to the current operating condition parameter, the target operating parameter, and the The second energy efficiency ratio is the step of generating samples. In this way, it can be ensured that the air conditioner compressor can operate stably when the air conditioner is operating with the target operating parameters output by the neural network model.
  • the sample is generated according to the current operating condition parameter, the target operating parameter and the second energy efficiency ratio After the steps, it also includes:
  • Step S81 The sample satisfies the preset condition, and then the step of adding the sample as the training sample to the neural network model to update the neural network model is performed.
  • a new sample may be generated, that is, the sample is added to the neural network model as a training sample, the neural network model is updated, and the neural network model is further trained according to the new training sample and the original training sample; or
  • all samples not added to the neural network model are used as training samples, added to the neural network model, the neural network model is updated, and the nerve is further trained together with the original training sample according to the new training sample Network model.
  • the preset condition includes any one of the following: the ratio between the total number of the samples and the first preset value is greater than the reference value; the sum of the total number of the samples and the total number of the training samples and the second preset value The ratio between is greater than the reference value.
  • the preset condition is that when the ratio between the total number of samples and the first preset value is greater than the reference value, the condition formula is as follows:
  • n1 is the total number of samples not added to the neural network model
  • X is the first preset value
  • N is the reference value. The reference value is cumulatively increased by one every time the neural network model is updated.
  • the preset condition is that when the ratio between the total number of the samples and the total number of the training samples and the second preset value is greater than the reference value, the conditional formula is as follows:
  • n1 is the total number of samples not added to the neural network model
  • n2 is the total number of existing training samples in the neural network model
  • Y is the first preset value
  • N is the reference value, each time the neural network model is updated, reference The value is incremented by one.
  • the reference value is cumulatively increased by one
  • the initial value of the required reference value can be selected as 0; the value range of the first preset value can be selected from 1-1000, and the value can be 100; the value range of the second preset value can be Select 1-1000, the value can be 100.
  • the preset condition may be characterized as that the number of samples not added to the neural network model reaches a certain number.
  • the terminal may store new samples in the database after each new sample is generated, and when the samples in the database reach a certain number (for example, the samples in the database satisfy (n1 / 100)> N, N follows
  • the step of adding the sample to the neural network model as the training sample to update the neural network model is performed. In this way, frequent updates of the neural network model can be avoided, which can play a role in saving system resources.
  • control method of the air conditioner further includes:
  • Step S100 periodically obtaining the required cooling / heating capacity in the room
  • the required cooling / heating capacity of the room can be obtained according to the difference between the current indoor temperature and the indoor target temperature, and after the air conditioner operates according to the current operating parameters for a preset period of time, the current indoor temperature is detected to obtain the indoor Required cooling / heating, to avoid the air conditioner running in accordance with the current operating parameters is too short, resulting in inaccurate indoor cooling / heating required, and to avoid too much fluctuation when the air conditioner starts operating according to the current operating parameters.
  • the required cooling / heating capacity obtained is inaccurate; the preset duration is greater than 0.5h, such as 1h, 1.5h, etc.
  • Step S200 Acquire target operating parameters of the air conditioner according to the required cooling / heating capacity and a preset neural network model, which is generated by training based on sample data in the sample database;
  • the neural network model includes an input layer, a hidden layer, and an output layer. Connection weights are set between the input layer and the hidden layer, and between the hidden layer and the output layer Connection weights are also set, the hidden layer is also set with a hidden layer threshold, and the output layer is set with an output layer threshold.
  • the sample database includes multiple sets of sample data, each sample data includes input parameters and output parameters, the input parameters include environmental parameters and air conditioning operating parameters, the environmental parameters include but are not limited to collected as fixed value input Indoor ambient temperature, indoor ambient humidity, outdoor ambient temperature, air outlet temperature, air outlet humidity, indoor fan speed, total operating voltage of air conditioner, total operating current of air conditioner; operating parameters of the air conditioner include but are not limited to adjustable Variable input compressor frequency, indoor fan speed, outdoor fan speed, electronic expansion valve opening.
  • the air conditioner When the air conditioner is turned on, it operates according to preset initial operating parameters, but the operating parameters include, but are not limited to, compressor frequency, indoor fan speed, outdoor fan speed, electronic expansion valve opening, etc .;
  • the cooling / heating capacity Q of the air conditioner can be calculated from the air conditioner outlet air parameters and the indoor environmental parameters. Calculations include but are not limited to the enthalpy difference method and the heat balance method.
  • the sample database uses the corresponding air-conditioning operating parameters (for example: compressor frequency, indoor fan speed, outdoor fan speed, electronic expansion valve opening, indoor ambient temperature T1, indoor ambient humidity ⁇ room, outdoor ambient temperature T4, etc.) as A set of inputs on the network, the cooling / heating capacity Q of the air conditioner, and the energy efficiency ratio COP as a set of output, the input and output form a set of sample data, and store the sample data in the database;
  • air-conditioning operating parameters for example: compressor frequency, indoor fan speed, outdoor fan speed, electronic expansion valve opening, indoor ambient temperature T1, indoor ambient humidity ⁇ room, outdoor ambient temperature T4, etc.
  • the sample data in the sample data can be retrieved to perform network training on the neural network model, and input parameters of the sample data and input parameters ( Obtained from the sample data) includes environmental parameters (such as indoor ambient temperature, indoor ambient humidity, outdoor ambient temperature, etc.) and operating parameters (such as compressor frequency, indoor fan speed, outdoor fan speed, electronic expansion valve opening, etc.) ;
  • the output parameters obtained from the same set of sample data
  • the output parameters include the cooling / heating capacity Q of the air conditioner and the energy efficiency ratio COP. Due to the input parameters of the sample data and The output parameters are determined.
  • connection weight between the input layer and the hidden layer in the neural network model Value the connection weight between the hidden layer and the output layer; the more groups of sample data input, the more accurate the connection weight will be adjusted, and the neural network model will be realized. Training.
  • the preset neural network model is a neural network model that has been trained / trained.
  • the indoor cooling / heating required is used as the output parameter of the preset neural network model.
  • the preset neural network model combines optimization algorithms such as genetic algorithm, particle swarm optimization algorithm, etc. to reversely solve the input parameters of the preset neural network model, and the input parameters include the target operating parameters such as compressor frequency, indoor Extension speed, outdoor fan speed, electronic expansion valve opening, etc.
  • step S300 the air conditioner is controlled to operate according to the target operating parameter, where the energy efficiency ratio of the air conditioner operating with the target operating parameter is greater than the energy efficiency ratio corresponding to other operating parameters on the premise that the cooling / heating capacity is reached.
  • the energy efficiency ratio of the air conditioner operating with the target operating parameter is greater than that of other operating parameters Energy efficiency ratio, so when the air conditioner operates according to the target operating parameters, the energy efficiency ratio of the air conditioner will be maximized, that is, when the air conditioner reaches the required cooling / heating capacity, the power of the air conditioner is the smallest, and the air conditioner is the most economical. Electricity.
  • this application will obtain the required cooling / heating capacity indoors at regular intervals, and the target operating parameter is the initial solution condition: on the premise that the cooling / heating capacity is reached, the air conditioner uses all The energy efficiency ratio of the target operating parameter during operation is greater than the energy efficiency ratio corresponding to other operating parameters.
  • the target operating parameter is obtained through a preset trained neural network model and the above-mentioned solving conditions, and the air conditioner is controlled to operate according to the target operating parameter After that, the energy efficiency ratio of the air conditioner will be maximized, that is, when the air conditioner reaches the required cooling / heating capacity, the power of the air conditioner is the smallest, and the air conditioner is most power-saving at this time, thereby achieving the purpose of lower power consumption and delay
  • the loss of various components of the air conditioner extends the service life of the air conditioner.
  • the ninth embodiment of the air conditioner control method of the present application includes:
  • Step S400 the sample data in the sample database is divided into training samples and test samples according to a preset ratio, the sample data includes input parameters and output parameters, and the input parameters include environment parameters input as fixed values and as adjustable variables Input air conditioning operating parameters, the output parameters include cooling / heating capacity and energy efficiency ratio;
  • Step S500 Train the neural network model according to the training samples, and update the connection weight of the input layer and the hidden layer, the connection weight of the hidden layer and the output layer, and the hidden layer threshold in the neural network model And output layer threshold.
  • the sample data in the sample database is divided into training samples and test samples according to a preset ratio, for example, 85% of the total amount of sample data is used as the training sample, and the remaining 15% of the total amount of sample data is used as the test Samples, the training samples are used to iteratively train the neural network model.
  • the tenth embodiment of the air conditioner control method of the present application includes:
  • Step S600 Determine whether the neural network model has been trained according to the test sample
  • Step S700 if the neural network model is trained, the preset neural network model is generated;
  • step S50 If the neural network model has not been trained, then proceed to step S50.
  • test samples are used to determine whether the neural network model has been trained / completed to ensure the neural network model ’s accuracy.
  • the step S600 includes:
  • Step S610 Obtain the difference between the output parameter of the output sample and the target parameter corresponding to the input parameter of the test sample and the neural network model, where the target parameter includes the target cooling / heating capacity and the target energy efficiency ratio;
  • Step S620 if the difference is less than the preset difference, it is determined that the neural network model has been trained.
  • the output parameters corresponding to the output layer after the neural network model include cooling / heating capacity and energy efficiency ratio
  • the target parameters include target cooling / heating capacity and target energy efficiency ratio.
  • the difference between the cooling / heating capacity (such as 2900) and the target cooling / heating capacity (such as 3000) corresponding to the input parameters of the test sample (the difference is 100 at this time), if the preset difference is 120, at this time, If the difference is less than the preset difference, it is determined that the neural network model has been trained; if the difference is greater than the preset difference, the neural network model also needs to continue iterative training through training samples.
  • the twelfth embodiment of the air conditioner control method of the present application based on the embodiment shown in FIGS. 10-13 above, before the step S400, further includes:
  • Step S800 Initialize the neural network model before training, set the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes in the neural network, and the initial connection weight of the input layer and the hidden layer. The initial connection weight of the layer and the output layer, the initial hidden layer threshold and the initial output layer threshold.
  • the neural network model before training is initialized, for example, the initial connection weight, the initial hidden layer threshold, and the initial output layer threshold are all assigned 1 or 0, or a random number.
  • a thirteenth embodiment of the air conditioner control method of the present application based on the embodiment shown in FIGS. 10-14 above, before the step S300, further includes:
  • Step S900 Determine whether the target operating parameter is within the preset operating range of the air conditioner
  • step S30 is executed.
  • the air conditioner has a certain operating range.
  • the operating frequency of the air conditioner compressor is generally 10-98 Hz
  • the opening of the electronic expansion valve is 50-400 degrees
  • the speed of the outdoor fan is 500-3000 rpm.
  • the indoor fan speed is 400-1000 rpm / h, so before controlling the air conditioner to operate according to the target operating parameter, it is necessary to determine whether the target operating parameter is within the preset operating range of the air conditioner, if it is within the operating range , The air conditioner may execute the target operating parameter, and if it is not within the operating range, re-solve the target operating parameter or retrain the neural network model.
  • a fourteenth embodiment of the air conditioner control method of the present application further includes:
  • Step S110 when receiving the stop optimization instruction, stop executing the step of acquiring the target operating parameters of the air conditioner according to the required cooling / heating capacity and the preset neural network model;
  • Step S120 controlling the air conditioner to operate according to the finally obtained target operating parameter.
  • the solution process of the target operating parameters is stopped to avoid the constant solution of energy consumption, and at the same time, the air conditioner is controlled to operate according to the finally obtained target operating parameters, which can also ensure the air conditioner It is in the most power-saving state within a certain period of time.
  • the present application also proposes an air conditioner including a memory, a processor, and a control program of the air conditioner stored on the memory and operable on the processor, the processor performs control of the air conditioner
  • the program implements the steps of the air conditioner control method described in the above embodiment.
  • the present application also proposes a computer-readable storage medium, the computer-readable storage medium includes a control program of an air conditioner, and the control program of the air conditioner is executed by a processor to implement the air conditioner as described in the above embodiments Steps of the control method.
  • the methods in the above embodiments can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware, but in many cases the former is optional Implementation.
  • the technical solution of the present application can be embodied in the form of a software product in essence or part that contributes to the existing technology, and the computer software product is stored in a storage medium (such as ROM / RAM as described above) , Disk, CD), including several instructions to make a terminal device (which can be a TV, mobile phone, computer, server, air conditioner, or network equipment, etc.) to perform the method described in each embodiment of the present application.

Abstract

一种空调器的控制方法、空调器以及计算机可读存储介质，所述控制方法包括以下步骤：获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数；控制所述空调器以所述目标运行参数运行。

Description

空调器的控制方法、空调器及存储介质
相关申请
本申请要求2019年9月17日申请的，申请号为201910878640.5，名称为“空调器的控制方法、空调器及存储介质”的中国专利申请的优先权，以及要求2018年11月15日申请的，申请号为201811364391.X，名称为“空调的控制方法、装置及可读存储介质”的中国专利申请的优先权，在此将其全文引入作为参考。
技术领域
本申请涉及空调器技术领域，尤其涉及一种空调器的控制方法、空调器以及计算机可读存储介质。
背景技术
目前，主要是根据用户设定的设定参数查取控制表，以获取空调器的运行参数。而控制表往往是有工程师根据工作经验制定的，且空调器在不同工况条件下运行时的能效亦不同，因此控制表的制定思路往往只足以保证空调器正常运行及空气调节能力达到目标值，而忽略对空调器运行时的能效的考量。这就使得空调器往往不能以更优良的能效运行。
上述内容仅用于辅助理解本申请的技术方案，并不代表承认上述内容是现有技术。
发明内容
本申请的主要目的在于提供一种空调器的控制方法、空调器以及计算机可读存储介质，解决了如何优化空调器的能效比的问题。
为实现上述目的，本申请提供一种空调器的控制方法，所述空调器的控制方法包括以下步骤：
获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；
根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数；
控制所述空调器以所述目标运行参数运行。
可选地，所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤包括：
将所述当前工况参数对应的多个历史运行参数中，所述第一能效比最大的历史运行参数作为所述目标运行参数。
可选地，所述获取空调器的当前工况参数的步骤之后，还包括：
检测与所述当前工况参数对应的历史运行参数的数量；
所述历史运行参数的数量大于或者等于预设数量，则执行所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤；
所述历史运行参数的数量小于预设数量，根据所述当前工况参数和预设的工况参数与运行参数对应关系获取目标运行参数。
可选地，所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤包括：
将所述当前工况参数输入神经网络模型，以得到所述目标运行参数，其中，所述神经网络模型根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比进行训练或更新。
可选地，所述神经网络模型根据多个训练样本训练或者更新，每一训练样本包括历史工况参数、所述历史工况参数对应的所述历史运行参数和所述第一能效比，所述神经网络模型被配置为将所述历史工况参数与所述当前工况参数对应的多个训练样本中，所述第一能效比最大的历史运行参数作为所述目标运行参数输出。
可选地，所述控制所述空调器以所述目标运行参数运行的步骤之后，还包括：
确定所述目标运行参数对应的第二能效比；
根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本；
将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型。
可选地，所述确定所述目标运行参数对应的第二能效比的步骤之后，还包括：
获取所述第二能效比对应的制冷量或者制热量；
所述制冷量处于预设制冷量区间内，或者所述制热量处于预设制热量区间内，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
可选地，所述控制所述空调器以所述目标运行参数运行的步骤之后，还包括：
获取所述空调器的压缩机工况参数；
所述压缩机工况参数小于预设阈值，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
可选地，所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤之后，还包括：
所述样本满足预设条件，则执行所述将所述样本作为所述训练样本加入到所述神经网 络模型中，以更新所述神经网络模型的步骤。
可选地，所述预设条件包括以下任一个：
所述样本的总数与第一预设值之间的比值大于参考值；
所述样本的总数和所述训练样本的总数之和与第二预设值之间的比值大于参考值；
其中，所述将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型的步骤每执行一次，所述参考值累计加一。
可选地，所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤之前，还包括：
将所述训练样本按照预设比例分为第一训练样本及第二训练样本；以及
根据所述第一训练样本对所述神经网络模型进行网络训练，并更新所述神经网络模型中输入层与隐含层的连接权值、隐含层与输出层的连接权值、隐含层阈值及输出层阈值。
可选地，所述根据所述第一训练样本对所述神经网络模型进行网络训练的步骤之后，还包括：
所述神经网络模型训练完毕，则执行所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤；以及
所述神经网络模型未训练完毕，则继续执行所述根据所述第一训练样本对所述神经网络模型进行网络训练的步骤。
可选地，所述空调器的控制方法还包括：
获取所述第二训练样本的输入参数输入所述神经网络模型后对应的输出层输出参数与目标参数的差值；以及
所述差值小于预设差值，则所述神经网络模型训练完毕。
可选地，所述将所述训练样本按照预设比例分为第一训练样本及第二训练样本的步骤之前，还包括：
对训练前的神经网络模型初始化，设定所述神经网络模型中的输入层节点数、隐含层节点数、输出层节点数以及输入层与隐含层的初始连接权值、隐含层与输出层的初始连接权值、初始隐含层阈值及初始输出层阈值。
可选地，所述控制所述空调器根据所述目标运行参数运行的步骤之后，还包括：
接收到停止优化指令，停止执行所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤；以及
控制所述空调器根据最后得到的目标运行参数运行。
可选地，所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历 史运行参数对应的第一能效比获取目标运行参数的步骤之后，还包括：
所述目标运行参数处于所述空调器预设的运行范围内，则执行所述控制所述空调器根据所述目标运行参数运行的步骤。
为实现上述目的，本申请还提供一种空调器，所述空调器包括：
所述空调器包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的空调器的控制程序，所述空调器的控制程序被所述处理器执行时实现如上述空调器的控制方法的步骤。
为实现上述目的，本申请还提供一种计算机可读存储介质，所述计算机可读存储介质上存储有空调器的控制程序，所述空调器的控制程序被处理器执行时实现如上述空调器的控制方法的步骤。
本申请提供的空调器的控制方法、空调器以及计算机可读存储介质，获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数；控制所述空调器以所述目标运行参数运行。这样，通过根据运行参数对应的能效比寻找空调器运行参数，并控制空调器运行，解决了如何优化空调器的能效比的问题。
附图说明
图1为本申请实施例方案涉及的实施例终端的硬件运行环境示意图；
图2为本申请空调器的控制方法第一实施例的流程示意图；
图3为本申请空调器的控制方法第二实施例的流程示意图；
图4为本申请空调器的控制方法第三实施例的流程示意图；
图5为本申请空调器的控制方法第四实施例的流程示意图；
图6为本申请空调器的控制方法第五实施例的流程示意图；
图7为本申请空调器的控制方法第六实施例的流程示意图；
图8为本申请空调器的控制方法第七实施例的流程示意图；
图9为本申请空调器的控制方法一实施例的神经网络模型的框架图；
图10为本申请空调器的控制方法第八实施例的流程示意图；
图11为本申请空调器的控制方法第九实施例的流程示意图；
图12为本申请空调器的控制方法第十实施例的流程示意图；
图13为本申请空调器的控制方法第十一实施例的流程示意图；
图14为本申请空调器的控制方法第十二实施例的流程示意图；
图15为本申请空调器的控制方法第十三实施例的流程示意图；
图16为本申请空调器的控制方法第十四实施例的流程示意图。
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
本申请提供一种空调器的控制方法，解决了如何优化空调器的能效比的问题。
如图1所示，图1是本申请实施例方案涉及的实施例终端的硬件运行环境示意图；
本申请实施例终端可以是空调器，也可以是控制空调器的控制终端或者服务器。
如图1所示，该终端可以包括：处理器1001，例如CPU中央处理器(central processing unit)，存储器1002，通信总线1003。其中，通信总线1003用于实现该终端中各组成部件之间的连接通信。存储器1002可以是高速RAM随机存储器(random-access memory)，也可以是稳定的存储器(non-volatile memory)，例如磁盘存储器。存储器1002可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解，图1中示出的终端的结构并不构成对本申请实施例终端的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
如图1所示，作为一种计算机存储介质的存储器1002中可以包括空调器的控制程序。
在图1所示的终端中，处理器1001可以用于调用存储器1002中存储的空调器的控制程序，并执行以下操作：
获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；
根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数；
控制所述空调器以所述目标运行参数运行。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
将所述当前工况参数对应的多个历史运行参数中，所述第一能效比最大的历史运行参数作为所述目标运行参数。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
检测与所述当前工况参数对应的历史运行参数的数量；
所述历史运行参数的数量大于或者等于预设数量，则执行所述根据所述当前工况参数 对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤；
所述历史运行参数的数量小于预设数量，根据所述当前工况参数和预设的工况参数与运行参数对应关系获取目标运行参数。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
将所述当前工况参数输入神经网络模型，以得到所述目标运行参数，其中，所述神经网络模型根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比进行训练或更新。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
所述神经网络模型根据多个训练样本训练或者更新，每一训练样本包括历史工况参数、所述历史工况参数对应的所述历史运行参数和所述第一能效比，所述神经网络模型被配置为将所述历史工况参数与所述当前工况参数对应的多个训练样本中，所述第一能效比最大的历史运行参数作为所述目标运行参数输出。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
确定所述目标运行参数对应的第二能效比；
根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本；
将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
获取所述第二能效比对应的制冷量或者制热量；
所述制冷量处于预设制冷量区间内，或者所述制热量处于预设制热量区间内，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
获取所述空调器的压缩机工况参数；
所述压缩机工况参数小于预设阈值，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以 下操作：
所述样本满足预设条件，则执行所述将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型的步骤。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
所述样本的总数与第一预设值之间的比值大于参考值；
所述样本的总数和所述训练样本的总数之和与第二预设值之间的比值大于参考值；
其中，所述将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型的步骤每执行一次，所述参考值累计加一。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
将所述训练样本按照预设比例分为第一训练样本及第二训练样本；以及
根据所述第一训练样本对所述神经网络模型进行网络训练，并更新所述神经网络模型中输入层与隐含层的连接权值、隐含层与输出层的连接权值、隐含层阈值及输出层阈值。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
所述神经网络模型训练完毕，则执行所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤；以及
所述神经网络模型未训练完毕，则继续执行所述根据所述第一训练样本对所述神经网络模型进行网络训练的步骤。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
获取所述第二训练样本的输入参数输入所述神经网络模型后对应的输出层输出参数与目标参数的差值；以及
所述差值小于预设差值，则所述神经网络模型训练完毕。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
对训练前的神经网络模型初始化，设定所述神经网络模型中的输入层节点数、隐含层节点数、输出层节点数以及输入层与隐含层的初始连接权值、隐含层与输出层的初始连接权值、初始隐含层阈值及初始输出层阈值。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以 下操作：
接收到停止优化指令，停止执行所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤；以及
控制所述空调器根据最后得到的目标运行参数运行。
进一步地，处理器1001可以调用存储器1002中存储的空调器的控制程序，还执行以下操作：
所述目标运行参数处于所述空调器预设的运行范围内，则执行所述控制所述空调器根据所述目标运行参数运行的步骤。
参照图2，在一实施例中，所述空调器的控制方法包括：
步骤S10、获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数。
本实施例中，实施例终端可以是空调器，也可以是控制空调器的控制终端或者服务器。以下以实施例终端为空调器为例进行说明。
可选地，所述当前环境参数包括当前室外温度、当前室内温度、当前室外湿度和当前室内湿度中的至少一个；所述当前设定参数包括当前设定温度和当前设定湿度中的至少一个。
可选地，终端设置或者连通有数据采集模块，数据采集模块具有室外温度传感器、室内温度传感器、室内湿度传感器和室外湿度传感器，所述室外温度传感器用于检测当前室外温度、所述室内温度传感器用于检测当前室内温度、所述室外湿度传感器用于检测当前室外湿度、所述室内湿度传感器用于检测当前室内湿度。
可选地，终端可以是在空调器开启时，获取空调器关机时的上一次开机运行时的设定参数作为当前设定参数；也可以是在检测到设定参数变更后(0如用户更改空调器的设定参数)，获取变更后的设定参数作为当前设定参数。
可选地，在空调器开机运行的过程中，终端可定时或实时获取空调器的当前工况参数。
步骤S20、根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数。
可选地，当前工况参数对应的多个历史运行参数中，每个历史运行参数对应的工况参数与当前工况参数均相同，或者每个历史运行参数对应的工况参数与当前工况参数处于同一数值范围内。第一能效比则为空调器基于历史运行参数对应的工况参数或者当前工况参数的环境下，以历史运行参数运行时，空调器所达到的能效比。
可选地，终端可以是将当前工况参数对应的多个历史运行参数中，第一能效比最大的 历史运行参数作为目标运行参数。
可选地，终端设置或者连通有神经网络模型。需要说明的是，神经网络(Neural Networks,NN)是由大量的、简单的处理单元(称为神经元)广泛地互相连接而形成的复杂网络系统，神经网络模型反映了人脑功能的许多基本特征，是一个高度复杂的非线性动力学习系统。神经网络模型具有大规模并行、分布式存储和处理、自组织、自适应和自学能力，特别适合处理需要同时考虑许多因素和条件的、不精确和模糊的信息处理问题。
可选地，终端可以是将当前工况参数作为输入参数，输入到预先基于当前工况参数对应的多个历史运行参数和历史运行参数对应的第一能效比进行训练得到的神经网络模型中，而神经网络模型则会输出一组运行参数作为输出值作为当前工况参数对应的目标运行参数。
例如，在当前工况参数中，当前环境参数包括当前室外温度和当前室内温度，当前设定参数包括当前设定温度时，神经网络模型对应输出的目标运行参数可以是包括压缩机频率、电子膨胀阀开度、内风机转速和外风机转速；在当前工况参数中，当前环境参数包括当前室外温度、当前室内温度、当前室外湿度和当前室内湿度，当前设定参数包括当前设定温度和当前设定湿度时，神经网络模型对应输出的目标运行参数可以是包括压缩机频率、电子膨胀阀开度、内风机转速、外风机转速和加湿频率。
需要说明的是，神经网络模型中的历史运行参数，可以是由工程师输入到神经网络模型中的训练样本中的历史运行参数，也可以是由神经网络模型基于训练样本自学习得到的新的训练样本中的历史运行参数。
这样，可以通过利用预先构建的神经网络模型实现空调器的能效比的自动寻优，并可进一步通过神经网络模型的不断迭代训练，实现空调器的能效比的优化。
步骤S30、根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数。
可选地，终端在得到当前工况参数对应的目标运行参数后，即可控制空调器以目标运行参数运行。
需要说明的是，终端可以是在检测到空调器的设定参数变更时，则执行步骤S10至S30；终端也可以是在检测到所有当前环境参数中的任一当前环境参数的变化量超出预设变化量范围时，则执行步骤S10至S30，其中，温度值对应的预设变化量范围可以是0-1℃，湿度值对应的预设变化量可以是0-5％。
在一实施例中，获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；根据所述当前工况参数对应的多个历史运行参数，以及根据 各个所述历史运行参数对应的第一能效比获取目标运行参数；控制所述空调器以所述目标运行参数运行。这样，通过根据运行参数对应的能效比寻找空调器运行参数，并控制空调器运行，解决了如何优化空调器的能效比的问题。
在第二实施例中，如图3所示，在上述图2所示的实施例基础上，所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤包括：
步骤S21、将所述当前工况参数输入神经网络模型，以得到所述目标运行参数，其中，所述神经网络模型根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比进行训练或更新。
本实施例中，终端设置或者连通有神经网络模型。当将当前工况参数作为输入值输入至预先训练好的神经网络模型中，神经网络模型可以将一组运行参数作为输出值输出作为当前工况参数对应的目标运行参数。
可选地，所述神经网络模型根据多个训练样本训练或者更新，每一训练样本包括历史工况参数、所述历史工况参数对应的所述历史运行参数和所述第一能效比。
可选地，参见图9，预先构建的神经网络模型采用五层的全连接神经网络，隐含层激活函数为relu，输出层采用线性激活函数。采用遍历所有工况参数数值组合的多个数据组(此时的工况参数即为历史工况参数)，及数据组对应的历史运行参数、历史运行参数对应的第一能效比作为神经网络模型的训练样本(每一训练样本中的历史工况参数均有对应的历史运行参数，每一历史运行参数均有其对应的第一能效比)，训练神经网络至损失函数不再下降后停止迭代。
需要说明的是，所有训练样本中，同一历史工况参数的训练样本具有至少一个；历史运行参数对应的第一能效比即为空调器在历史运行参数对应的历史工况参数下，以历史运行参数运行时，空调器所输出的能效比。
例如，若遍历所有工况参数数值组合具有90个数据组，则将这90个数据组作为训练样本加入至神经网络模型中迭代训练。可选地，采用另外21组包含全工况范围的训练样本来验证神经网络模型的表现，若发现21条训练样本上的预测值与实际值的偏差在±5％，即说明该神经网络模型为一个可行的训练好的神经网络模型。
应当理解的是，历史工况参数包括历史环境参数和历史设定参数，其中，历史环境参数包括历史室外温度、历史室内温度、历史室外湿度和历史室内湿度中的至少一个；所述历史设定参数包括历史设定温度和历史设定湿度中的至少一个。历史运行参数包括压缩机频率、电子膨胀阀开度、内风机转速、外风机转速和加湿频率中的至少一个。
需要说明的是，神经网络模型基于输入到模型中的训练样本进行迭代训练的过程中，还可以产生新的训练样本，自学习产生的每一训练样本中同样具有历史工况参数、历史运行参数和第一能效比。
可选地，终端在获取到当前工况参数后，将当前工况参数作为输入值输入至神经网络模型中，神经网络模型可在训练样本中查询与当前工况参数对应的所有历史工况参数(这些历史工况参数与当前工况参数均相同，或者与当前工况参数处于同一数值范围内)，并将这些历史工况参数中，对应的第一能效比最大的历史运行参数作为输出值输出作为当前工况参数对应的目标运行参数。
需要说明的是，在历史工况参数与当前工况参数对应的所有训练样本中，同一历史工况参数的训练样本具有至少一个，而与历史工况参数与当前工况参数对应的训练样本具有至少两个。在同一历史工况参数的训练样本具有多个时，则历史工况参数相同的训练样本中的历史运行参数可以不同，例如，第一训练样本中具有历史工况参数A和历史运行参数B，而第二训练样本中具有历史工况参数A和历史运行参数C，则第一训练样本和第二训练样本相比，历史工况参数相同，而历史运行参数不同，当然历史运行参数B对应的能效比与历史运行参数C亦不同。
在一实施例中，将所述当前工况参数输入神经网络模型，以得到所述目标运行参数，其中，所述神经网络模型根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比进行训练或更新。这样，通过训练神经网络模型，基于当前工况参数寻找最优能效比下的空调器的运行参数，并控制空调器运行，解决了如何优化空调器的能效比的问题。
在第三实施例中，如图4所示，在上述图2至图3的实施例基础上，所述获取空调器的当前工况参数的步骤之后，还包括：
步骤S40、检测与所述当前工况参数对应的历史运行参数的数量。
步骤S50、所述历史运行参数的数量小于预设数量，根据所述当前工况参数和预设的工况参数与运行参数对应关系获取目标运行参数。
本实施例中，终端在获取到当前工况参数后，可查询与当前工况参数对应的历史运行参数的数量，或者查询神经网络模型中现有的训练样本中，历史工况参数与当前工况参数的训练样本的数量，作为与当前工况参数对应的历史运行参数的数量。
可选地，终端再获取到与当前工况参数对应的历史运行参数的数量后，检测与当前工况参数对应的历史运行参数的数量是否大于或者等于预设数量，并在检测到与当前工况参数对应的历史运行参数的数量小于预设数量时，则获取预设的工况参数与运行参数对应关 系，然后根据所述当前工况参数和预设的工况参数与运行参数对应关系获取所述当前工况参数对应的目标运行参数；此外，终端在检测到与当前工况参数对应的历史运行参数的数量大于或者等于预设数量时，则执行所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤(即步骤20)。
需要说明的是，所述预设数量可以是根据实际情况需要设置，如10-50个，可选为30个。
可选地，当前工况参数对应的历史运行参数的数量小于预设数量时，表征为神经网络模型当前的历史运行参数与当前工况参数对应的训练样本的数量还不够多，神经网络模型不能达到空调器的能效比的自动寻优的要求，因此，此时需要根据所述当前工况参数和预设的工况参数与运行参数对应关系获取目标运行参数。
可选地，预设的工况参数与运行参数对应关系，可以是由工程师根据工作经验制定的具有多种工况参数与运行参数的对应关系的空调器控制表得到。终端可根据当前工况参数，在空调器控制表中查询数值与当前工况参数对应的工况参数(数值与当前工况参数相同，或者数值与当前工况参数处于同一数值范围)，然后获取查询的得到的工况参数对应的运行参数，作为目标运行参数。
在一实施例中，检测与所述当前工况参数对应的历史运行参数的数量；所述历史运行参数的数量小于预设数量，根据所述当前工况参数和预设的工况参数与运行参数对应关系获取目标运行参数。这样，提高了获取空调器的目标运行参数时的稳定性。
在第四实施例中，如图5所示，在上述图2至图4的实施例基础上，所述控制所述空调器以所述目标运行参数运行的步骤之后，还包括：
步骤S60、确定所述目标运行参数对应的第二能效比。
步骤S70、根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本。
步骤S80、将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型。
本实施例中，终端可以是将所述当前工况参数输入神经网络模型，以得到所述目标运行参数；也可以是根据所述当前工况参数和预设的工况参数与运行参数对应关系获取目标运行参数。终端在获取到目标运行参数后，控制空调器以目标运行参数运行。
可选地，在空调器以目标运行参数运行后，终端可以是在室内环境参数达到目标运行参数后，检测空调器的能效参数，所述能效参数包括出风干球温度、出风湿球温度、室内干球温度、室内湿球温度和空调器的当前功率。
可选地，终端进一步根据目标运行参数和能效参数确定目标运行参数对应的第二能效 比。
可选地，终端可先根据目标运行参数和能效参数确定空调器当前的制冷量或者制热量。然后根据制冷量或者制热量，以及根据空调器的当前功率确定目标运行参数对应的第二能效比。需要说明的是，在空调器运行于制冷模式时，则确定制冷量；在空调器运行于制热模式时，则确定制热量。
可选地，第二能效比的计算公式如下：
Figure PCTCN2019109080-appb-000001
其中，E为第二能效比，Q为制冷量或者制热量，W为空调器的当前功率。
可选地，制冷量或制热量Q的计算公式如下:
Figure PCTCN2019109080-appb-000002
其中，根据出风干球温度、出风湿球温度、室内干球温度、室内湿球温度，通过查找焓湿图得到出风焓值ha和进风焓值hb；Wn为室内空气湿度，通过出风干球和出风湿球确定；V1为空调器的出风量；V2为出风比容，V2可以通过查找焓湿图确定。
可选地，空调器的出风量的计算公式如下：
Figure PCTCN2019109080-appb-000003
其中，n为内风机转速，N为外风机转速，V3为额定风量。
可选地，终端在得到第二能效比后，可根据当前工况参数、目标运行参数和第二能效比生成样本，其中，在样本中当前工况参数更变为历史工况参数，目标运行参数更变为历史运行参数，第二能效比更变为第一能效比。
可选地，终端在生成新的样本后，可将样本作为训练样本加入到神经网络模型中，以更新神经网络模型。
其中，在目标运行参数是根据所述当前工况参数和预设的工况参数与运行参数对应关系得到时，通过生成该目标运行参数对应的样本，随着神经网络模型的训练样本的不断更新迭代，即可实现与当前工况参数对应的训练样本的数量大于或者等于预设数量，这样，这样完成神经网络模型训练，使训练完成后的神经网络模型达到空调器的能效比的自动寻优的要求；在目标运行参数是根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比得到时，通过生成该目标运行参数对应的样本，作为训练样本加入至神经网络模型中训练时，亦可实现神经网络模型的不断优化。
在第五实施例中，如图6所示，在上述图2至图5的实施例基础上，所述确定所述目标运行参数对应的第二能效比的步骤之后，还包括：
步骤S61、获取所述第二能效比对应的制冷量或者制热量。
步骤S71、所述制冷量处于预设制冷量区间内，或者所述制热量处于预设制热量区间内，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
本实施例中，终端可根据目标运行参数和能效参数确定空调器当前的制冷量或者制热量。需要说明的是，在空调器当前运行于制冷模式时，则确定空调器当前的制冷量；在空调器当前运行于制热模式时，则确定空调器当前的制热量。
需要说明的是，由于第二能效比与目标运行参数对应，因此，获取根据目标运行参数和能效参数确定得到的制冷量或者制热量，即为第二能效比对应的制冷量或者制热量。
可选地，终端得到当前工况参数后，可在空调器控制表中查询数值与当前工况参数对应的工况参数(数值与当前工况参数相同，或者数值与当前工况参数处于同一数值范围)，然后获取查询的得到的工况参数对应的运行参数，然后确定根据该工况参数和运行参数确定该工况参数对应的制冷量或者制热量。将该制冷量前后加减预设值，得到预设制冷量区间的两个大小端值，这样即可得到预设制冷量区间；将该制热量前后加减预设值，得到预设制热量区间的两个大小端值，这样即可得到预设制热量区间。
需要说明的是，所述预设值的取值范围可以是50W-200W，可选为100W。
可选地，在空调器当前运行于制冷模式时，当第二能效比对于的制冷量处于预设制冷量区间内，判定当前满足样本的生成条件，并执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤(即步骤70)。而在第二能效比对于的制冷量不处于预设制冷量区间时，则可舍弃这组数据，不再将当前工况参数、所述目标运行参数和所述第二能效比生成样本。
可选地，在空调器当前运行于制热模式时，当第二能效比对于的制热量处于预设制热量区间内，判定当前满足样本的生成条件，并执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤(即步骤70)。而在第二能效比对于的制热量不处于预设制热量区间时，则可舍弃这组数据，不再将当前工况参数、所述目标运行参数和所述第二能效比生成样本。
可选地，基于预设制冷量区间和预设制热量区间，神经网络模型的训练过程也可以是：当有一组历史工况参数输入到神经网络模型后，对空调器所有可能的控制参数按照参数的数值范围和步长遍历所有可能的取值，得到所有可能的控制参数组合，作为该历史工况参数对应的历史运行参数。对于该历史运行参数每一组可能控制参数的组合，依次输入到神经网络模型，计算按该组参数运行后得到的制冷量或者制热量为Q，当制冷量处于预设制冷量区间内，或者制热量处于预设制热量区间内时，则将该历史工况参数，及历史工况参 数对应的制冷量处于预设制冷量区间内，或者制热量处于预设制热量区间内的历史运行参数，及该历史运行参数对应的第一能效值组成一个合格的训练样本，将该训练样本连同神经网络的输出值一起作为一条数据记录在列表list中。
在一实施例中，获取所述第二能效比对应的制冷量或者制热量；所述制冷量处于预设制冷量区间内，或者所述制热量处于预设制热量区间内，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。这样，可以保证神经网络模型输出的目标运行参数所能达到的制冷量或者制热量能够满足空调器正常的制冷或制热需求。
在第六实施例中，如图7所示，在上述图2至图6的实施例基础上，所述控制所述空调器以所述目标运行参数运行的步骤之后，还包括：
步骤S62、获取所述空调器的压缩机工况参数。
步骤S72、所述压缩机工况参数小于预设阈值，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
本实施例中，在空调器以目标运行参数运行后，终端可以是在室内环境参数达到目标运行参数后，检测并获取空调器的压缩机工况参数；终端也可以是获取由神经网络模型预测得到的，空调器在当前工况参数对应的环境下以目标运行参数运行时，空调器的压缩机的压缩机工况参数。
可选地，所述压缩机工况参数包括压缩机排气温度、压缩机电流和压缩机冷媒压力中的至少一个。需要说明的是，不同类型的压缩机工况参数对应的预设阈值不同，终端在检测到压缩机工况参数小于压缩机工况参数对应的预设阈值时，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。以下以压缩机工况参数为压缩机排气温度为例进行说明。
可选地，终端检测到获取的压缩机排气温度小于压缩机排气温度对应的预设阈值(即预设温度)时，则判定在空调器以当前的目标运行温度运行时，空调器的负载处于合理的运行范围内，即压缩机处于稳定运行的状态，这时，可执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤(即步骤70)；在终端检测到获取的压缩机排气温度大于或者等于压缩机排气温度对应的预设阈值时，则判定在空调器以当前的目标运行温度运行时，空调器的负载并未处于合理的运行范围内，即压缩机处于不稳定运行的状态，这时，则可舍弃这组数据，不再将当前工况参数、所述目标运行参数和所述第二能效比生成样本。
可选地，压缩机排气温度对应的预设阈值的取值范围可以是90℃-100℃，可选为95℃。
可选地，基于预设制冷量区间或预设制热量区间，以及压缩机工况参数和对应的预设阈值，神经网络模型的训练过程也可以是：当有一组历史工况参数输入到神经网络模型后，对空调器所有可能的控制参数按照参数的数值范围和步长遍历所有可能的取值，得到所有可能的控制参数组合，作为该历史工况参数对应的历史运行参数。对于该历史运行参数每一组可能控制参数的组合，依次输入到神经网络模型，计算按该组参数运行后得到的制冷量或者制热量为Q，当制冷量处于预设制冷量区间内，或者制热量处于预设制热量区间内时，进一步判断这组数据中，历史运行参数对应的压缩机工况参数是否小于预设阈值，并在压缩机工况参数小于预设阈值时，则将该历史工况参数，及历史工况参数对应的制冷量处于预设制冷量区间内，或者制热量处于预设制热量区间内的历史运行参数，及该历史运行参数对应的第一能效值组成一个合格的训练样本，将该训练样本连同神经网络的输出值一起作为一条数据记录在列表list中。
在一实施例中，获取所述空调器的压缩机工况参数；所述压缩机工况参数小于预设阈值，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。这样，可以保证空调器在以神经网络模型输出的目标运行参数运行时，空调器的压缩机能够稳定运行。
在第七实施例中，如图8所示，在上述图2至图7的实施例基础上，所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤之后，还包括：
步骤S81、所述样本满足预设条件，则执行所述将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型的步骤。
本实施例中，可以是一生成有新的样本，即将该样本作为训练样本加入到神经网络模型中，更新神经网络模型，进一步根据新的训练样本和原训练样本一起训练神经网络模型；也可以是在样本满足预设条件时，再将所有未加入到神经网络模型中的样本作为训练样本，加入到神经网络模型中，更新神经网络模型，进一步根据新的训练样本和原训练样本一起训练神经网络模型。
所述预设条件包括以下任一个：所述样本的总数与第一预设值之间的比值大于参考值；所述样本的总数和所述训练样本的总数之和与第二预设值之间的比值大于参考值。
可选地，预设条件为所述样本的总数与第一预设值之间的比值大于参考值时，条件公式如下：
Figure PCTCN2019109080-appb-000004
其中，n1为未加入到神经网络模型中的样本的总数，X为第一预设值，N为参考值，神经网络模型每更新一次，参考值累计加一。
可选地，可选地，预设条件为所述样本的总数和所述训练样本的总数之和与第二预设值之间的比值大于参考值时，条件公式如下：
Figure PCTCN2019109080-appb-000005
其中，n1为未加入到神经网络模型中的样本的总数，n2为神经网络模型中现存的训练样本的总数，Y为第一预设值，N为参考值，神经网络模型每更新一次，参考值累计加一。
需要说明的是，所述将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型的步骤(即步骤80)每执行一次，所述参考值累计加一；所需参考值的初始值可选为0值；所述第一预设值的取值范围可选为1-1000，可取值为100；所述第二预设值的取值范围可选为1-1000，可取值为100。
可选地，所述预设条件可表征为未加入到神经网络模型中的样本的数量达到了一定数量。
可选地，终端可以是每生成一个新的样本后，则将新的样本存入数据库，当数据库中的样本达到一定数量后(例如数据库中样本满足(n1/100)>N，N随着神经网络模块跟新一次进行累加，N＝N+1，初始值可设置为0)，对神经网络进行更新一次，利用最新的神经网络模型替代上一次的神经网络模型。
在一实施例中，所述样本满足预设条件，则执行所述将所述样本作为所述训练样本加入到所述神经网络模型中，以更新所述神经网络模型的步骤。这样，可以避免神经网络模型的频繁更新，可以起到节约系统资源的作用。
在第八实施例中，如图10所示，在上述图2至图8的实施例基础上，所述空调器的控制方法还包括：
步骤S100，定时获取室内所需的制冷/制热量；
在本实施例中，可根据当前室内温度及室内目标温度的差值，获取室内所需的制冷/制热量，在空调按照当前运行参数运行预设时长后，再检测当前室内温度，从而获取室内所需的制冷/制热量，避免因空调按照当前运行参数运行时长过短，导致获取到的室内所需制冷/制热量不准确，也可以避免空调刚开始按照当前运行参数运行时波动太大，导致获取的所需制冷/制热量不准确；所述预设时长大于0.5h，如可为1h、1.5h等。
步骤S200，根据所需制冷/制热量以及预设的神经网络模型获取空调的目标运行参数，所述预设的神经网络模型根据样本数据库中的样本数据训练生成；
在本实施例中，所述神经网络模型包括输入层，隐含层及输出层，所述输入层与隐含层之间设置有连接权值，所述隐含层与所述输出层之间也设置有连接权值，所述隐含层还 设置有隐含层阈值，所述输出层设置有输出层阈值。
所述样本数据库中包括多组样本数据，每个样本数据包括输入参数及输出参数，所述输入参数包括环境参数及空调的运行参数，所述环境参数包括但不限于采集得到的作为定值输入的室内环境温度、室内环境湿度、室外环境温度、空气出风温度、空气出风湿度、室内风机转速、空调总运行电压、空调总运行电流；所述空调的运行参数包括但不限于作为可调变量输入的压缩机频率、室内风机转速、室外风机转速、电子膨胀阀开度。
在所述样本数据库的形成阶段，包括如下步骤：
(1)空调开启时，按照预设的初始的运行参数运行，该运行参数但不限于包括压缩机频率、室内风机转速、室外风机转速、电子膨胀阀开度等；
(2)在空调的运行过程中，采集环境参数，所述环境参数包括但不限于室内环境温度、室内环境湿度、室外环境温度、空气出风温度、空气出风湿度、室内风机转速、空调运行电压、空调总运行电流；
(3)将采集到的环境参数及初始运行参数导入到样本数据库中，并计算空调的制冷/制热量及能效比。
具体地，空调的制冷/制热量Q可由空调出口空气参数和室内环境参数计算得出。计算包括但不限于焓差法和热平衡法。例如焓差法可根据公式：Q＝m|h1(T1,θ1)-h2(T2,θ2)|计算得到，其中，m表示空调风量，h表示单位质量空气的焓值。
空调的能效比COP＝Q/P；而空调的消耗功率P＝IU；其中，COP表示能效比，Q表示空调的制冷/制热量，P表示空调的消耗功率；I表示空调总运行电流，U表示空调总运行电压。
(4)样本数据库将对应的空调运行参数(例如：压缩机频率、室内风机转速、室外风机转速、电子膨胀阀开度、室内环境温度T1、室内环境湿度θ室、室外环境温度T4等)作为网络的一组输入，将空调的制冷/制热量Q、能效比COP作为一组输出，输入和输出构成一组样本数据，将样本数据存储在数据库中；
(5)空调按照不同运行参数运行一次，将会得到一组新的样本数据，新的实际运行产生的新样本数据和历史样本数据(部分历史数据可来自空调产品开发阶段实验室测试数据)共同构成神经网络的样本数据库，空调实际运行越久，新的样本数据所占比重越大，根据所述样本数据库训练完毕的网络越能真实反映实际运行状态。
在所述神经网络模型的训练阶段，调取所述样本数据可中的样本数据对所述神经网络模型进行网络训练，在所述神经网络模型的输入层输入样本数据的输入参数，输入参数(从所述样本数据中获取)包括环境参数(如室内环境温度、室内环境湿度、室外环境温度等) 及运行参数(如压缩机频率、室内风机转速、室外风机转速、电子膨胀阀开度等)；在所述神经网络模型的输入层以输出参数(从同一组样本数据中获取)作为限制条件，输出参数包括空调的制冷/制热量Q和能效比COP，由于所述样本数据的输入参数以及输出参数是确定的，将所述输入参数输入到所述神经网络模型后，为了得到该确定的输出参数，就需要不断调整所述神经网络模型中的输入层与隐含层之间的连接权值、所述隐含层与所述输出层之间的连接权值；输入越多组的样本数据，所述连接权值也就被调整的越准确，也就实现了对所述神经网络模型的训练。
在所述目标运行的获取阶段，所述预设的神经网络模型是已经训练好/训练完毕的神经网络模型，室内所需制冷/制热量作为所述预设的神经网络模型输出参数，通过所述预设的神经网络模型结合优化算法如遗传算法、粒子群优化算法等反向求解所述预设的神经网络模型的输入参数，而该输入参数包括所述目标运行参数如压缩机频率、室内分机转速、室外风机转速、电子膨胀阀开度等。
步骤S300，控制空调按照所述目标运行参数运行，其中，在达到所述制冷/制热量的前提下，所述空调器以所述目标运行参数运行的能效比大于其它运行参数对应的能效比。
在本实施例中，由于所述目标运行参数是最初的求解条件为：在达到所述制冷/制热量的前提下，所述空调器以所述目标运行参数运行的能效比大于其它运行参数对应的能效比，故当空调按照所述目标运行参数运行时，空调的能效比将最大化，也即，所述空调在达到所需制冷/制热量时，空调的功率最小，此时空调最省电。
综上，本申请通过将定时获取室内所需的制冷/制热量，并将所述目标运行参数是最初的求解条件为：在达到所述制冷/制热量的前提下，所述空调器以所述目标运行参数运行时的能效比大于其它运行参数对应的能效比，通过预设的训练好的神经网络模型以及上述求解条件求得所述目标运行参数，并控制空调按照所述目标运行参数运行后，所述空调的能效比将最大化，也即，所述空调在达到所需制冷/制热量，空调的功率最小，此时空调最省电，从而实现耗电少的目的，还可以延缓空调各元件的损耗，延长空调的使用寿命。
进一步的，参照图11，本申请空调器的控制方法第九实施例，基于上述图10所示实施例，所述步骤S200之前，包括：
步骤S400，将样本数据库中的样本数据按照预设比例分为训练样本及测试样本，所述样本数据包括输入参数以及输出参数，所述输入参数包括作为定值输入的环境参数及作为可调变量输入的空调运行参数，所述输出参数包括制冷/制热量以及能效比；
步骤S500，根据所述训练样本对所述神经网络模型进行训练，更新所述神经网络模型中输入层与隐含层的连接权值、隐含层与输出层的连接权值、隐含层阈值及输出层阈值。
在本实施例中，将样本数据库中的样本数据按照预设比例分为训练样本及测试样本，例如，将样本数据总数量的85％作为训练样本，样本数据总数量剩下的15％作为测试样本，所述训练样本用于不断对所述神经网络模型进行迭代训练。
进一步的，参照图12，本申请空调器的控制方法第十实施例，基于上述图10-11所示实施例，所述步骤500之后包括：
步骤S600，根据所述测试样本判断所述神经网络模型是否训练完毕；
步骤S700，若所述神经网络模型训练完毕，则生成所述预设的神经网络模型；
若所述神经网络模型未训练完毕，则继续执行步骤S50。
在本实施例中，在占总数量85％的训练样本对所述神经网络模型进行训练后，通过所述测试样本判断所述神经网络模型是否训练好/完毕，从而保证所述神经网络模型的准确性。
进一步的，参照图13，本申请空调器的控制方法第十一实施例，基于上述图10-12所示实施例，所述步骤S600包括：
步骤S610，获取所述测试样本的输入参数输入所述神经网络模型后对应的输出层输出参数与目标参数的差值，所述目标参数包括目标制冷/制热量以及目标能效比；
步骤S620，若所述差值小于预设差值，则判定所述神经网络模型训练完毕。
在本实施例中，如前所述，所述神经网络模型后对应的输出层输出参数包括制冷/制热量及能效比，而所述目标参数包括目标制冷/制热量以及目标能效比，求取所述测试样本的输入参数对应的制冷/制热量(如2900)与目标制冷/制热量(如3000)的差值(此时差值为100)，若预设差值为120，此时，所述差值小于预设差值，则判定所述神经网络模型训练完毕；若所述差值大于预设差值，则所述神经网络模型还需通过训练样本继续进行迭代训练。
进一步的，参照图14，本申请空调器的控制方法第十二实施例，基于上述图10-13所示实施例，所述步骤S400之前，还包括：
步骤S800，对训练前的神经网络模型初始化，设定所述神经网络中的输入层节点数、隐含层节点数、输出层节点数以及输入层与隐含层的初始连接权值、隐含层与输出层的初始连接权值、初始隐含层阈值及初始输出层阈值。
在本实施例中，对训练前的神经网络模型初始化，例如，将初始连接权值、初始隐含层阈值及初始输出层阈值均赋值为1或者0，或者赋值随机数。
进一步的，参照图15，本申请空调器的控制方法第十三实施例，基于上述图10-14所示实施例，所述步骤S300之前，还包括：
步骤S900，判断所述目标运行参数是否处于空调预设的运行范围内；
若所述目标运行参数处于所述预设的运行范围内，则执行步骤S30。
在本实施例中，空调都有一定的运行范围，例如，空调压缩机的运行频率一般为10-98Hz，电子膨胀阀的开度为50-400度，室外风机转速为500-3000转/h，室内风机转速为400-1000转/h，故控制所述空调按照所述目标运行参数运行之前，需判断所述目标运行参数是否处于空调预设的运行范围内，若在所述运行范围内，则可空调执行所述目标运行参数，若不在所述运行范围内，则对所述目标运行参数重新求解或者对所述神经网络模型重新进行训练。
进一步的，参照图16，本申请空调器的控制方法第十四实施例，基于上述图10-15所示实施例，所述步骤S300之后，还包括：
步骤S110，在接收到停止优化指令时，停止执行根据所需制冷/制热量以及预设的神经网络模型获取空调的目标运行参数的步骤；
步骤S120，控制所述空调按照最后得到的目标运行参数运行。
在本实施例中，在接收到停止优化指令时，停止对所述目标运行参数的求解过程，避免一直求解消耗能量，同时，控制所述空调按照最后得到的目标运行参数运行，也能保证空调在一定时间内处于最省电的状态。
此外，本申请还提出一种空调器，所述空调器包括存储器、处理器及存储在存储器上并可在处理器上运行的空调器的控制程序，所述处理器执行所述空调器的控制程序时实现如以上实施例所述的空调器的控制方法的步骤。
此外，本申请还提出一种计算机可读存储介质，所述计算机可读存储介质包括空调器的控制程序，所述空调器的控制程序被处理器执行时实现如以上实施例所述的空调器的控制方法的步骤。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是可选实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在如上所述的一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是电视机，手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的可选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (18)

  1. 一种空调器的控制方法，其中，所述空调器的控制方法包括以下步骤：
    获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；
    根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数；以及
    控制所述空调器根据所述目标运行参数运行。
  2. 如权利要求1所述的空调器的控制方法，其中，所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤包括：
    将所述当前工况参数对应的多个历史运行参数中，所述第一能效比最大的历史运行参数作为所述目标运行参数。
  3. 如权利要求1所述的空调器的控制方法，其中，所述获取空调器的当前工况参数的步骤之后，还包括：
    检测与所述当前工况参数对应的历史运行参数的数量；
    所述历史运行参数的数量大于或者等于预设数量，则执行所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤；以及
    所述历史运行参数的数量小于预设数量，根据所述当前工况参数和预设的工况参数与运行参数对应关系获取目标运行参数。
  4. 如权利要求1所述的空调器的控制方法，其中，所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤包括：
    将所述当前工况参数输入神经网络模型得到所述目标运行参数，其中，所述神经网络模型根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比进行训练或更新。
  5. 如权利要求4所述的空调器的控制方法，其中，所述神经网络模型根据多个训练样本训练或者更新，每一训练样本包括历史工况参数、所述历史工况参数对应的所述历史运行参数和所述第一能效比，所述神经网络模型被配置为将所述历史工况参数与所述当前 工况参数对应的多个训练样本中，所述第一能效比最大的历史运行参数作为所述目标运行参数输出。
  6. 如权利要求5所述的空调器的控制方法，其中，所述控制所述空调器根据所述目标运行参数运行的步骤之后，还包括：
    确定所述目标运行参数对应的第二能效比；
    根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本；以及
    将所述样本作为所述训练样本加入到所述神经网络模型中，并更新所述神经网络模型。
  7. 如权利要求6所述的空调器的控制方法，其中，所述确定所述目标运行参数对应的第二能效比的步骤之后，还包括：
    获取所述第二能效比对应的制冷量或者制热量；以及
    所述制冷量处于预设制冷量区间内，或者所述制热量处于预设制热量区间内，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
  8. 如权利要求6所述的空调器的控制方法，其中，所述控制所述空调器以所述目标运行参数运行的步骤之后，还包括：
    获取所述空调器的压缩机工况参数；以及
    所述压缩机工况参数小于预设阈值，则执行所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤。
  9. 如权利要求6所述的空调器的控制方法，其中，所述根据所述当前工况参数、所述目标运行参数和所述第二能效比生成样本的步骤之后，还包括：
    所述样本满足预设条件，则执行所述将所述样本作为所述训练样本加入到所述神经网络模型中，并更新所述神经网络模型的步骤。
  10. 如权利要求9所述的空调器的控制方法，其中，所述预设条件包括以下任一个：
    所述样本的总数与第一预设值之间的比值大于参考值；
    所述样本的总数和所述训练样本的总数之和与第二预设值之间的比值大于参考值；以及
    其中，所述将所述样本作为所述训练样本加入到所述神经网络模型中，并更新所述神经网络模型的步骤每执行一次，所述参考值累计加一。
  11. 如权利要求5所述的空调器的控制方法，其中，所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤之前，还包括：
    将所述训练样本按照预设比例分为第一训练样本及第二训练样本；以及
    根据所述第一训练样本对所述神经网络模型进行网络训练，并更新所述神经网络模型 中输入层与隐含层的连接权值、隐含层与输出层的连接权值、隐含层阈值及输出层阈值。
  12. 如权利要求11所述的空调器的控制方法，其中，所述根据所述第一训练样本对所述神经网络模型进行网络训练的步骤之后，还包括：
    所述神经网络模型训练完毕，则执行所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤；以及
    所述神经网络模型未训练完毕，则继续执行所述根据所述第一训练样本对所述神经网络模型进行网络训练的步骤。
  13. 如权利要求12所述的空调器的控制方法，其中，所述空调器的控制方法还包括：
    获取所述第二训练样本的输入参数输入所述神经网络模型后对应的输出层输出参数与目标参数的差值；以及
    所述差值小于预设差值，则所述神经网络模型训练完毕。
  14. 如权利要求11所述的空调器的控制方法，其中，所述将所述训练样本按照预设比例分为第一训练样本及第二训练样本的步骤之前，还包括：
    对训练前的神经网络模型初始化，设定所述神经网络模型中的输入层节点数、隐含层节点数、输出层节点数以及输入层与隐含层的初始连接权值、隐含层与输出层的初始连接权值、初始隐含层阈值及初始输出层阈值。
  15. 如权利要求5所述的空调器的控制方法，其中，所述控制所述空调器根据所述目标运行参数运行的步骤之后，还包括：
    接收到停止优化指令，停止执行所述将所述当前工况参数输入神经网络模型得到所述目标运行参数的步骤；以及
    控制所述空调器根据最后得到的目标运行参数运行。
  16. 如权利要求1所述的空调器的控制方法，其中，所述根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数的步骤之后，还包括：
    所述目标运行参数处于所述空调器预设的运行范围内，则执行所述控制所述空调器根据所述目标运行参数运行的步骤。
  17. 一种空调器，其中，所述空调器包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的空调器的控制程序，所述空调器的控制程序被所述处理器执行时实现如下所述的空调器的控制方法的步骤：
    获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；
    根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数；以及
    控制所述空调器根据所述目标运行参数运行。
  18. 一种计算机可读存储介质，其中，所述计算机可读存储介质上存储有空调器的控制程序，所述空调器的控制程序被处理器执行时实现如下所述的空调器的控制方法的步骤：
    获取空调器的当前工况参数，所述当前工况参数包括当前环境参数和所述空调器的当前设定参数；
    根据所述当前工况参数对应的多个历史运行参数，以及根据各个所述历史运行参数对应的第一能效比获取目标运行参数；以及
    控制所述空调器根据所述目标运行参数运行。
PCT/CN2019/109080 2018-11-15 2019-09-29 空调器的控制方法、空调器及存储介质 WO2020098405A1 (zh)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN201811364391.XA CN109323425B (zh) 2018-11-15 2018-11-15 空调的控制方法、装置及可读存储介质
CN201811364391.X 2018-11-15
CN201910878640.5 2019-09-17
CN201910878640.5A CN110553353B (zh) 2019-09-17 2019-09-17 空调器的控制方法、空调器及存储介质

Publications (1)

Publication Number Publication Date
WO2020098405A1 true WO2020098405A1 (zh) 2020-05-22

Family

ID=70730761

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/109080 WO2020098405A1 (zh) 2018-11-15 2019-09-29 空调器的控制方法、空调器及存储介质

Country Status (1)

Country Link
WO (1) WO2020098405A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11573023B2 (en) 2021-03-07 2023-02-07 Mitsubishi Electric Research Laboratories, Inc. Controlling vapor compression system using probabilistic surrogate model

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0933089A (ja) * 1995-07-19 1997-02-07 Daikin Ind Ltd 空調機の運転制御装置
JP2006078009A (ja) * 2004-09-07 2006-03-23 Shimizu Corp 空調負荷予測方法
CN102620378A (zh) * 2011-01-27 2012-08-01 国际商业机器公司 数据中心节能控制方法和系统
CN103542489A (zh) * 2012-07-12 2014-01-29 珠海格力节能环保制冷技术研究中心有限公司 空调系统的控制方法、装置及系统
CN104091045A (zh) * 2014-06-16 2014-10-08 华南理工大学 一种基于bp神经网络的空调长效性能预测方法
CN106369739A (zh) * 2016-08-23 2017-02-01 海信（山东）空调有限公司 一种空调控制方法及空调控制器和空调系统
CN106765932A (zh) * 2016-12-14 2017-05-31 深圳达实智能股份有限公司 中央空调系统制冷主机的能效比预测方法及装置
CN107883536A (zh) * 2017-09-30 2018-04-06 珠海格力电器股份有限公司 空调设备的参数调整方法及装置、终端
CN109323425A (zh) * 2018-11-15 2019-02-12 广东美的制冷设备有限公司 空调的控制方法、装置及可读存储介质

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0933089A (ja) * 1995-07-19 1997-02-07 Daikin Ind Ltd 空調機の運転制御装置
JP2006078009A (ja) * 2004-09-07 2006-03-23 Shimizu Corp 空調負荷予測方法
CN102620378A (zh) * 2011-01-27 2012-08-01 国际商业机器公司 数据中心节能控制方法和系统
CN103542489A (zh) * 2012-07-12 2014-01-29 珠海格力节能环保制冷技术研究中心有限公司 空调系统的控制方法、装置及系统
CN104091045A (zh) * 2014-06-16 2014-10-08 华南理工大学 一种基于bp神经网络的空调长效性能预测方法
CN106369739A (zh) * 2016-08-23 2017-02-01 海信（山东）空调有限公司 一种空调控制方法及空调控制器和空调系统
CN106765932A (zh) * 2016-12-14 2017-05-31 深圳达实智能股份有限公司 中央空调系统制冷主机的能效比预测方法及装置
CN107883536A (zh) * 2017-09-30 2018-04-06 珠海格力电器股份有限公司 空调设备的参数调整方法及装置、终端
CN109323425A (zh) * 2018-11-15 2019-02-12 广东美的制冷设备有限公司 空调的控制方法、装置及可读存储介质

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11573023B2 (en) 2021-03-07 2023-02-07 Mitsubishi Electric Research Laboratories, Inc. Controlling vapor compression system using probabilistic surrogate model

Similar Documents

Publication Publication Date Title
WO2021063033A1 (zh) 空调能耗模型训练方法与空调系统控制方法
CN110553353B (zh) 空调器的控制方法、空调器及存储介质
CN109323425B (zh) 空调的控制方法、装置及可读存储介质
CN111537252B (zh) 空调器测试方法、装置、计算机设备及存储介质
CN104698843B (zh) 一种基于模型预测控制的数据中心节能控制方法
CN110726218B (zh) 空调器及其控制方法、装置、存储介质和处理器
WO2023030522A1 (zh) 一种数据中心空调系统诊断方法及装置
WO2020135490A1 (zh) 一种智能家居中动态构建冲突校验规则表的方法及装置
CN114279042B (zh) 一种基于多智能体深度强化学习的中央空调控制方法
WO2023065755A1 (zh) 空气源热泵机组的控制方法、装置和电子设备
WO2020098405A1 (zh) 空调器的控制方法、空调器及存储介质
CN114543274B (zh) 一种建筑中央空调温湿度优化控制方法及系统
CN108507126B (zh) 一种中央空调冷冻水的模糊控制方法、装置及空调
CN113268913B (zh) 一种基于pso-elm算法的智能建筑空调冷机系统运行优化方法
CN110873446A (zh) 控制空调的方法及装置、存储介质、处理器
CN113791538A (zh) 一种机房设备的控制方法、控制设备及控制系统
WO2023159977A1 (zh) 空调节能的控制方法、控制系统、电子设备和存储介质
WO2023159942A1 (zh) 空调节能的控制方法、控制系统、电子设备和存储介质
CN113531827B (zh) 变频空调控制方法、装置、电子设备和介质
CN115264768A (zh) 一种基于物联网的楼宇空调节能控制方法及系统
CN112781181A (zh) 空调群组节能控制方法及装置
US20200003436A1 (en) Systems and methods for evaporative cooling control
CN111274733A (zh) 数据处理方法、数据处理装置、空调系统和存储介质
Zhang et al. Rule-based optimal control for the cooling water subsystem of chiller plants
CN113747756B (zh) 数据中心机房供冷设备控制方法、装置、设备和供冷系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19885992

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 17/09/2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19885992

Country of ref document: EP

Kind code of ref document: A1