Intelligent air conditioner optimization control method and device

Abstract

The invention relates to the technical field of air conditioner control, and discloses an intelligent air conditioner optimization control method and device, and the method comprises the steps: building an air conditioner control strategy library, collecting and coding environment data and air conditioner setting data, and carrying out the index building and storage through a dictionary tree. And in combination with a machine learning algorithm, training an air conditioner parameter prediction model to realize intelligent decision making. Indoor environment data is collected in real time, after cleaning and coding are conducted, a matching index is searched in the dictionary tree, and corresponding air conditioner settings are retrieved. And if the complete matching index is not found, the optimal air conditioner setting is generated by utilizing the prediction model. And finally, the generated control strategy is sent to an air conditioner controller, and intelligent adjustment of the air conditioner temperature, the air speed, the operation mode and the air purification setting is achieved. The self-adaptive capacity and the intelligent level of the air conditioner are improved, the control strategy can be dynamically adjusted according to using habits and preferences of different users and complex changes of the external environment, the efficient and comfortable effect is achieved, and the user experience is remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of air conditioner control, in particular to an intelligent air conditioner optimization control method and device. BACKGROUND

[0002] With the rapid development of modern science and technology, smart home has become an important part of improving the quality of life, and the intelligent degree of the control method of intelligent air conditioner as an important part of smart home directly affects the user's comfortable experience and the effective use of energy. However, although the current market is full of intelligent air conditioner products, the control method still has many deficiencies, which is difficult to meet the user's all-round demand for high efficiency, comfort and energy saving.

[0003] The control strategy of the existing intelligent air conditioner is often based on fixed algorithm or rule, lacking self-learning and self-adaptive ability. This means that in the face of different user habits, preferences and complex changes of external environment, the traditional intelligent air conditioner cannot dynamically adjust its control strategy to achieve optimal performance. For example, in the season alternation or special weather conditions, the fixed control strategy may not adapt to the environmental changes in time, resulting in low efficiency of air conditioner operation or reduction of user comfort. The control method of the existing intelligent air conditioner has obvious defects in environment perception, strategy adjustment and decision making, which cannot meet the user's demand for intelligent, personalized and efficient energy-saving air conditioner control. SUMMARY

[0004] The purpose of the present application is to provide an intelligent air conditioner optimization control method and device to solve the problems raised in the background.

[0005] To achieve the above purpose, the present application provides the following technical scheme: an intelligent air conditioner optimization control method, the method comprising:

[0006] Constructing a standard air conditioner control strategy library: acquiring all environmental data and its corresponding air conditioner setting data, encoding the acquired data, and indexing and storing the encoded data by using a dictionary tree; the environmental data includes environmental temperature, humidity and air quality data; the air conditioner setting data includes air conditioner temperature, wind speed, running mode and air purification setting parameters;

[0007] Using machine learning algorithm and combining the data in the air conditioner control strategy library, training the air conditioner parameter prediction model;

[0008] Real-time acquisition of current indoor temperature, humidity and air quality data; and cleaning the collected data;

[0009] The cleaned data is encoded, and an index matching the encoded data is found in the dictionary tree; the corresponding air conditioner settings are retrieved from the dictionary tree according to the found index; if a completely matching index is found, the corresponding air conditioner settings are directly applied; if no completely matching index is found, the air conditioner parameter prediction model is used to predict the optimal air conditioner settings according to the current data;

[0010] The generated control strategy is sent to the air conditioner controller to adjust the temperature, air speed, operation mode and air purification settings of the air conditioner.

[0011] Preferably, the specific method for constructing the air conditioner control strategy library is as follows:

[0012] Define the data structure: set the dictionary tree as the data structure for storing the air conditioner control strategy, wherein each node contains a private two-dimensional array, which is used to record the character set represented by the path from the root node to the current node; the leaf node contains a target text, which represents a specific air conditioner control strategy, including the temperature setting, air speed, operation mode and air purification setting parameters of the air conditioner; the dictionary tree uses a hash table structure to efficiently store and retrieve node information;

[0013] Character encoding: encode the environmental data as a combination of English letters to represent the environmental data;

[0014] Design a character matcher: the matcher has an ASCII code calculator built-in; using the ASCII code calculator, the character matcher quickly locates a specific index in the two-dimensional array, and then finds the complete path matching the environmental data encoding in the dictionary tree;

[0015] Implement a tree constructor: the tree constructor provides addition and deletion functions for users to add or delete nodes and paths in the dictionary tree according to actual needs.

[0016] Preferably, the steps for constructing and training the air conditioner parameter prediction model include:

[0017] Step 1: Form a regular data set from the data in the air conditioner control strategy library; and divide the data set into a training set, a validation set and a test set, wherein the training set is used for model training, the validation set is used for model effect verification and optimization, and the test set is used for verifying the generalization ability of the model on unknown data sets;

[0018] Step 2: Use a linear regression model to construct a parameter adjustment prediction model, and the algorithm formula of the linear regression model is as follows:

[0019] y = β0 + β1x1 + β2x2 +... + β n x n + ∈

[0020] Where y is the target variable, representing the setting parameter of the air conditioner, x1, x2,..., x n are input variables, representing environmental parameters; β1, β2,..., β n are regression coefficients determined through model training; β0 is the intercept term, which represents the base value of the air conditioner setting parameter without any environmental impact; ∈ is the error term; the purpose of model training is to estimate the regression coefficients β0, β1, β1, β2,..., β n , so that the error between the predicted value and the true value is minimized;

[0021] Step three: set initial parameters for the model, assign an initial value to each feature x i in the linear regression model, corresponding to the parameter β i , which is set to a random number close to 0, and also assign an initial value to the intercept term β0; these initial parameter values will serve as the starting point for model training, which will be continuously optimized and adjusted through subsequent training processes;

[0022] Step four: use the training set to adjust the prediction model, continuously adjust the model parameters through optimization algorithms to minimize the difference between the predicted setting parameter and the actual setting parameter; use the validation set to validate the trained model, adjust the model parameters or structure according to the validation results; use the test set to comprehensively evaluate the optimized model through the set evaluation indicators, and ensure that the model performance meets the expected standards;

[0023] Step five: save the trained and validated air conditioner parameter prediction model.

[0024] Preferably, the way of cleaning the real-time collected data includes:

[0025] Missing value processing, using linear interpolation algorithm for filling;

[0026] Outlier processing, using threshold judgment method, regarding data points exceeding the preset threshold range as outliers, and replacing them with the mean value of their neighboring points.

[0027] Preferably, the mean square error (MSE) algorithm is used to evaluate the performance of the air conditioner parameter prediction model, and the specific algorithm is:

[0028]

[0029] Where n is the number of data points, Y pred,i is the model predicted value, Y actual,i is the actual value; the prediction accuracy of the model is evaluated according to the MSE value, and the smaller the MSE value, the higher the prediction accuracy of the model.

[0030] Preferably, the method further comprises: identifying different use scenarios, and pre-setting multiple sets of air conditioner control strategies according to the characteristics and needs of different scenarios; real-time detection of the current scenario, automatic switching to the corresponding control strategy, and ensuring that the air conditioner settings match the needs of the scenario.

[0031] Preferably, the method further comprises: periodically updating the air conditioner control strategy library, incorporating new environmental data and corresponding optimized air conditioner setting data, and ensuring the effectiveness and adaptability of the strategy library.

[0032] Preferably, the method further comprises: monitoring the actual operation effect of the air conditioner, including the indoor temperature change rate, humidity adjustment efficiency and air purification effect; according to the deviation of the monitoring result and the expected target, using a feedback mechanism to dynamically adjust the air conditioner control strategy library and the air conditioner parameter prediction model.

[0033] Preferably, an intelligent air conditioner optimization control device comprises a data acquisition layer, a data processing layer, a strategy generation layer and a control execution layer.

[0034] The data acquisition layer comprises an environmental sensor module and a data cleaning module; the environmental sensor module is used to collect indoor temperature, humidity and air quality data in real time; the data cleaning module is used to preprocess the collected data to remove outliers and noise;

[0035] The data processing layer comprises a data encoding module and a dictionary tree index module; the data encoding module is used to encode the cleaned data; the dictionary tree index module is used to establish and store the index of the encoded data;

[0036] The strategy generation layer comprises a model training module, an index matching module and a parameter prediction module; the model training module uses machine learning technology to train the air conditioner parameter prediction model in combination with the data in the air conditioner control strategy library; the index matching module is used to find the index matching the encoded data in the dictionary tree; the parameter prediction module is used to predict the best air conditioner settings according to the current data using the air conditioner parameter prediction model when no completely matching index is found;

[0037] The control execution layer comprises a control strategy sending module and an air conditioner controller module; the control strategy sending module is used to send the generated control strategy to the air conditioner controller; the air conditioner controller module is responsible for adjusting the temperature, wind speed, operation mode and air purification settings of the air conditioner according to the received control strategy.

[0038] Compared with the prior art, the present application has the following advantages:

[0039] The present application constructs an air conditioner control strategy library and combines machine learning algorithms, enabling the air conditioner to dynamically adjust control strategies based on real-time collected environmental data and user habits. This adaptive capability ensures that the air conditioner can provide optimal operating conditions in different seasons, weather, and user usage habits, significantly improving user comfort.

[0040] The present application takes into account the usage habits and preferences of different users, gradually learns and optimizes control strategies through machine learning algorithms, enabling the air conditioner to be personalized according to the unique needs of each user. This personalized customization not only improves user satisfaction, but also further enhances the intelligence and practicality of the air conditioner. BRIEF DESCRIPTION OF DRAWINGS

[0041] Fig. 1 Step diagram of the intelligent air conditioner optimization control method described in the present application;

[0042] Fig. 2 Dictionary tree structure diagram;

[0043] Fig. 3 Storage structure diagram for air conditioner control strategy library. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be described in detail below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present application.

[0045] Please refer to Figs. 1-3 The present application provides an intelligent air conditioner optimization control method, which realizes intelligent, personalized and energy-efficient air conditioner control by constructing an air conditioner control strategy library, applying machine learning algorithms and real-time environmental data perception. The following is a detailed implementation of the method:

[0046] Step 1: Construct a standard air conditioner control strategy library:

[0047] Data collection: First, through the arrangement of environmental sensors (such as temperature sensors, humidity sensors and air quality sensors) in the air conditioner and the recording function of the air conditioner itself, a large amount of environmental data and corresponding air conditioner setting data are collected. Environmental data specifically includes environmental temperature, humidity and air quality data (such as PM2.5 concentration, CO2 concentration, etc.), and air conditioner setting data includes air conditioner temperature, wind speed, operating mode (such as cooling, heating, dehumidification, air supply, etc.) and air purification setting parameters.

[0048] Data Encoding: Preprocess the collected environmental data and air conditioner setting data, including data cleaning, format unification, and encoding. The purpose of encoding is to convert continuous or discrete numerical data into a format that is easily processed by machines.

[0049] Indexing and Storage: Use a Trie tree structure to index and store the encoded data. Trie tree is a tree structure that can effectively match and retrieve strings. Through the Trie tree, you can quickly find the air conditioner settings that match the specific environmental data.

[0050] Step 2: Train air conditioner parameter prediction model:

[0051] Select machine learning algorithm: According to the characteristics of the data in the air conditioner control strategy library and the prediction requirements, select appropriate machine learning algorithms such as random forest, support vector machine or neural network, etc.

[0052] Model training: Use the data in the air conditioner control strategy library to train the selected machine learning algorithm, and build an air conditioner parameter prediction model. During the training process, adjust the algorithm parameters and optimize the model structure to improve the prediction accuracy and generalization ability of the model.

[0053] Step 3: Real-time data collection and processing:

[0054] Real-time collection: Collect real-time indoor temperature, humidity and air quality data through environmental sensors.

[0055] Data cleaning: Clean the collected data, including removing outliers, filling missing values and normalizing, etc., to ensure the accuracy and consistency of the data.

[0056] Step 4: Data encoding and index lookup

[0057] Data encoding: Encode the cleaned data to match the index in the Trie tree.

[0058] Index lookup: Find the index that matches the encoded data in the Trie tree. If a complete match is found, apply the corresponding air conditioner settings directly; if no complete match is found, go to the next step.

[0059] Step 5: Air conditioner parameter prediction and application:

[0060] Parameter prediction: For cases where no complete match index is found, use the trained air conditioner parameter prediction model to predict the best air conditioner settings based on the current environmental data.

[0061] Control strategy generation: According to the found index or the predicted air conditioner settings, generate a control strategy, including temperature, wind speed, running mode and air purification settings of the air conditioner, etc.

[0062] Strategy application: Send the generated control strategy to the air conditioner controller, adjust the operation of the air conditioner through the controller to realize intelligent, personalized and efficient air conditioner control.

[0063] The intelligent air conditioner optimization control method provided by the application can realize real-time sensing of environmental changes, dynamically adjust the air conditioner control strategy, and meet the user's all-round demand for high efficiency, comfort and energy saving. At the same time, the method also has self-learning and self-adaptive ability, which can continuously optimize the control strategy, improve the operation efficiency of the air conditioner and the comfort of the user.

[0064] The application will be further described below in combination with examples 1 to 3:

[0065] Example 1:

[0066] The specific method for constructing the air conditioner control strategy library is:

[0067] i) Define data structure:

[0068] Select the dictionary tree (Trie tree) as the core data structure for storing air conditioner control strategies. The dictionary tree is a multi-way tree, and its characteristic is that each node contains a private two-dimensional array, which records the character set passed through the path from the root node to the current node. In this application, this two-dimensional array is designed to correspond to the characters in the ASCII code table, so as to facilitate fast matching and retrieval.

[0069] Leaf nodes play a special role in the dictionary tree, they not only contain path information, but also contain a target text. This target text specifically represents a specific air conditioner control strategy, covering detailed information such as temperature setting, wind speed, running mode and air purification setting parameters of the air conditioner.

[0070] In order to improve the storage and retrieval efficiency, the dictionary tree adopts a hash table structure to store and retrieve node information. In this way, when looking for a specific node, the target node can be quickly located through hash operation, thereby greatly shortening the search time.

[0071] ii) Character encoding: encode the environment data as a combination of English letters to represent the environment data; the following is a specific implementation of the encoding rule:

[0072] Temperature Encoding: Convert the ambient temperature value to Celsius and take the integer part. Map each digit of the integer part to a specific English letter. For example, use 'A' to 'Z' to represent 0 to 25 (assuming the temperature range is between -9 and 49, adjust the mapping range as needed). If the temperature value is negative, add a specific prefix letter at the beginning, such as 'M' to represent negative. For the decimal part, ignore it or process it similarly according to the precision requirement.

[0073] Humidity Encoding: Convert the humidity value to a percentage and take the integer part. Use a similar method as temperature encoding to map each digit of the integer part to an English letter. For example, use 'a' to 'z' to represent 0 to 25 (assuming the humidity range is between 0% and 100%).

[0074] Air Quality Encoding: According to the Air Quality Index (AQI) or specific air quality parameters (such as PM2.5 concentration), convert it into a level or range. Assign a specific combination of English letters to each level or range. For example, use 'y' for good,'m' for medium, 'p' for poor, 'v' for very poor, etc. If more detailed encoding is needed, process the specific numerical value of the air quality parameter similarly as temperature and humidity.

[0075] According to the above encoding rules, convert the temperature, humidity and air quality data into combinations of English letters respectively. You can concatenate these combinations to form a complete environment data encoding string.

[0076] Example: Suppose the current ambient temperature is 25°C, humidity is 60%, and air quality is good (for example, PM2.5 concentration is considered good within 0-35 μg / m 3

[0077] Temperature Encoding: Map 25 to 'y' (assuming 'a' = 0, 'b' = 1,..., 'y' = 25).

[0078] Humidity Encoding: Map 60 to 'f' (assuming 'a' = 0, 'b' = 1,..., 'f' = 6).

[0079] Air Quality Encoding: Map good to 'g'.

[0080] Therefore, the complete environment data encoding is "yfg". This encoding string can be used to retrieve the corresponding air conditioner control strategy in the dictionary tree.

[0081] iii) Design character matcher:

[0082] ​The character matcher is a key component in the invention, which has an ASCII code calculator built-in. This calculator can quickly calculate the ASCII code value of the input character and use this value to locate the specific index in the two-dimensional array.

[0083] When it is necessary to find a complete path that matches the environment data encoding, the character matcher will traverse the input environment data encoding characters one by one, use the ASCII code calculator to find the corresponding characters at each node of the dictionary tree, and gradually deepen along the matched path until the leaf node is found or the matching cannot continue.

[0084] iv) Implement the tree constructor:

[0085] The tree constructor is a tool responsible for building and managing the dictionary tree. It provides addition and deletion functions, allowing users to flexibly add or delete nodes and paths in the dictionary tree according to actual needs.

[0086] When adding nodes, the tree constructor will first check whether the current path already exists in the dictionary tree. If not, it will create new nodes one by one and connect them in path order, finally storing the corresponding air conditioner control strategy at the leaf node.

[0087] When deleting nodes, the tree constructor will first locate the target node, then recursively delete its child nodes and related path information. If the target node is a leaf node, it will be directly deleted; if the target node is a non-leaf node, it needs to delete all its child nodes first, then delete the target node itself.

[0088] Specific example: Suppose the following air conditioner control strategy needs to be added to the dictionary tree: when the environment temperature is 25℃, the humidity is 50%, and the air quality is good, the air conditioner temperature is set to 24℃, the wind speed is medium, the operation mode is cooling, and the air purification setting is on.

[0089] First, encode these environment data and air conditioner setting data into character sequences, such as "yeg". Then, use the tree constructor's addition function to create corresponding nodes in the dictionary tree one by one, and connect them to form a path. At the leaf node, store the specific information of the air conditioner control strategy. When you need to retrieve this control strategy, input the same environment data encoding, and the character matcher will quickly find the matching path in the dictionary tree and return the control strategy stored at the leaf node.

[0090] Example 2:

[0091] The following steps are used to build and train the air conditioner parameter prediction model:

[0092] Step 1, data preparation:

[0093] Extract data from the air conditioning control strategy library to form a structured dataset. The dataset should include environmental parameters (such as temperature, humidity, air quality, etc.) and corresponding air conditioning setting parameters (such as temperature settings, wind speed, operating mode, etc.).

[0094] Randomly divide the dataset into training set, validation set and test set. The proportion can be adjusted according to actual needs, but generally it is recommended that the training set accounts for 70%, the validation set accounts for 15%, and the test set accounts for 15%. In this way, the performance of the model on different datasets can be fully evaluated.

[0095] Step 2, model construction:

[0096] Use a linear regression model as the parameter adjustment prediction model. Linear regression model can capture the linear relationship between environmental parameters and air conditioning setting parameters, and is commonly used in prediction problems. The algorithm formula of linear regression model is

[0097] y = β0 + β1x1 + β2x2 +... + β n x n + ∈

[0098] Where y is the target variable, representing the setting parameters of the air conditioner; x1, x2,..., x n are input variables, representing environmental parameters; β1, β2,..., β n are regression coefficients determined through model training; β0 is the intercept term, which represents the base value of the air conditioning setting parameter when there is no environmental impact; ∈ is the error term; The purpose of model training is to estimate the regression coefficients β0, β1, β1, β2,..., β n , so that the error between the predicted value and the true value is minimized.

[0099] Step 3, parameter initialization:

[0100] Assign an initial value to each feature x i in the linear regression model corresponding to the parameter β i . The initial value is usually set to a random number close to 0 to ensure that the starting point of model training is random and avoid falling into local optimal solution. Assign an initial value to the intercept term β0, which represents the base value of the air conditioning setting parameter when there is no environmental impact (such as β0 = 0.5, β1 = 0.1, β2 = 0.2, β3 = 0.05, etc.).

[0101] Step 4, model training and validation:

[0102] Input the training set data into the linear regression model, and continuously adjust the model parameters through optimization algorithms (such as gradient descent method) to minimize the difference between the predicted setting parameters and the actual setting parameters. Specifically, for each parameter βj (including the intercept term β0), whose update formula is as follows:

[0103]

[0104] where η is the learning rate, controlling the step size of parameter updates. is the partial derivative of the loss function with respect to the parameter β j , i.e., the gradient.

[0105] The trained model is verified using the validation set data. According to the verification result, the model parameters or structure can be adjusted to improve the generalization ability of the model. For example, the learning rate, regularization parameter and other hyperparameters can be adjusted. The optimized model is comprehensively evaluated using the test set data through the set evaluation index (such as mean square error MSE). The specific algorithm of mean square error MSE is as follows:

[0106]

[0107] where n is the number of data points, Y pred,i is the predicted value of the model, and Y actual,i is the actual value; the prediction accuracy of the model is evaluated according to the MSE value, and the smaller the MSE value, the higher the prediction accuracy of the model.

[0108] Step 5, model saving:

[0109] The air conditioner parameter prediction model that has completed training and verification is saved as a file or database record. This can facilitate the loading and use of the model in subsequent applications without the need for retraining.

[0110] As new data is continuously generated and accumulated, the model can be updated and retrained regularly to ensure that the model always maintains optimal performance.

[0111] Example 3:

[0112] The way to clean the real-time collected data includes:

[0113] Missing value processing, specifically:

[0114] i) Identify missing values: First, traverse the real-time collected data to identify missing values. Missing values may appear as null, NaN (not a number), or other special markers.

[0115] ii) Linear interpolation filling: For the identified missing values, linear interpolation algorithm is used for filling. Linear interpolation is a method of estimating unknown data points based on known data points. In specific implementation, the estimated value of the missing value position can be calculated by taking the two valid data points before and after the missing value through a linear equation, and the estimated value is used to replace the missing value.

[0116] Suppose the real-time collected temperature data sequence is [22, 23,?, 25, 26], where "?" represents a missing value. Using the linear interpolation algorithm, the estimated value at the missing value position can be calculated as 24 (i.e., (23+25) / 2), so the missing value is replaced with 24, resulting in the complete temperature data sequence [22, 23, 24, 25, 26].

[0117] Anomaly value processing, specifically:

[0118] i) Set threshold range: According to the characteristics of real-time collected data and business requirements, set a reasonable threshold range for each data point. The threshold range can be set according to historical data, empirical values or business requirements.

[0119] ii) Identify abnormal values: Traverse the real-time collected data, and consider the data points that exceed the pre-set threshold range as abnormal values. Abnormal values may be caused by sensor failure, data transmission error or other reasons.

[0120] iii) Neighboring point mean replacement: For the identified abnormal values, use the mean of its neighboring points to replace it. In specific implementation, you can take several valid data points before and after the abnormal value (such as 3 points before and after), calculate the mean of these points, and replace the abnormal value with the mean. If the abnormal value is at the beginning or end of the data sequence, only consider the valid data points on one side.

[0121] Suppose the real-time collected humidity data sequence is [50, 55, 60, 1000, 65, 70], where 1000 is obviously beyond the normal threshold range (such as 50-100). Using the neighboring point mean replacement method, the mean of the valid data points before and after the abnormal value can be calculated as 60 (i.e., (60+65) / 2, ignoring the abnormal value itself), so the abnormal value 1000 is replaced with 60, resulting in the corrected humidity data sequence [50, 55, 60, 60, 65, 70].

[0122] The air conditioning optimization control method further comprises:

[0123] i) Identify different use scenarios and pre-set multiple sets of air conditioning control strategies:

[0124] Through built-in sensors (such as infrared sensors, sound sensors, image recognition modules, etc.) and user behavior analysis, real-time perception of indoor environment characteristics (such as number of people, activity state, light intensity, etc.) and user habits, identify different use scenarios, such as living alone, family gatherings, office, sleep, etc.

[0125] For each identified scene, one or more sets of air conditioner control strategies are preset. These strategies include but are not limited to temperature settings, humidity control, wind speed adjustment, air purification mode, etc., to ensure that the air conditioner settings are highly matched with the scene requirements. For example, in the sleep scene, the temperature setting gradually decreases to adapt to the physiological changes of the human body during sleep, while reducing noise and wind speed, providing a comfortable sleep environment.

[0126] ii) Real-time detection of current scene and automatic switching of control strategy:

[0127] Utilize sensor networks to continuously monitor indoor environmental parameters and user behavior, forming real-time scene description data. Match real-time monitoring data with preset scene characteristics, and when a scene change is detected, automatically switch to the corresponding air conditioner control strategy. The switching process should be smooth and fast to avoid discomfort to the user.

[0128] iii) Regularly update the air conditioner control strategy library:

[0129] Collect environmental data, user feedback, and optimized air conditioner setting data during air conditioner operation through cloud or local storage. Regularly analyze the collected data to extract the association rules between new environmental characteristics and optimized settings, and update the air conditioner control strategy library.

[0130] iv) Monitor the actual running effect of the air conditioner and dynamically adjust:

[0131] Monitor the actual running effect of the air conditioner in real time through built-in sensors, including indoor temperature change rate, humidity adjustment efficiency, air purification effect, etc. Compare the monitoring results with the expected target, calculate the deviation value. When the deviation exceeds the preset threshold, trigger the feedback mechanism to dynamically adjust the relevant strategies in the air conditioner control strategy library or optimize the air conditioner parameter prediction model.

[0132] An intelligent air conditioner optimization control device, including a data acquisition layer, a data processing layer, a strategy generation layer, and a control execution layer;

[0133] The data acquisition layer includes an environmental sensor module and a data cleaning module; the environmental sensor module is used to collect indoor temperature, humidity and air quality data in real time; the data cleaning module is used to preprocess the collected data to remove outliers and noise;

[0134] The data processing layer includes a data encoding module and a dictionary tree index module; the data encoding module is used to encode the cleaned data; the dictionary tree index module is used to establish and store the index of the encoded data;

[0135] The strategy generation layer comprises a model training module, an index matching module and a parameter prediction module; the model training module trains an air conditioner parameter prediction model by using a machine learning technique in combination with data in an air conditioner control strategy library; the index matching module is used to find an index matching the coded data in a dictionary tree; and the parameter prediction module is used to predict the best air conditioner settings according to current data by using the air conditioner parameter prediction model when no completely matching index is found.

[0136] The control execution layer comprises a control strategy sending module and an air conditioner controller module; the control strategy sending module is used to send the generated control strategy to an air conditioner controller; and the air conditioner controller module is responsible for adjusting the temperature, air speed, operation mode and air purification settings of the air conditioner according to the received control strategy.

[0137] It should be noted that, in this document, the terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed or inherent to such a process, method, article or device.

[0138] Although the embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made thereto without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. An intelligent air conditioning optimization control method, characterized in that, The method includes: Construct a standard air conditioning control strategy library: acquire all environmental data and their corresponding air conditioning setting data, encode the acquired data, and use a trie to index and store the encoded data; the environmental data includes ambient temperature, humidity, and air quality data; the air conditioning setting data includes air conditioning temperature, fan speed, operating mode, and air purification setting parameters; Using machine learning algorithms and combining data from the air conditioning control strategy library, an air conditioning parameter prediction model is trained. Real-time collection of indoor temperature, humidity, and air quality data; and data cleaning of the collected data; The cleaned data is encoded, and an index matching the encoded data is searched in the trie. Based on the index found, the corresponding air conditioning settings are retrieved from the trie. If a perfect match is found, the corresponding air conditioning settings are applied directly. If no perfect match is found, the air conditioning parameter prediction model is used to predict the optimal air conditioning settings based on the current data. The generated control strategy is sent to the air conditioning controller to adjust the air conditioner's temperature, fan speed, operating mode, and air purification settings.

2. The intelligent air conditioning optimization control method according to claim 1, characterized in that, The specific method for constructing an air conditioning control strategy library is as follows: Define the data structure: Use a trie as the data structure to store air conditioning control strategies. Each node contains a private two-dimensional array that records the set of characters represented by the path from the root node to the current node. Leaf nodes contain a target text that represents a specific air conditioning control strategy, including the air conditioning temperature setting, fan speed, operating mode, and air purification settings. The trie uses a hash table structure to efficiently store and retrieve node information. Character encoding: Encode environmental data using characters, employing a combination of English letters to represent the environmental data; Design a character matcher: This matcher has a built-in ASCII code calculator; using the ASCII code calculator, the character matcher quickly locates a specific index in a two-dimensional array, and then finds the complete path that matches the environment data encoding in the trie; Implement a tree builder: The tree builder provides add and delete functions, allowing users to add or delete nodes and paths in the trie according to their actual needs.

3. The intelligent air conditioning optimization control method according to claim 2, characterized in that, The steps for building and training an air conditioning parameter prediction model include: Step 1: Organize the data in the air conditioning control strategy library into a regular dataset; and divide the dataset into a training set, a validation set, and a test set. The training set is used for training the model, the validation set is used for validating and optimizing the model's performance, and the test set is used to validate the model's generalization ability on unknown datasets. Step 2: Construct a parameter-adjusted prediction model using a linear regression model. The algorithm formula for the linear regression model is as follows: y=β0+β1x1+β2x2+...+β n x n +∈ Where y is the target variable, representing the air conditioner's setting parameters, x1, x2, ..., x... n These are input variables, representing environment parameters; β1, β2, ..., β n These are the regression coefficients, determined through model training; β0 is the intercept term, which represents the baseline value of the air conditioning setting parameters when there is no environmental influence; ∈ is the error term; the purpose of model training is to estimate the regression coefficients β0, β1, β2, ..., β1. n This minimizes the error between the predicted and actual values. Step 3: Set initial parameters for the model, for each feature x in the linear regression model. i The corresponding parameter β i An initial value is assigned to each parameter, which is a random number close to 0. An initial value is also assigned to the intercept term β0. These initial parameter values ​​will serve as the starting point for model training and will be continuously optimized and adjusted through subsequent training processes. Step 4: Train the prediction model using the training set, continuously adjusting the model parameters through optimization algorithms to minimize the gap between the predicted and actual parameter settings; validate the trained model using the validation set, adjusting the model parameters or structure based on the validation results; and comprehensively evaluate the optimized model using the test set through the set evaluation metrics to ensure that the model performance meets the expected standards. Step 5: Save and validate the completed air conditioning parameter prediction model.

4. The intelligent air conditioning optimization control method according to claim 1, characterized in that, Methods for cleaning real-time collected data include: Missing values ​​are filled using a linear interpolation algorithm; Outlier handling employs a threshold judgment method, where data points exceeding a preset threshold range are considered outliers and replaced with the average of their nearest neighbors.

5. The intelligent air conditioning optimization control method according to claim 3, characterized in that, The mean squared error (MSE) algorithm is used to evaluate the performance of the air conditioning parameter prediction model. The specific algorithm is as follows: Where n is the number of data points, Y pred,i Y is the model's predicted value. actual,i The actual value is used to evaluate the model's prediction accuracy. The smaller the MSE value, the higher the model's prediction accuracy.

6. The intelligent air conditioning optimization control method according to claim 1, characterized in that, The method also includes: identifying different usage scenarios and pre-setting multiple air conditioning control strategies according to the characteristics and needs of different scenarios; detecting the current scenario in real time and automatically switching to the corresponding control strategy to ensure that the air conditioning settings match the scenario requirements.

7. The intelligent air conditioning optimization control method according to claim 1, characterized in that, The method also includes: regularly updating the air conditioning control strategy library, incorporating new environmental data and corresponding optimized air conditioning settings data, to ensure the effectiveness and adaptability of the strategy library.

8. The intelligent air conditioning optimization control method according to claim 1, characterized in that, The method further includes: monitoring the actual operating effect of the air conditioner, including the rate of change of indoor temperature, humidity regulation efficiency and air purification effect; and dynamically adjusting the air conditioner control strategy library and air conditioner parameter prediction model using a feedback mechanism based on the deviation between the monitoring results and the expected targets.

9. An intelligent air conditioning optimization control device, characterized in that: It includes a data acquisition layer, a data processing layer, a strategy generation layer, and a control execution layer; The data acquisition layer includes an environmental sensor module and a data cleaning module; the environmental sensor module is used to collect indoor temperature, humidity and air quality data in real time; the data cleaning module is used to preprocess the collected data to remove outliers and noise. The data processing layer includes a data encoding module and a trie index module; the data encoding module is used to encode the cleaned data; the trie index module is used to create and store an index of the encoded data. The strategy generation layer includes a model training module, an index matching module, and a parameter prediction module. The model training module uses machine learning technology and data from the air conditioning control strategy library to train an air conditioning parameter prediction model. The index matching module is used to find an index in the trie that matches the coded data. The parameter prediction module is used to predict the optimal air conditioning settings based on the current data when no perfectly matching index can be found. The control execution layer includes a control strategy sending module and an air conditioning controller module; the control strategy sending module is used to send the generated control strategy to the air conditioning controller; the air conditioning controller module is responsible for adjusting the air conditioner's temperature, fan speed, operating mode, and air purification settings according to the received control strategy.

Images