CN107860099A

CN107860099A - Air conditioner frosting detection method and device, storage medium and equipment

Info

Publication number: CN107860099A
Application number: CN201710829628.6A
Authority: CN
Inventors: 刘佰兰; 黄辉; 宋德超; 陈翀; 田涛
Original assignee: Gree Electric Appliances Inc of Zhuhai
Current assignee: Gree Electric Appliances Inc of Zhuhai
Priority date: 2017-09-15
Filing date: 2017-09-15
Publication date: 2018-03-30
Anticipated expiration: 2037-09-15
Also published as: CN107860099B

Abstract

The invention discloses a frosting detection method, a frosting detection device, a storage medium and equipment of an air conditioner, wherein the method comprises the following steps: establishing a corresponding relation between the operating parameters of the air conditioner and the frosting state of the condenser in the heating mode by utilizing the self-learning capability of the artificial neural network; acquiring current operating parameters of the air conditioner; and determining the current frosting state corresponding to the current operation parameters according to the corresponding relation. According to the scheme provided by the invention, the defects of poor frosting judgment accuracy, poor heating effect, poor user experience and the like in the prior art can be overcome, and the beneficial effects of good frosting judgment accuracy, good heating effect and good user experience are realized.

Description

Air conditioner frosting detection method and device, storage medium and equipment

Technical Field

The invention belongs to the technical field of defrosting, and particularly relates to a frosting detection method, a frosting detection device, a frosting detection storage medium and frosting detection equipment for an air conditioner, in particular to a method for identifying the frosting condition of an air conditioner condenser by using an artificial neural network algorithm, a device corresponding to the method, a computer readable form storage medium storing instructions corresponding to the method, and equipment capable of loading and executing the instructions corresponding to the method.

Background

An air conditioner (i.e., an air conditioner) may be a device that manually adjusts and controls parameters such as temperature, humidity, cleanliness, and speed of ambient air within a building/structure. After the air conditioner operates for a period of time in a heating state, the condenser of the outdoor unit is frosted, which affects the heating effect, and thus, the defrosting operation is performed on the condenser.

However, in the defrosting control strategy of the air conditioner in the industry at present, the defrosting is started when the heating operation of the air conditioner is fully performed for a period of time and the temperature point of a condenser pipeline reaches a certain degree based on certain preset rules and time of experience. However, this method has a limited accuracy in determining the frosting condition, and sometimes defrosting is started without starting frosting clearly, and sometimes defrosting is not started even after the frost has already been formed to a large thickness, and thus an ideal defrosting effect cannot be obtained.

In the prior art, the defects of poor frosting judgment accuracy, poor heating effect, poor user experience and the like exist.

Disclosure of Invention

The invention aims to provide a frosting detection method, a frosting detection device, a storage medium and equipment of an air conditioner, aiming at the defects, so as to solve the problem that the judgment accuracy of the frosting condition in the prior art is limited and achieve the effect of good frosting judgment accuracy.

The invention provides a frosting detection method of an air conditioner, which comprises the following steps: establishing a corresponding relation between the operating parameters of the air conditioner and the frosting state of the condenser in the heating mode by utilizing the self-learning capability of the artificial neural network; acquiring current operating parameters of the air conditioner; and determining the current frosting state corresponding to the current operation parameters according to the corresponding relation.

Optionally, wherein the operating parameters include: the system comprises environment parameters and/or working parameters and/or one-dimensional or more than two-dimensional arrays consisting of features extracted from the environment parameters and the working parameters according to a set rule; wherein the environmental parameters include: at least one of indoor and outdoor ambient temperature, user body surface temperature and indoor and outdoor ambient humidity; and/or, the operating parameters include: at least one of set temperature, set wind speed, set wind gear, condenser temperature and outdoor fan running current; and/or, the corresponding relation comprises: a functional relationship; the operation parameters are input parameters of the functional relation, and the frosting state is output parameters of the functional relation; determining a current frosting status corresponding to the current operating parameter, including: determining the frosting state corresponding to the operation parameter which is the same as the current operation parameter in the corresponding relation as the current frosting state; and/or when the corresponding relation comprises a functional relation, inputting the current operation parameter into the functional relation, and determining the output parameter of the functional relation as the current frosting state.

Optionally, establishing a corresponding relationship between an operation parameter of the air conditioner in the heating mode and a frosting state of the condenser includes: acquiring sample data for establishing a corresponding relation between the operating parameters and the frosting state; analyzing the characteristics and the rules of the operating parameters, and determining the network structure and the network parameters of the artificial neural network according to the characteristics and the rules; training and testing the network structure and the network parameters by using the sample data, and determining the corresponding relation between the operation parameters and the frosting state.

Optionally, acquiring sample data for establishing a correspondence between the operating parameter and the frosting status includes: collecting operation data and frosting states of the air conditioner in different use environments; analyzing the operating data, and selecting data related to the frosting state as the operating parameters by combining prestored expert experience information; and taking the frosting state and the data pair formed by the selected operation parameters as sample data.

Optionally, collecting operation data of the air conditioner in different use environments includes: acquiring the operating parameters of the air conditioner in a laboratory simulation environment, and/or acquiring the operating parameters collected by the Internet of things technology when the air conditioner is actually used by a user; and/or, the network architecture, comprising: at least one of a BP neural network, a CNN neural network, an RNN neural network, and a residual error neural network; and/or, the network parameters include: at least one of the number of input nodes, the number of output nodes, the number of hidden layers, the number of hidden nodes, an initial weight and a bias value.

Optionally, training the network structure and the network parameters includes: selecting a part of data in the sample data as a training sample, inputting the operation parameters in the training sample into the network structure, and training through an activation function of the network structure and the network parameters to obtain an actual training result; determining whether an actual training error between the actual training result and a corresponding frosting state in the training sample satisfies a set training error; determining that the training of the network structure and the network parameters is completed when the actual training error satisfies the set training error; and/or, testing the network structure and the network parameters, including: selecting another part of data in the sample data as a test sample, inputting the operation parameters in the test sample into the trained network structure, and testing by using the activation function and the trained network parameters to obtain an actual test result; determining whether an actual test error between the actual test result and a corresponding frosting condition in the test sample satisfies a set test error; and when the actual test error meets the set test error, determining that the test on the network structure and the network parameters is finished.

Optionally, training the network structure and the network parameters further includes: when the actual training error does not meet the set training error, updating the network parameters through an error energy function of the network structure; retraining through the activation function of the network structure and the updated network parameters until the retrained actual training error meets the set training error; and/or, testing the network structure and the network parameters, further comprising: and when the actual test error does not meet the set test error, retraining the network structure and the network parameters until the retrained actual test error is slower than the set test error.

Optionally, the method further comprises: determining whether the current frosting state reaches a set defrosting degree; when the current frosting state reaches the defrosting degree, entering a set defrosting mode, and/or initiating a prompt that the current frosting state reaches the defrosting degree; and/or determining the entering time and/or exiting time of the defrosting mode after the set defrosting mode needs to be entered according to the current frosting state; and/or, displaying and/or outputting at least one of a current operation mode of the air conditioner, the current operation parameter, the current frosting state, the entering time of a defrosting mode and the exiting time of the defrosting mode; and/or when a verification result that the current frosting state is not consistent with the actual frosting state is received, and/or the corresponding relation does not have the operation parameter which is the same as the current operation parameter, performing at least one maintenance operation of updating, correcting and relearning on the corresponding relation.

In accordance with the above method, another aspect of the present invention provides a frost formation detection apparatus for an air conditioner, including: the building unit is used for building a corresponding relation between the operating parameters of the air conditioner and the frosting state of the condenser in the heating mode by utilizing the self-learning capability of the artificial neural network;

the acquiring unit is used for acquiring the current operating parameters of the air conditioner;

and the determining unit is used for determining the current frosting state corresponding to the current operation parameters through the corresponding relation.

Optionally, wherein the operating parameters include: the system comprises environment parameters and/or working parameters and/or one-dimensional or more than two-dimensional arrays consisting of features extracted from the environment parameters and the working parameters according to a set rule; wherein the environmental parameters include: at least one of indoor and outdoor ambient temperature, user body surface temperature and indoor and outdoor ambient humidity; and/or, the operating parameters include: at least one of set temperature, set wind speed, set wind gear, condenser temperature and outdoor fan running current; and/or, the corresponding relation comprises: a functional relationship; the operation parameters are input parameters of the functional relation, and the frosting state is output parameters of the functional relation; the determining unit determines a current frosting state corresponding to the current operating parameter, and specifically includes: determining the frosting state corresponding to the operation parameter which is the same as the current operation parameter in the corresponding relation as the current frosting state; and/or when the corresponding relation comprises a functional relation, inputting the current operation parameter into the functional relation, and determining the output parameter of the functional relation as the current frosting state.

Optionally, the establishing unit establishes a correspondence between an operation parameter of the air conditioner in the heating mode and a frosting state of the condenser, and specifically includes: acquiring sample data for establishing a corresponding relation between the operating parameters and the frosting state; analyzing the characteristics and the rules of the operating parameters, and determining the network structure and the network parameters of the artificial neural network according to the characteristics and the rules; training and testing the network structure and the network parameters by using the sample data, and determining the corresponding relation between the operation parameters and the frosting state.

Optionally, the establishing unit obtains sample data for establishing a correspondence between the operating parameter and the frosting state, and specifically includes: collecting operation data and frosting states of the air conditioner in different use environments; analyzing the operating data, and selecting data related to the frosting state as the operating parameters by combining prestored expert experience information; and taking the frosting state and the data pair formed by the selected operation parameters as sample data.

Optionally, the establishing unit collects operation data of the air conditioner in different use environments, and specifically includes: acquiring the operating parameters of the air conditioner in a laboratory simulation environment, and/or acquiring the operating parameters collected by the Internet of things technology when the air conditioner is actually used by a user; and/or, the network architecture, comprising: at least one of a BP neural network, a CNN neural network, an RNN neural network, and a residual error neural network; and/or, the network parameters include: at least one of the number of input nodes, the number of output nodes, the number of hidden layers, the number of hidden nodes, an initial weight and a bias value.

Optionally, the training of the network structure and the network parameters by the establishing unit specifically includes: selecting a part of data in the sample data as a training sample, inputting the operation parameters in the training sample into the network structure, and training through an activation function of the network structure and the network parameters to obtain an actual training result; determining whether an actual training error between the actual training result and a corresponding frosting state in the training sample satisfies a set training error; determining that the training of the network structure and the network parameters is completed when the actual training error satisfies the set training error; and/or, the establishing unit tests the network structure and the network parameters, and specifically includes: selecting another part of data in the sample data as a test sample, inputting the operation parameters in the test sample into the trained network structure, and testing by using the activation function and the trained network parameters to obtain an actual test result; determining whether an actual test error between the actual test result and a corresponding frosting condition in the test sample satisfies a set test error; and when the actual test error meets the set test error, determining that the test on the network structure and the network parameters is finished.

Optionally, the establishing unit trains the network structure and the network parameters, and specifically includes: when the actual training error does not meet the set training error, updating the network parameters through an error energy function of the network structure; retraining through the activation function of the network structure and the updated network parameters until the retrained actual training error meets the set training error; and/or, the establishing unit tests the network structure and the network parameters, and specifically further includes: and when the actual test error does not meet the set test error, retraining the network structure and the network parameters until the retrained actual test error is slower than the set test error.

Optionally, the method further comprises: the determining unit is further used for determining whether the current frosting state reaches a set defrosting degree; when the current frosting state reaches the defrosting degree, entering a set defrosting mode, and/or initiating a prompt that the current frosting state reaches the defrosting degree; and/or the determining unit is further configured to determine, according to the current frosting state, entry time and/or exit time of the defrosting mode after a set defrosting mode needs to be entered; and/or the determining unit is further configured to display and/or output at least one of a current operation mode of the air conditioner, the current operation parameter, the current frosting state, an entering time of a defrosting mode, and an exiting time of the defrosting mode; and/or the determining unit is further configured to perform at least one maintenance operation of updating, correcting, and relearning the corresponding relationship through the establishing unit when a verification result that the current frosting state does not conform to the actual frosting state is received and/or the corresponding relationship does not have the same operation parameter as the current operation parameter.

In accordance with the above method, a further aspect of the present invention provides a storage medium comprising: the storage medium has stored therein a plurality of instructions; the plurality of instructions are used for loading and executing the frosting detection method of the air conditioner by the processor.

In accordance with the above method or apparatus, a further aspect of the present invention provides an apparatus comprising: a processor for executing a plurality of instructions; a memory to store a plurality of instructions; the plurality of instructions are stored by the memory, and are loaded and executed by the processor; alternatively, it comprises: the frosting detection device of the air conditioner is described above.

Optionally, the apparatus comprises: the air conditioner comprises an air conditioner body and/or an external control end, wherein the air conditioner body is used for controlling the air conditioner body; wherein, the external control end includes: at least one of a wireless communication module, a router, a server and a terminal.

According to the scheme, the self-learning function of the neural network is utilized to establish the corresponding relation between the operating parameters of the air conditioner and the frosting state of the condenser; according to the current operation parameters of the air conditioner, the current frosting state can be determined through the corresponding relation, the determination mode is reliable, and the determination result is good in accuracy.

Furthermore, according to the scheme of the invention, the self-learning function of the neural network is utilized, the acquired data are trained and learned, and the operation parameters of the air conditioner and the corresponding relation function between the sensor parameters and the frosting state of the condenser are mastered, so that the frosting state of the condenser of the air conditioner is recognized in real time, the accuracy of defrosting detection is improved, the heating effect of the whole air conditioner is improved, the heating efficiency is improved, and the comfort of a user is improved.

Furthermore, according to the scheme of the invention, the operation state rule of the frosting process of the air conditioner condenser is analyzed by utilizing the artificial neural network algorithm, the mapping rule between the frosting condition of the condenser and the operation parameter state of the air conditioner in the heating operation of the air conditioner is found through the self-learning and self-adaptive characteristics of the artificial neural network, the current frosting condition of the air conditioner can be effectively identified, so that an accurate judgment basis is provided for the start and the end of defrosting, and the judgment result is good in accuracy.

Furthermore, according to the scheme of the invention, the current frosting state corresponding to the current operation parameter is determined according to the corresponding relation between the operation parameter and the frosting state, the operation mode is simple and convenient, and the reliability of the operation result is high.

Therefore, according to the scheme of the invention, the corresponding relation between the frosting state of the condenser in the heating operation of the air conditioner and the operation parameters of the air conditioner is determined by utilizing a neural network algorithm, and the current frosting state of the condenser is further determined according to the current operation parameters of the air conditioner; the problem of the limited judgement accuracy of the frosting condition among the prior art is solved to, overcome the frosting among the prior art and judge the defect that the accuracy is poor, the effect of heating is poor and user experience is poor, realize that frosting judges that the accuracy is good, the good beneficial effect of heating effect and user experience is good.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.

Drawings

FIG. 1 is a schematic flow chart illustrating a frosting detection method of an air conditioner according to an embodiment of the present invention;

FIG. 2 is a schematic flow chart illustrating an embodiment of the method for establishing a corresponding relationship between an operating parameter of an air conditioner in a heating mode and a frosting state of a condenser according to the present invention;

FIG. 3 is a schematic flow chart diagram illustrating one embodiment of obtaining sample data for establishing a correspondence between the operating parameters and the frosting status in the method of the present invention;

FIG. 4 is a schematic flow chart illustrating an embodiment of training the network structure and the network parameters in the method of the present invention;

FIG. 5 is a schematic flow chart illustrating an embodiment of testing the network structure and the network parameters in the method of the present invention;

FIG. 6 is a flowchart illustrating an embodiment of retraining the network structure and the network parameters in the method of the present invention;

FIG. 7 is a flowchart illustrating an embodiment of determining whether the current frosting status reaches the defrosting level according to the method of the present invention;

FIG. 8 is a schematic flow chart illustrating an embodiment of verifying whether the current frosting status and the actual frosting status are consistent according to the method of the present invention;

FIG. 9 is a schematic structural diagram illustrating an embodiment of a frost formation detection apparatus of an air conditioner according to the present invention;

FIG. 10 is a schematic structural diagram of an embodiment of a first algorithm structure of an artificial neural network (e.g., BP neural network) in the method of the present invention;

FIG. 11 is a schematic structural diagram of an embodiment of a second artificial neural network algorithm (CNN convolutional neural network) structure in the method of the present invention;

FIG. 12 is a schematic diagram of an embodiment of a third artificial neural network algorithm (e.g., a residual neural network) in the method of the present invention;

FIG. 13 is a schematic structural diagram of an embodiment of an artificial neural network algorithm network structure IV in the method of the present invention;

FIG. 14 is a schematic diagram of an embodiment of an intelligent device (e.g., a terminal) for operating an artificial neural network algorithm in the apparatus of the present invention.

The reference numbers in the embodiments of the present invention are as follows, in combination with the accompanying drawings:

102-a building unit; 104-an obtaining unit; 106-determination unit.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

According to an embodiment of the present invention, a method for detecting frosting of an air conditioner is provided, as shown in fig. 1, which is a schematic flow chart of an embodiment of the method of the present invention. The frosting detection method of the air conditioner can comprise the following steps:

at step S110, a correspondence relationship between the operation parameters of the air conditioner in the heating mode and the frosting state of the condenser is established using the self-learning capability of the artificial neural network.

For example: by utilizing the self-learning function of the neural network, the operating parameters of the air conditioner and the corresponding relation function between the sensor parameters and the frosting state of the condenser are mastered by training and learning the acquired data.

The sensor parameters may refer to temperature and humidity sensors for detecting external environment conditions, and the like, which are already part of the operation parameters.

For example: the operating state rule of the frosting process of the air conditioner condenser is analyzed by utilizing an artificial neural network algorithm, and the mapping rule between the frosting condition of the condenser and the operating parameter state of the air conditioner in the heating operation of the air conditioner is found through the self-learning and self-adaptive characteristics of the artificial neural network.

For example: the method can utilize an artificial neural network algorithm to collect air conditioner operation data in a large number of different use environments (including but not limited to one or more of indoor and outdoor environment temperature and humidity, set temperature and the like), select operation state parameters and frosting state parameters of a plurality of air conditioners as sample data, learn and train the neural network, enable the neural network to fit the relationship between the air conditioner operation parameters and the frosting state by adjusting the weight between the network structure and the network nodes, and finally enable the neural network to accurately fit the corresponding relationship between the operation parameters of the air conditioner and the frosting state of the condenser.

In an alternative example, the operating parameters may include: the system comprises environment parameters and/or working parameters and/or one-dimensional or more than two-dimensional arrays consisting of features extracted from the environment parameters and the working parameters according to a set rule.

Optionally, the environmental parameter may include: at least one of indoor and outdoor ambient temperature, user body surface temperature, and indoor and outdoor ambient humidity.

Optionally, the operating parameters may include: at least one of set temperature, set wind speed, set wind gear, condenser temperature and outdoor fan running current.

For example: input parameters include, but are not limited to, one or more of the following: indoor and outdoor environment temperature, air conditioner set temperature, set wind speed (gear), condenser temperature, outdoor fan running current and the like. The input parameters are not only single parameters, but also one-dimensional or multi-dimensional arrays of the input parameters formed by extracting features according to a certain rule.

Therefore, the accuracy and the reliability of determining the corresponding relation between the operation parameters and the frosting state are improved through the operation parameters in various forms.

In an optional example, the correspondence may include: and (4) functional relation.

Optionally, the operating parameter is an input parameter of the functional relationship, and the frosting state is an output parameter of the functional relationship.

Therefore, the flexibility and convenience for determining the current frosting state can be improved through the corresponding relations in various forms.

In an alternative example, a specific process of establishing the correspondence between the operating parameters of the air conditioner in the heating mode and the frosting state of the condenser in step S110 may be further described with reference to a flowchart of an embodiment of the method of the present invention shown in fig. 2, where the correspondence between the operating parameters of the air conditioner in the heating mode and the frosting state of the condenser is established.

Step S210, sample data that can be used to establish a correspondence between the operating parameter and the frosting status is acquired.

Optionally, a specific process of acquiring sample data that can be used for establishing a correspondence between the operating parameter and the frosting state in step S210 may be further described with reference to a flowchart of an embodiment of acquiring sample data used for establishing a correspondence between the operating parameter and the frosting state in the method of the present invention shown in fig. 3.

And step S310, collecting operation data and frosting states of the air conditioner in different use environments.

In an alternative specific example, the collecting operation data of the air conditioner in different use environments in step S310 may include: and acquiring the operating parameters of the air conditioner in a laboratory simulation environment.

In an optional specific example, the collecting operation data of the air conditioner in different use environments in step S310 may further include: and acquiring the operating parameters collected by the Internet of things technology when the user actually uses the air conditioner.

For example: data collection: collecting the self running state parameters of the air conditioner in different use environments and the corresponding frosting condition. The specific collection mode includes, but is not limited to, modes of collecting operation parameters of the air conditioner in a laboratory simulation environment, collecting operation parameters of the air conditioner when the air conditioner is used by an actual user through the internet of things technology, and the like.

Therefore, the operation data are collected through multiple ways, the quantity of the operation data is increased, the learning capacity of the artificial neural network is improved, and the accuracy and the reliability of the determined corresponding relation are improved.

And step S320, analyzing the operation data, and selecting data related to the frosting state as the operation parameters by combining with prestored expert experience information (for example, selecting parameters influencing the frosting state of the air conditioner as input parameters, and using the frosting state as output parameters).

For example: by analyzing the data and combining expert experience knowledge, parameters influencing the frosting state of the air conditioner are selected as input parameters, and the frosting state is taken as output parameters.

And step S330, taking the frosting state and the data pair formed by the selected operation parameters as sample data.

For example: and using part of the obtained input and output parameter pairs as training sample data and using part of the obtained input and output parameter pairs as test sample data.

Therefore, the collected operation data are analyzed and processed to determine the operation parameters, so that sample data is obtained, the operation process is simple, and the reliability of the operation result is high.

Step S220, analyzing the characteristics and the rules of the operation parameters, and determining the network structure and the network parameters of the artificial neural network according to the characteristics and the rules.

For example: according to the data characteristics and the rules of the data characteristics, which affect the operation of the air conditioner, such as different set temperatures, environment temperatures, body surface temperatures and the like, the basic structure of the network, the number of input and output nodes of the network, the number of network hidden layers, the number of hidden nodes, the initial weight of the network and the like can be preliminarily determined.

Optionally, the network structure may include: at least one of a BP neural network, a CNN neural network, an RNN neural network, and a residual neural network.

Optionally, the network parameter may include: at least one of the number of input nodes, the number of output nodes, the number of hidden layers, the number of hidden nodes, an initial weight and a bias value.

For example: fig. 10 shows a schematic structural diagram of the BP neural network, and in actual application, the number of nodes of the input layer, the hidden layer, the output layer, and the number of hidden layers can be adjusted as needed.

For example: the multilayer convolution network is to continuously extract and combine low-dimensional features to obtain higher-dimensional features, so that the higher-dimensional features can be used for classification or correlation tasks. For example: the schematic diagram of the multilayer convolutional network structure is shown in fig. 11, and the network structure can be adjusted according to actual conditions in practical application.

For example: fig. 12 shows a schematic diagram of a residual neural network structure, and the network structure can be adjusted according to actual situations in practical application.

For example: when the CNN network is debugged, the network performance cannot be improved by deepening the network layer number and changing the size of the convolution kernel. The residual block is added, so that the data before and after connection can be better, and the feature expression capability is enhanced, so that the learning capability of the convolutional network can be enhanced. As shown in fig. 13, the input of a certain segment of neural network is x, the desired output is h (x), and after the input x is input to the output as an initial structure, the target to be learned becomes f (x) ═ h (x) -x.

For example: the artificial neural network used in the scheme is not limited to a certain network structure, and may be a classical artificial neural network such as a BP neural network, a high-level artificial neural network, or a deep learning network such as a CNN (convolutional neural network) and an RNN (recurrent neural network), and the specific scheme may be selected according to an actual application scenario.

Therefore, through the artificial neural network structures and the network parameters in various forms, the flexibility of selecting the network structures can be improved, and the convenience and the reliability of determining the corresponding relation between the operation parameters and the frosting state can be improved.

Step S230, using the sample data, training and testing the network structure and the network parameters, and determining the corresponding relationship between the operation parameters and the frosting status.

For example: and training and testing the sample data to obtain a training result and a testing result. Selecting sample data of which the training result and the test result both meet set requirements, and determining the corresponding relation between the operating parameters and the frosting state according to the selected sample data.

For example: after the network design is completed, training the network by using training sample data. The training method can be adjusted according to the actual network structure and the problems found in training.

Therefore, by collecting the operation data, sample data is selected, training and testing are carried out based on the sample data, the corresponding relation between the operation parameters and the frosting state is determined, and convenience and accuracy in judging the frosting state are improved.

Optionally, a specific process of training the network structure and the network parameters in step S230 may be further described with reference to a flowchart of an embodiment of training the network structure and the network parameters in the method of the present invention shown in fig. 4.

Step S410, selecting a part of data in the sample data as a training sample, inputting the operation parameters in the training sample into the network structure, and training through an activation function of the network structure and the network parameters to obtain an actual training result.

For example, ⑴ imports input data x, counts according to activation function, initialized weight and offsetCalculating the actual output a (x) of the network, i.e. a (x) is 1/(1+ e)^-z) Wherein Z ═ W_k*x+b_l。

Step S420, determining whether an actual training error between the actual training result and a corresponding frosting state in the training sample satisfies a set training error. For example: taking the respective frosting status in the sample data as an expected training result.

For example: judging whether the expected output y (x) and the actual output a (x) of the network meet the output precision requirement, namely: | (|) (x) a (x) | <e, e is the target minimum error.

Step S430, when the actual training error satisfies the set training error, determining that the training of the network structure and the network parameters is completed.

For example: the training is ended if the desired output y (x) of the network and the actual output a (x) meet the accuracy requirement.

Therefore, the training samples are used for training the selected network structure and the selected network parameters, so that the more reliable network structure and network parameters can be obtained, and the accuracy and the reliability of determining the corresponding relation between the operation parameters and the frosting state are improved.

More optionally, a specific process of retraining the network structure and the network parameters in the step S230 when training the network structure and the network parameters may be further described with reference to a flowchart of an embodiment of retraining the network structure and the network parameters in the method of the present invention shown in fig. 6.

And step S610, when the actual training error does not meet the set training error, updating the network parameters through an error energy function of the network structure.

Step S620, retraining through the activation function of the network structure and the updated network parameters until the retrained actual training error meets the set training error.

For example: if the expected output y (x) and the actual output a (x) of the network are not satisfied, the weight W of the network is updated according to the following mode_kOffset b from_l：

C (w, b) is the error energy function (taking the standard deviation function as an example), n is the total number of training samples, and the summation is performed over the total training samples x:

updating the weight of each layer:

updating each layer bias:

wherein:

W_kas an initial weight value, the weight value,is the partial derivative of the error energy function to the weight, b_lIn order to be the initial bias, the bias is,is the partial derivative of the error energy function to the bias;until the output error of the network is less than e (target minimum error).

Therefore, the network parameters are modified and retrained when the training errors are large, so that a more accurate and reliable network structure can be obtained, and a more accurate and reliable corresponding relation can be obtained.

Optionally, a specific process of testing the network structure and the network parameters in step S230 may be further described with reference to a flowchart of an embodiment of testing the network structure and the network parameters in the method of the present invention shown in fig. 5.

Step S510, selecting another part of the data in the sample data as a test sample, inputting the operation parameters in the test sample into the trained network structure, and performing a test with the activation function and the trained network parameters to obtain an actual test result.

For example: and after the network training is finished, the network is positively tested by using the test sample.

Step S520, determining whether an actual test error between the actual test result and a corresponding frosting status in the test sample (i.e. the frosting status in the sample data is taken as an expected output parameter) satisfies a set test error.

Step S530, when the actual test error satisfies the set test error, determining that the test on the network structure and the network parameter is completed.

For example: and if the test error meets the requirement, finishing the network training test.

Therefore, the reliability of the network structure and the network parameters is further verified by using the test sample for testing the network structure and the network parameters obtained by training.

More optionally, in the step S230, in testing the network structure and the network parameter, the method may further include: and when the actual test error does not meet the set test error, retraining the network structure and the network parameters until the retrained actual test error is slower than the set test error.

For example: and when the test error does not meet the requirement, repeating the steps and retraining the network.

Therefore, the network structure is retrained to be retested when the test error is large, so that the network structure which is more accurate and reliable is obtained, and the accuracy of determining the frosting state is improved.

At step S120, current operating parameters of the air conditioner are acquired.

At step S130, a current frosting state corresponding to the current operation parameter is determined through the correspondence.

For example: and identifying the frosting state of the condenser of the air conditioner in real time.

The comfort is achieved by controlling the control state of the air conditioner (for example, setting parameters), and the difference between input and output parameters is determined by recognizing the air conditioner state and the frosting condition.

Therefore, the current frosting condition of the air conditioner is effectively identified according to the current operation parameters based on the corresponding relation, so that accurate judgment basis is provided for the start and the end of defrosting, and the judgment result is good in accuracy.

In an optional example, the determining the current frosting status corresponding to the current operation parameter in step S130 may include: and determining the frosting state corresponding to the operation parameter which is the same as the current operation parameter in the corresponding relation as the current frosting state.

In an optional example, the determining, in step S130, a current frosting state corresponding to the current operating parameter may further include: when the corresponding relation can comprise a functional relation, inputting the current operation parameter into the functional relation, and determining the output parameter of the functional relation as the current frosting state.

Therefore, the current frosting state is determined according to the current operation parameters based on the corresponding relation or the functional relation, the determination mode is simple and convenient, and the reliability of the determination result is high.

In an alternative embodiment, the method may further include: and judging whether the current frosting state reaches the defrosting degree.

Optionally, a specific process of determining whether the current frosting state reaches the defrosting degree may be further described with reference to a flowchart of an embodiment of determining whether the current frosting state reaches the defrosting degree in the method of the present invention shown in fig. 7.

And step S710, determining whether the current frosting state reaches a set defrosting degree.

Step S720, when the current frosting state reaches the defrosting degree, entering a set defrosting mode; and/or initiating a prompt that the current frosting state reaches the defrosting degree.

And when the current frosting state does not reach the defrosting degree, maintaining the current operation.

Therefore, whether defrosting is needed or not is judged based on the current frosting state, the accuracy of defrosting detection is improved, the heating effect of the whole air conditioner is improved beneficially, the heating efficiency is improved, and the comfort of a user is improved.

In an alternative embodiment, the method may further include: and determining the entering time and the exiting time of the defrosting mode.

Optionally, according to the current frosting state, when a set defrosting mode needs to be entered, the entering time and/or exiting time of the defrosting mode (for example, the defrosting starting time and/or defrosting ending time is determined) can be determined.

For example: the defrosting operation can be preset according to the existing experience, for example, when the thickness reaches F; and when the thickness is reduced to A in the defrosting process, the defrosting operation is quitted.

Therefore, the defrosting process can be better mastered by determining the entering time, the exiting time and the like of the defrosting mode, the defrosting reliability can be improved, the running state of the air conditioner can be clearer for a user, the use is further facilitated, and the user experience is good.

In an alternative embodiment, the method may further include: and displaying and/or outputting the corresponding parameters and the like.

Optionally, at least one of a current operation mode of the air conditioner, the current operation parameter, the current frosting state, an entering time of a defrosting mode, and an exiting time of the defrosting mode may be displayed and/or output.

For example: at least one of the operation mode of the air conditioner, the current operation parameter, the current frosting state, the defrosting start time, and the defrosting end time may be further transmitted. For example: if the air conditioner is arranged, the information can be sent to the client. If the terminal side is available, the information can be sent to an air conditioner or other clients.

Therefore, the understanding and the checking of the user on the operation state of the air conditioner can be improved by displaying, outputting and other display operations on the corresponding parameters, and the method is strong in intuition and good in humanization.

In an alternative embodiment, the method may further include: and verifying whether the current frosting state and the actual frosting state are consistent or not.

Optionally, when a verification result that the current frosting state is not in accordance with the actual frosting state is received and/or it is determined that the corresponding relationship does not have the operation parameter the same as the current operation parameter, at least one maintenance operation of updating, correcting and relearning the corresponding relationship may be performed.

For example: the air conditioner can not know the actual frosting state, and needs feedback operation of a user, namely, if the frosting state is intelligently judged by the air conditioner, the user feeds back that the frosting state is not consistent with the actual state through operations such as a remote controller, and the air conditioner can know the frosting state.

For example: the authentication may be performed by a remote control. When the remote controller verifies, a specific process of verifying whether the current frosting state and the actual frosting state are consistent or not may be further described with reference to a flowchart of an embodiment of verifying whether the current frosting state and the actual frosting state are consistent or not in the method of the present invention shown in fig. 8.

Step S810, verifying whether the current frosting state and the actual frosting state are consistent (for example, the actual frosting state may be displayed by the AR display module to verify whether the determined current frosting state and the actual frosting state are consistent).

Step S820, when the current frosting state does not conform to the actual frosting state and/or the corresponding relationship does not have the same operation parameter as the current operation parameter, performing at least one maintenance operation of updating, correcting and relearning the corresponding relationship.

For example: the current frosting state can be determined according to the maintained corresponding relation and the current operation parameters. For example: and determining the frosting state corresponding to the operation parameter which is the same as the current operation parameter in the maintained corresponding relation as the current frosting state.

Therefore, the accuracy and the reliability of the frosting state determination are favorably improved through maintaining the corresponding relation between the determined operation parameters and the frosting state.

Through a large number of tests, the technical scheme of the embodiment is adopted, and the self-learning function of the neural network is utilized to establish the corresponding relation between the operating parameters of the air conditioner and the frosting state of the condenser; according to the current operation parameters of the air conditioner, the current frosting state can be determined through the corresponding relation, the determination mode is reliable, and the determination result is good in accuracy.

According to the embodiment of the invention, a frosting detection device (for example, a condenser frosting condition detection system) of the air conditioner is also provided, which corresponds to the frosting detection method of the air conditioner. Referring to fig. 9, a schematic diagram of an embodiment of the apparatus of the present invention is shown. The frosting detection device of the air conditioner may include: a establishing unit 102, an obtaining unit 104 and a determining unit 106.

In an optional example, the establishing unit 102 may be configured to utilize a self-learning capability of the artificial neural network to establish a correspondence relationship between an operation parameter of the air conditioner in the heating mode and a frosting state of the condenser. The specific function and processing of the creating unit 102 are shown in step S110.

Optionally, the operating parameters may include: the system comprises environment parameters and/or working parameters and/or one-dimensional or more than two-dimensional arrays consisting of features extracted from the environment parameters and the working parameters according to a set rule.

In an alternative specific example, the environment parameter may include: at least one of indoor and outdoor ambient temperature, user body surface temperature, and indoor and outdoor ambient humidity.

In an alternative specific example, the operating parameters may include: at least one of set temperature, set wind speed, set wind gear, condenser temperature and outdoor fan running current.

Optionally, the correspondence relationship may include: and (4) functional relation.

In an alternative specific example, the operation parameter is an input parameter of the functional relationship, and the frosting state is an output parameter of the functional relationship.

Optionally, the establishing unit 102 establishes a corresponding relationship between an operation parameter of the air conditioner in the heating mode and a frosting state of the condenser, and specifically may include: sample data that may be used to establish a correspondence between the operating parameters and the frosting condition is obtained. The specific functions and processes of the establishing unit 102 are also referred to in step S210.

More optionally, the establishing unit 102 obtains sample data that may be used to establish a correspondence between the operating parameter and the frosting state, and specifically may include: collecting the operation data and the frosting state of the air conditioner under different use environments. The specific functions and processes of the establishing unit 102 are also referred to in step S310.

In a more optional specific example, the establishing unit 102 collects operation data of the air conditioner in different use environments, and specifically may include: and acquiring the operating parameters of the air conditioner in a laboratory simulation environment.

In a more optional specific example, the establishing unit 102 collects operation data of the air conditioner in different use environments, and may further include: and acquiring the operating parameters collected by the Internet of things technology when the user actually uses the air conditioner.

More optionally, the establishing unit 102 may acquire sample data that may be used to establish a correspondence between the operating parameter and the frosting state, and specifically may further include: and analyzing the operating data, and selecting data related to the frosting state as the operating parameters by combining with prestored expert experience information (for example, selecting parameters influencing the frosting state of the air conditioner as input parameters, and using the frosting state as output parameters). The specific functions and processes of the establishing unit 102 are also referred to in step S320.

More optionally, the establishing unit 102 may acquire sample data that may be used to establish a correspondence between the operating parameter and the frosting state, and specifically may further include: and taking the frosting state and the data pair formed by the selected operation parameters as sample data. The specific functions and processes of the establishing unit 102 are also referred to in step S330.

Optionally, the establishing unit 102 establishes a corresponding relationship between an operation parameter of the air conditioner in the heating mode and a frosting state of the condenser, and may specifically include: and analyzing the characteristics and the rules of the operating parameters, and determining the network structure and the network parameters of the artificial neural network according to the characteristics and the rules. The specific functions and processes of the establishing unit 102 are also referred to in step S220.

More optionally, the network structure may include: at least one of a BP neural network, a CNN neural network, an RNN neural network, and a residual neural network.

More optionally, the network parameter may include: at least one of the number of input nodes, the number of output nodes, the number of hidden layers, the number of hidden nodes, an initial weight and a bias value.

Optionally, the establishing unit 102 establishes a corresponding relationship between an operation parameter of the air conditioner in the heating mode and a frosting state of the condenser, and may specifically include: training and testing the network structure and the network parameters by using the sample data, and determining the corresponding relation between the operation parameters and the frosting state. The specific function and processing of the establishing unit 102 are also referred to in step S230.

More optionally, the establishing unit 102 trains the network structure and the network parameters, and specifically may include: selecting a part of data in the sample data as a training sample, inputting the operation parameters in the training sample into the network structure, and training through an activation function of the network structure and the network parameters to obtain an actual training result. The specific functions and processes of the establishing unit 102 are also referred to in step S410.

For example, ⑴ introduces input data x, calculates the actual output a (x) of the network based on the activation function, the initialized weight and the offset, i.e., a (x) is 1/(1+ e)^-z) Wherein Z ═ W_k*x+b_l。

In a more optional specific example, the establishing unit 102 trains the network structure and the network parameters, and may further include: determining whether an actual training error between the actual training result and a corresponding frosting condition in the training sample satisfies a set training error. The specific function and processing of the establishing unit 102 are also referred to in step S420. For example: taking the respective frosting status in the sample data as an expected training result.

In a more optional specific example, the establishing unit 102 trains the network structure and the network parameters, and may further include: determining that the training of the network structure and the network parameters is complete when the actual training error satisfies the set training error. The specific functions and processes of the establishing unit 102 are also referred to in step S430.

More optionally, the establishing unit 102 trains the network structure and the network parameters, and may specifically include: and when the actual training error does not meet the set training error, updating the network parameters through an error energy function of the network structure. The specific functions and processes of the creating unit 102 are also referred to in step S610.

In a more optional specific example, the establishing unit 102 trains the network structure and the network parameters, and may further include: retraining through the activation function of the network structure and the updated network parameters until the retrained actual training error meets the set training error. The specific functions and processes of the establishing unit 102 are also referred to in step S620.

For example: if the expected output y (x) and the actual output a (x) of the network are not satisfied, the network is updated according to the following mannerWeight W of the collateral_kOffset b from_l：

updating the weight of each layer:

updating each layer bias:

wherein:

More optionally, the establishing unit 102 tests the network structure and the network parameters, and specifically may include: selecting another part of data in the sample data as a test sample, inputting the operation parameters in the test sample into the trained network structure, and testing by using the activation function and the trained network parameters to obtain an actual test result. The specific functions and processes of the establishing unit 102 are also referred to in step S510.

In a more optional specific example, the establishing unit 102 may perform a test on the network structure and the network parameter, and specifically may further include: it is determined whether an actual test error between the actual test result and a corresponding frosting condition in the test sample (i.e., with the frosting condition in the sample data as an expected output parameter) satisfies a set test error. The specific functions and processes of the establishing unit 102 are also referred to in step S520.

In a more optional specific example, the establishing unit 102 may perform a test on the network structure and the network parameter, and specifically may further include: and when the actual test error meets the set test error, determining that the test on the network structure and the network parameters is finished. The specific functions and processes of the establishing unit 102 are also referred to in step S530.

More optionally, the establishing unit 102 tests the network structure and the network parameters, and may further include: and when the actual test error does not meet the set test error, retraining the network structure and the network parameters until the retrained actual test error is slower than the set test error.

Therefore, the network structure is retrained to be retested when the test error is larger, so that the network structure which is more accurate and reliable is favorably obtained, and the accuracy of determining the frosting state is further improved

In an alternative example, the obtaining unit 104 may be configured to obtain current operating parameters of the air conditioner. The specific function and processing of the acquisition unit 104 are referred to in step S120.

In an optional example, the determining unit 106 may be configured to determine, through the correspondence, a current frosting status corresponding to the current operating parameter. The specific function and processing of the determination unit 106 are referred to in step S130.

Optionally, the determining unit 106 determines the current frosting state corresponding to the current operating parameter, which may specifically include: and determining the frosting state corresponding to the operation parameter which is the same as the current operation parameter in the corresponding relation as the current frosting state.

Optionally, the determining unit 106 determines the current frosting state corresponding to the current operating parameter, and may specifically include: when the corresponding relation can comprise a functional relation, inputting the current operation parameter into the functional relation, and determining the output parameter of the functional relation as the current frosting state.

In an optional example, the determining unit 106 may be further configured to determine whether the current frosting state reaches a set defrosting degree. The specific functions and processes of the establishing unit 102 are also referred to in step S710.

In an optional example, the determining unit 106 may be further configured to enter a set defrosting mode when the current frost formation state reaches the defrosting degree, and/or initiate a prompt that the current frost formation state reaches the defrosting degree. The specific functions and processes of the establishing unit 102 are also referred to in step S720.

And maintaining the current operation when the current frosting state does not reach the defrosting degree.

In an optional example, the determining unit 106 may be further configured to determine, according to the current frost formation state, an entry time and/or an exit time of the defrosting mode (e.g., determine a defrosting start time and/or a defrosting end time) after a set defrosting mode needs to be entered.

In an optional example, the determining unit 106 may be further configured to display and/or output at least one of a current operation mode of the air conditioner, the current operation parameter, the current frosting state, an entering time of a defrosting mode, and an exiting time of the defrosting mode.

In an optional example, the determining unit 106 may be further configured to, when receiving a verification result that the current frosting state does not conform to the actual frosting state and/or determining that there is no operating parameter in the corresponding relationship that is the same as the current operating parameter, perform at least one maintenance operation of updating, correcting, and relearning on the corresponding relationship.

For example: the authentication may be performed by a remote control. When the remote controller is verified, the following operations can be executed:

for example: the determining unit 106 may be further configured to verify whether the current frosting state and the actual frosting state are consistent (for example, the actual frosting state may be displayed by an AR display module to verify whether the determined current frosting state and the actual frosting state are consistent). The specific function and processing of the determination unit 106 are also referred to in step S810.

For example: the determining unit 106 may be further configured to perform at least one of maintenance operations of updating, correcting, and relearning the corresponding relationship through the establishing unit 102 when the current frosting state does not conform to the actual frosting state and/or the corresponding relationship does not have the same operation parameter as the current operation parameter. The specific function and processing of the determination unit 106 are also referred to in step S820.

Since the processes and functions implemented by the apparatus of this embodiment substantially correspond to the embodiments, principles and examples of the method shown in fig. 1 to 8, the description of this embodiment is not detailed, and reference may be made to the related descriptions in the foregoing embodiments, which are not repeated herein.

Through a large number of tests, the technical scheme of the invention is adopted, the acquired data are trained and learned by utilizing the self-learning function of the neural network, and the operation parameters of the air conditioner and the corresponding relation function between the sensor parameters and the frosting state of the condenser are mastered, so that the frosting state of the condenser of the air conditioner is recognized in real time, the accuracy of frosting detection is improved, the heating effect of the whole air conditioner is improved, the heating efficiency is improved, and the comfort of a user is improved.

According to an embodiment of the present invention, there is also provided a storage medium corresponding to a frost formation detection method of an air conditioner. The storage medium may include: the storage medium has stored therein a plurality of instructions; the plurality of instructions are used for loading and executing the frosting detection method of the air conditioner by the processor.

Since the processing and functions implemented by the storage medium of this embodiment substantially correspond to the embodiments, principles, and examples of the methods shown in fig. 1 to 8, the description of this embodiment is not detailed, and reference may be made to the related descriptions in the foregoing embodiments, which are not described herein again.

Through a large number of tests, the technical scheme of the invention is adopted, the operation state rule of the air conditioner condenser in the frosting process is analyzed by utilizing the artificial neural network algorithm, the mapping rule between the frosting condition of the condenser and the operation parameter state of the air conditioner in the heating operation of the air conditioner is found through the self-learning and self-adaptive characteristics of the artificial neural network, the current frosting condition of the air conditioner can be effectively identified, so that an accurate judgment basis is provided for the start and the end of defrosting, and the judgment result is good in accuracy.

According to an embodiment of the present invention, there is also provided an apparatus corresponding to a frost formation detection method of an air conditioner or a frost formation detection method of an air conditioner. The apparatus may include: a processor for executing a plurality of instructions; a memory to store a plurality of instructions; the plurality of instructions are stored by the memory, and are loaded and executed by the processor. Alternatively, the apparatus may comprise: the frosting detection device of the air conditioner is described above.

For example: a method and a device for recognizing the frosting condition of an air conditioner condenser based on a neural network are provided. As shown in fig. 14, when the air conditioner with wireless communication is operated, the operation parameters of the air conditioner are uploaded to the intelligent device.

For example: and the intelligent device inputs the operation parameters into the trained network algorithm, and sends the frosting condition to the air conditioner or directly sends a defrosting or defrosting ending command after judging the frosting condition.

Optionally, the apparatus may comprise: the air conditioner comprises an air conditioner body used for controlling the air conditioner body and/or an external control end used for controlling the air conditioner body.

For example: the algorithm can also be directly integrated in the controller of the air conditioner, and an intelligent device is not additionally connected.

Wherein, the external control terminal may include: at least one of a wireless communication module, a router, a server and a terminal.

For example: smart devices include, but are not limited to, wireless communication modules, routers, servers, smart phones, and the like. For example: a smart device, may include: at least one of wireless communication module, router, server, smart mobile phone.

In an optional example, the device may use an artificial neural network algorithm to collect air conditioner operation data in a large number of different use environments (including but not limited to one or more of indoor and outdoor environment temperature and humidity, set temperature, and the like), select operation state parameters and frosting state parameters of a plurality of air conditioners as sample data, learn and train the neural network, and fit the neural network to the relationship between the air conditioner operation parameters and frosting state by adjusting the network structure and the weight values between network nodes, so that the neural network can accurately fit the corresponding relationship between the operation parameters of the air conditioner and the frosting state of the condenser.

For example: the network structure and the network nodes are terms in the field of neural networks, the nodes, namely neurons, represent a specific output function, and the structure is the summary of the composition condition and the connection condition of the whole neural network.

For example: the relationship between the network structure and the network nodes may include: the network structure can be determined by the connection condition among each network node, and different network structures can be represented by different node numbers, node-node connection modes, node layers and the like.

For example: adjusting the weights between the network structure and the network nodes may include: the overall network structure may be preset (or several similar structures may be preset, and the most suitable structure is selected according to the network training condition in the later stage), and the adjustment of the weight may be referred to the network training and testing in the following step 4.

The method comprises the following specific implementation steps:

step 1, data collection:

collecting the self running state parameters of the air conditioner in different use environments and the corresponding frosting condition. The specific collection mode includes, but is not limited to, modes of collecting operation parameters of the air conditioner in a laboratory simulation environment, collecting operation parameters of the air conditioner when the air conditioner is used by an actual user through the internet of things technology, and the like.

Step 2, sample data selection:

by analyzing the data and combining expert experience knowledge, parameters influencing the frosting state of the air conditioner are selected as input parameters, and the frosting state is taken as output parameters. In this embodiment, the input parameters include, but are not limited to, one or more of the following: indoor and outdoor environment temperature, air conditioner set temperature, set wind speed (gear), condenser temperature, outdoor fan running current and the like. The input parameters are not only single parameters, but also one-dimensional or multi-dimensional arrays of the input parameters formed by extracting features according to a certain rule.

And using part of the obtained input and output parameter pairs as training sample data and using part of the obtained input and output parameter pairs as test sample data.

Step 3, network structure design:

according to the data characteristics and the rules of the data characteristics, which affect the operation of the air conditioner, such as different set temperatures, environment temperatures, body surface temperatures and the like, the basic structure of the network, the number of input and output nodes of the network, the number of network hidden layers, the number of hidden nodes, the initial weight of the network and the like can be preliminarily determined.

For example: the data affecting the operation of the air conditioner may be all the operation data collected in step 1, or some items thereof.

For example: the characteristics of data influencing the operation of the air conditioner and the rules contained in the data, such as frosting, the frosting is very visual and has a relation with the temperature of the condenser, and the lower the temperature is, the higher the probability and thickness of frosting are, so that the temperature of the condenser can be used as an input parameter at first.

The schematic diagram of the neural network algorithm of the invention is shown in fig. 1, and the specific artificial neural network structure includes, but is not limited to, the following three network structures:

⑴ BP neural network

Fig. 10 shows a schematic structural diagram of the BP neural network, and in actual application, the number of nodes of the input layer, the hidden layer, the output layer, and the number of hidden layers can be adjusted as needed.

For example: for example, 8 output parameters are determined, and 10 output frosting conditions exist, the number of the nodes of the input layer is initially selected to be 8, and the number of the nodes of the output layer is selected to be 10; and the middle layer is selected to be 5 according to experience, and the later training result is not good and can be adjusted.

⑵ CNN convolutional neural network

The multilayer convolution network is to continuously extract and combine low-dimensional features to obtain higher-dimensional features, so that the higher-dimensional features can be used for classification or correlation tasks.

The schematic diagram of the multilayer convolutional network structure is shown in fig. 11, and the network structure can be adjusted according to actual conditions in practical application.

⑶ residual neural network

When the CNN network is debugged, the network performance cannot be improved by deepening the network layer number and changing the size of the convolution kernel. The residual block is added, so that the data before and after connection can be better, and the feature expression capability is enhanced, so that the learning capability of the convolutional network can be enhanced. As shown in fig. 13, the input of a certain segment of neural network is x, the desired output is h (x), and after the input x is input to the output as an initial structure, the target to be learned becomes f (x) ═ h (x) -x.

Fig. 12 shows a schematic diagram of a residual neural network structure, and the network structure can be adjusted according to actual situations in practical application.

That is to say, the artificial neural network used in the present solution is not limited to a certain network structure, and may be a classical artificial neural network such as a BP neural network, a high-level artificial neural network, or a deep learning network such as a CNN (convolutional neural network) and an RNN (recurrent neural network), and the specific solution may be selected according to an actual application scenario.

Step 4, network training and testing:

after the network design is completed, training the network by using training sample data.

The training method can be adjusted according to the actual network structure and the problems found in training. Only one of the methods of the present invention is illustrated herein as follows:

⑴ introduces input data x, calculates the actual output a (x) of the network based on the activation function, the initialized weight and the offset, i.e. a (x) is 1/(1+ e)^-z) Wherein Z ═ W_k*x+b_l。

⑵, determining whether the expected output y (x) and the actual output a (x) of the network meet the output accuracy requirement, namely:

| (|) (x) a (x) | <e, e is the target minimum error.

⑶ if the accuracy requirement is satisfied, ending the training, if not, updating the net according to the following modeWeight W of the collateral_kOffset b from_l：

updating the weight of each layer:

updating each layer bias:

wherein:

And after the network training is finished, the network is positively tested by using the test sample. When the test error does not meet the requirement, repeating the steps and retraining the network; and if the test error meets the requirement, finishing the network training test.

Since the processes and functions implemented by the device of this embodiment substantially correspond to the embodiments, principles and examples of the method shown in fig. 1 to 8 or the apparatus shown in fig. 9, no details are given in the description of this embodiment, and reference may be made to the related descriptions in the foregoing embodiments, which are not repeated herein.

Through a large number of tests, the technical scheme of the invention determines the current frosting state corresponding to the current operation parameter through the corresponding relation between the operation parameter and the frosting state, the operation mode is simple and convenient, and the operation result has high reliability.

In summary, it is readily understood by those skilled in the art that the advantageous modes described above can be freely combined and superimposed without conflict.

The above description is only an example of the present invention, and is not intended to limit the present invention, and it is obvious to those skilled in the art that various modifications and variations can be made in the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A frosting detection method of an air conditioner is characterized by comprising the following steps:

establishing a corresponding relation between the operating parameters of the air conditioner and the frosting state of the condenser in the heating mode by utilizing the self-learning capability of the artificial neural network;

acquiring current operating parameters of the air conditioner;

and determining the current frosting state corresponding to the current operation parameters according to the corresponding relation.

2. The method of claim 1, wherein,

the operating parameters include: the system comprises environment parameters and/or working parameters and/or one-dimensional or more than two-dimensional arrays consisting of features extracted from the environment parameters and the working parameters according to a set rule; wherein,

the environmental parameters comprise: at least one of indoor and outdoor ambient temperature, user body surface temperature and indoor and outdoor ambient humidity; and/or the presence of a gas in the gas,

the working parameters comprise: at least one of set temperature, set wind speed, set wind gear, condenser temperature and outdoor fan running current;

and/or the presence of a gas in the gas,

the corresponding relation comprises: a functional relationship;

the operation parameters are input parameters of the functional relation, and the frosting state is output parameters of the functional relation;

determining a current frosting status corresponding to the current operating parameter, including:

determining the frosting state corresponding to the operation parameter which is the same as the current operation parameter in the corresponding relation as the current frosting state; and/or the presence of a gas in the gas,

and when the corresponding relation comprises a functional relation, inputting the current operation parameter into the functional relation, and determining the output parameter of the functional relation as the current frosting state.

3. The method according to claim 1 or 2, wherein establishing the correspondence between the operating parameter of the air conditioner in the heating mode and the frosting state of the condenser comprises:

acquiring sample data for establishing a corresponding relation between the operating parameters and the frosting state;

analyzing the characteristics and the rules of the operating parameters, and determining the network structure and the network parameters of the artificial neural network according to the characteristics and the rules;

training and testing the network structure and the network parameters by using the sample data, and determining the corresponding relation between the operation parameters and the frosting state.

4. The method of claim 3, wherein obtaining sample data for establishing a correspondence between the operating parameter and the frosting condition comprises:

collecting operation data and frosting states of the air conditioner in different use environments;

analyzing the operating data, and selecting data related to the frosting state as the operating parameters by combining prestored expert experience information;

and taking the frosting state and the data pair formed by the selected operation parameters as sample data.

5. The method of claim 4,

collecting operation data of the air conditioner in different use environments comprises the following steps:

obtaining the operating parameters of the air conditioner in a laboratory simulation environment, and/or,

acquiring the operation parameters collected by the Internet of things technology when the user actually uses the air conditioner;

and/or the presence of a gas in the gas,

the network architecture, comprising: at least one of a BP neural network, a CNN neural network, an RNN neural network, and a residual error neural network; and/or the presence of a gas in the gas,

the network parameters comprise: at least one of the number of input nodes, the number of output nodes, the number of hidden layers, the number of hidden nodes, an initial weight and a bias value.

6. The method according to one of claims 3 to 5,

training the network structure and the network parameters, including:

selecting a part of data in the sample data as a training sample, inputting the operation parameters in the training sample into the network structure, and training through an activation function of the network structure and the network parameters to obtain an actual training result;

determining whether an actual training error between the actual training result and a corresponding frosting state in the training sample satisfies a set training error;

determining that the training of the network structure and the network parameters is completed when the actual training error satisfies the set training error;

and/or the presence of a gas in the gas,

testing the network structure and the network parameters, comprising:

selecting another part of data in the sample data as a test sample, inputting the operation parameters in the test sample into the trained network structure, and testing by using the activation function and the trained network parameters to obtain an actual test result;

determining whether an actual test error between the actual test result and a corresponding frosting condition in the test sample satisfies a set test error;

and when the actual test error meets the set test error, determining that the test on the network structure and the network parameters is finished.

7. The method of claim 6,

training the network structure and the network parameters, further comprising:

when the actual training error does not meet the set training error, updating the network parameters through an error energy function of the network structure;

retraining through the activation function of the network structure and the updated network parameters until the retrained actual training error meets the set training error;

and/or the presence of a gas in the gas,

testing the network structure and the network parameters, further comprising:

and when the actual test error does not meet the set test error, retraining the network structure and the network parameters until the retrained actual test error is slower than the set test error.

8. The method of any one of claims 1-7, further comprising:

determining whether the current frosting state reaches a set defrosting degree;

when the current frosting state reaches the defrosting degree, entering a set defrosting mode, and/or initiating a prompt that the current frosting state reaches the defrosting degree;

and/or the presence of a gas in the gas,

according to the current frosting state, when a set defrosting mode needs to be entered, determining the entering time and/or exiting time of the defrosting mode;

and/or the presence of a gas in the gas,

displaying and/or outputting at least one of a current operation mode of the air conditioner, the current operation parameters, the current frosting state, the entering time of a defrosting mode and the exiting time of the defrosting mode;

and/or the presence of a gas in the gas,

and when a verification result that the current frosting state is inconsistent with the actual frosting state is received and/or the corresponding relation does not have the operation parameter which is the same as the current operation parameter, performing at least one maintenance operation of updating, correcting and relearning on the corresponding relation.

9. A frost formation detection device for an air conditioner, comprising:

the building unit is used for building a corresponding relation between the operating parameters of the air conditioner and the frosting state of the condenser in the heating mode by utilizing the self-learning capability of the artificial neural network;

10. The apparatus of claim 9, wherein,

and/or the presence of a gas in the gas,

the corresponding relation comprises: a functional relationship;

the determining unit determines a current frosting state corresponding to the current operating parameter, and specifically includes:

11. The apparatus according to claim 9 or 10, wherein the establishing unit establishes a correspondence between an operation parameter of the air conditioner in the heating mode and a frosting state of the condenser, and specifically includes:

12. The apparatus according to claim 11, wherein the establishing unit obtains sample data for establishing a correspondence between the operating parameter and the frosting condition, and specifically includes:

13. The apparatus of claim 12,

the establishing unit collects operation data of the air conditioner in different use environments, and specifically comprises the following steps:

and/or the presence of a gas in the gas,

14. The apparatus according to one of claims 11 to 13,

the establishing unit trains the network structure and the network parameters, and specifically includes:

and/or the presence of a gas in the gas,

the establishing unit tests the network structure and the network parameters, and specifically includes:

15. The apparatus of claim 14,

and/or the presence of a gas in the gas,

16. The apparatus of any one of claims 9-15, further comprising:

the determining unit is further used for determining whether the current frosting state reaches a set defrosting degree;

and/or the presence of a gas in the gas,

the determining unit is further configured to determine, according to the current frosting state, entry time and/or exit time of the defrosting mode after a set defrosting mode needs to be entered;

and/or the presence of a gas in the gas,

the determining unit is further configured to display and/or output at least one of a current operation mode of the air conditioner, the current operation parameter, the current frosting state, entering time of a defrosting mode, and exiting time of the defrosting mode;

and/or the presence of a gas in the gas,

the determining unit is further configured to, when a verification result that the current frosting state does not conform to the actual frosting state is received and/or the corresponding relationship does not have an operating parameter that is the same as the current operating parameter, perform at least one of maintenance operations of updating, correcting, and relearning on the corresponding relationship through the establishing unit.

17. A storage medium, comprising: the storage medium has stored therein a plurality of instructions;

wherein the plurality of instructions are used for being loaded by the processor and executing the frosting detection method of the air conditioner according to any one of claims 1 to 8.

18. An apparatus, comprising:

a processor for executing a plurality of instructions;

a memory to store a plurality of instructions;

wherein the plurality of instructions are stored by the memory and loaded and executed by the processor to perform the frosting detection method of the air conditioner according to any of claims 1-8;

or,

the frost formation detecting apparatus of an air conditioner according to any of claims 9 to 16.

19. The apparatus of claim 18, characterized in that it comprises: the air conditioner comprises an air conditioner body and/or an external control end, wherein the air conditioner body is used for controlling the air conditioner body; wherein,

the external control terminal includes: at least one of a wireless communication module, a router, a server and a terminal.