Temperature self-adaptive adjusting method of annular air outlet warmer

Abstract

The invention relates to the technical field of temperature control, and discloses a temperature self-adaptive adjusting method of an annular air outlet warmer. The method comprises the steps that user information of a warmer user is obtained, and after the warmer is started, a corresponding first target strategy and a corresponding second target strategy are selected based on the user information, historical room temperature change conditions and historical room humidity change conditions; dividing the preset period into a plurality of first intervals by using a first target strategy, and dividing the preset period into a plurality of second intervals by using a second target strategy; selecting a corresponding prediction model to generate a temperature prediction sequence and a humidity prediction sequence of a future interval based on the environment information before the starting time point and the first interval and the second interval; and a temperature control scheme is generated based on the temperature prediction sequence, the humidity prediction sequence and the set target temperature, and self-adaptive adjustment of the temperature of the warmer is achieved. According to the invention, a personalized temperature adjustment strategy can be provided for a heater user.

Description

Technical Field

[0001] The present application relates to the field of temperature control technology, and in particular to a temperature adaptive adjustment method for a ring-shaped air outlet heater. Background Art

[0002] Traditional heaters usually have a constant temperature setting, and users need to manually adjust the temperature to achieve the desired comfort level. Although this method is simple and easy to use, it is inefficient and cannot automatically adapt to changes in ambient temperature. Many heaters are equipped with simple thermostats, but most can only achieve on-off cycling or constant temperature control, and cannot dynamically adjust according to external conditions.

[0003] In order to solve the above problems, a variety of technical solutions have been proposed in the prior art to adaptively adjust the temperature of the heater. For example, the Chinese patent document with publication number CN109611947A discloses a temperature control method and device for a heater. The method dynamically adjusts the heating state and power output of the heater by obtaining the ambient temperature, the shell temperature of the heater and the working mode, and combines the preset ambient temperature range and the shell temperature threshold, thereby achieving efficient temperature regulation for different scenes, and effectively improving the intelligence and energy-saving effect of the heater. For another example, the Chinese patent document with publication number CN104697039A discloses an electric heater and a control method for the electric heater. The method adopts a distance detection module and a temperature detection module to obtain user distance and ambient temperature data in real time or periodically, and dynamically adjusts the wind speed and heating power in combination with the preset functional relationship or table lookup method of the electric control module, thereby providing users with a constant comfortable temperature.

[0004] However, the above two methods ignore the personalized needs of users. Different users have different perceptions of comfortable temperature. Relying solely on static adjustment methods may not meet the comfort requirements of every user. Summary of the invention

[0005] In order to automatically adjust the temperature of the heater according to the user's personalized needs, the present application provides a temperature adaptive adjustment method for a ring-shaped air outlet heater.

[0006] In order to achieve the above-mentioned purpose of the invention, the present invention proposes a temperature adaptive adjustment method of a ring-shaped air outlet heater, comprising: Obtain user information of the heater user, including the user's life pattern, the area where the heater is located, and the room where the heater is located; Preset multiple first segmentation strategies and second segmentation strategies. After the heater is turned on, select the corresponding first segmentation strategy and second segmentation strategy based on user information and historical room temperature changes and historical room humidity changes collected by the heater, which are defined as the first target strategy and the second target strategy respectively; Using the first target strategy to divide the preset period into a plurality of first intervals, and using the second target strategy to divide the preset period into a plurality of second intervals; Locate the first interval and the second interval where the heater is turned on, generate a temperature prediction sequence for future intervals based on the environmental information before the turn-on time point and the prediction model corresponding to the first interval, and generate a humidity prediction sequence for future intervals based on the prediction model corresponding to the second interval; The heater generates a temperature control scheme based on the temperature prediction sequence, the humidity prediction sequence and the set target temperature to achieve adaptive adjustment of the heater temperature.

[0007] Furthermore, presetting the first segmentation strategy includes the following steps: Establishing a historical database, the historical database includes historical ambient temperatures collected by different heaters, each of the historical ambient temperatures corresponds to user information, integrating the historical ambient temperatures corresponding to the same user information into multiple initial temperature sequences within the preset period, and correcting the initial temperature sequences to obtain an ambient temperature sequence; Performing a primary clustering on the user information in the historical database to generate multiple user portraits, performing a secondary clustering on the ambient temperature sequence under the same user portrait to generate multiple temperature change patterns, and determining a representative sequence for each temperature change pattern; Based on the representative sequence and the ambient temperature sequence under the temperature change mode corresponding to the representative sequence, multiple prediction models are established, and the prediction time period with the best prediction effect of each prediction model within the preset period is determined, and the division method of the prediction time period within the preset period is set as the first segmentation strategy.

[0008] Furthermore, correcting the initial temperature sequence comprises the following steps: Divide each of the initial temperature sequences into a first sequence and a second sequence, wherein the first sequence is a temperature change sequence after the heater is turned on, and the second sequence is a temperature change sequence after the heater is turned off; A processing function is constructed based on the changing rules of the first sequence and the second sequence in each of the initial temperature sequences. Based on the processing function, the first sequence in each of the initial temperature sequences is corrected to a third sequence. The third sequence represents a virtual temperature sequence in which the heater is not turned on during the time period corresponding to the first sequence. The corrected initial temperature sequence is defined as the ambient temperature sequence.

[0009] Further, determining the predicted time period includes the following steps: Locate the extreme value point in the representative sequence, define the time point corresponding to the extreme value point as an initial segmentation point, define the ambient temperature sequence under the temperature change mode corresponding to the representative sequence as a target sequence, and use the initial segmentation point to divide all the target sequences into a plurality of initial intervals; Taking the environmental information in the initial interval as input features and the temperature change sequence in the initial interval as output features, training and generating the corresponding prediction model, and retaining the prediction accuracy of each prediction model for each initial interval; Adjust the position of the initial segmentation point to change the size of the initial interval, re-establish the prediction model based on the changed initial interval, repeat this step to a preset number of times, screen out the prediction model with the highest prediction accuracy and the corresponding initial interval, and define the time period covered by the screened initial interval as the prediction time period.

[0010] Further, selecting the first target strategy includes the following steps: Match the corresponding user profile according to the user information, obtain the representative sequences corresponding to the various temperature change modes under the user profile, generate an actual temperature sequence based on the historical room temperature, calculate the similarity between the actual temperature sequence and each representative sequence, and use the first segmentation strategy corresponding to the representative sequence with the highest similarity as the first target strategy.

[0011] Furthermore, performing secondary clustering on the ambient temperature sequence comprises the following steps: The time domain features in each of the ambient temperature sequences are extracted, the time domain features are used as clustering features, and the ambient temperature sequences are secondary clustered using a K-mean clustering algorithm.

[0012] Furthermore, the time domain features include average value, peak-to-peak value, standard deviation, method, maximum value, and minimum value.

[0013] Furthermore, in the secondary clustering result, the ambient temperature sequence closest to the cluster center in the cluster is used as the representative sequence.

[0014] Furthermore, the heater includes a temperature priority mode and a humidity priority mode, and the heater includes multiple air outlets. When in the temperature priority mode and before the target temperature is reached, the heater opens all the air outlets. When in the humidity priority mode and the current humidity is within the first range, the heater opens 50% of the air outlets. When the current humidity is within the second range, the heater opens 70% of the air outlets.

[0015] Furthermore, the room information includes room types, and each room type is pre-set with the corresponding target temperature.

[0016] The present invention combines the user's life pattern, area and room information, and historical room temperature and humidity changes, so that the heater can intelligently select the appropriate temperature and humidity control strategy to adapt to different environmental requirements and user preferences. The adaptive adjustment mechanism ensures that the indoor temperature and humidity are within a comfortable range to meet the personalized needs of users.

[0017] By dividing a day into multiple intervals and matching the best prediction model, the heater can accurately predict the future trend of temperature and humidity. Based on the partition strategy, the system can optimize the prediction model for the temperature and humidity characteristics of different time periods, improving the prediction accuracy and applicability. At the same time, using the environmental data at the start time point, combined with the predicted sequence of temperature and humidity, the heater can formulate a heating plan more scientifically, optimize the control effect, and improve the overall system intelligence level and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0019] Figure 1 A schematic diagram of a temperature adaptive adjustment method for a ring-shaped air outlet heater according to the present application; Figure 2 Schematic diagram of the principle of dividing the initial interval for this application. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0021] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0022] like Figure 1 As shown, a temperature adaptive adjustment method for a ring-shaped air outlet heater comprises: S1: Obtain user information of the heater user, the user information including the user's life pattern, the area where the heater is located, and the room where the heater is located.

[0023] User information can be actively filled in by the user, and part of it can be automatically generated by the heater. When automatically generated, the life pattern can be counted by the opening and closing time points during use. For example, the heater is turned off at 7 o'clock from Monday to Friday and turned on at 18 o'clock. Regional information can be obtained using an APP installed in the user's mobile phone. The regional information is determined by obtaining the mobile phone positioning information. The regional information is, for example, a certain province or city. Room information includes room area, room type and personnel activity information. Room information is, for example, bedroom, living room, and children's room. Personnel activity information can be obtained through linkage with other IoT devices, such as smart surveillance cameras that automatically identify human activities and determine the number of people in the room.

[0024] S2: Preset multiple first segmentation strategies and second segmentation strategies. After the heater is turned on, the corresponding first segmentation strategy and second segmentation strategy are selected based on user information and historical room temperature changes and historical room humidity changes collected by the heater, which are defined as the first target strategy and the second target strategy respectively.

[0025] S3: using a first target strategy to divide a preset period into a plurality of first intervals, and using a second target strategy to divide the preset period into a plurality of second intervals.

[0026] The first segmentation strategy and the second segmentation strategy are pre-set and stored in the server. The first segmentation strategy refers to the rule of dividing the preset period into multiple time periods based on historical temperature data. Each time period corresponds to a temperature prediction model. The second segmentation strategy refers to the rule of dividing time periods based on historical humidity data for humidity prediction. The preset period is one day. The first segmentation strategy is used to divide one day into multiple first intervals, and the second segmentation strategy is used to divide one day into multiple second intervals, such as 18:00-21:00, 21:00-23:00, etc. Finally, the temperature change sequence in one day is divided into multiple small segments through the first interval, and the humidity change sequence in one day is divided into multiple small segments through the second interval. The reasons for the separate division and how to choose the appropriate segmentation strategy will be introduced later.

[0027] S4: Locate the first interval and the second interval where the heater is turned on, generate a temperature prediction sequence for the future interval based on the environmental information before the turn-on time point and the corresponding prediction model selected in the first interval, and generate a humidity prediction sequence for the future interval based on the corresponding prediction model selected in the second interval.

[0028] The first interval is the time period divided in the temperature prediction model, and the second interval is the time period divided in the humidity prediction model. Assuming that the current heater is turned on at 18:05, it is within the first interval of 18:00-19:00 and within the second interval of 17:30-22:00, then the heater obtains environmental information before 18:05, including indoor ambient temperature, outdoor weather temperature, solar radiation data, precipitation data, number of personnel activities, etc., and uses a suitable prediction model to predict the temperature of the future interval according to the first interval where the start time point is located, and generates a temperature prediction sequence. The future interval of this embodiment is the remaining time of the first interval where the start time point is located, such as predicting the temperature from 18:06 to 19:00 based on the environmental information before 18:05. Similarly, the humidity of the future interval is predicted using a suitable prediction model based on the environmental information and the second interval.

[0029] Since the daily temperature and humidity changes are different, the prediction model used to predict the temperature may have a poor prediction effect on the humidity in the same time period. Therefore, it is necessary to divide the temperature and humidity into the first interval or the second interval with an appropriate time span length respectively, and select the best prediction model for each first interval and second interval.

[0030] S5: The heater generates a temperature control scheme based on the temperature prediction sequence, the humidity prediction sequence and the set target temperature to achieve adaptive adjustment of the heater temperature.

[0031] The heater in this embodiment has a variety of temperature control schemes, such as temperature priority mode and humidity priority mode. In the absence of interference from a humidifier, the faster the temperature rises, the faster the moisture content in the indoor air evaporates, and the faster the humidity decreases. Therefore, the temperature priority mode prioritizes rapid temperature increase with the room temperature reaching the target temperature, regardless of humidity. The humidity priority mode takes both temperature and humidity into consideration for steady temperature increase. In particular, the greater the power of the heater, the higher its own temperature. Subsequent adjustment of the heater power is equivalent to adjusting the heater temperature.

[0032] For example, if the target temperature of the living room is set to 24°C and the temperature priority mode is selected, the heater generates a temperature prediction sequence. If it is found that the temperature prediction sequence will rise steadily in the future, indicating that the temperature can be gradually increased without the intervention of the heater in the future, the heater will heat with moderate heating power, moderate wind speed, and open some air outlets, so as to increase the room temperature to 24°C under moderate power and in coordination with the original environment. If the humidity priority mode is selected, the heater obtains the humidity prediction sequence. If it is found that the humidity prediction sequence will rise steadily in the future (there may be humidifier intervention), it indicates that the heater can use a larger power for heating, such as 1000W, to achieve a gradual increase in temperature and a slow decrease in humidity or keep it stable. If it is found that the humidity prediction sequence will fluctuate within a certain value in the future and is always at a low level, such as 40%, the heater can use a lower power for heating, such as 500W, to avoid the indoor humidity being further reduced due to the excessively fast heating speed.

[0033] The present invention combines the user's life pattern, area and room information, as well as the historical room temperature and humidity changes, so that the heater can intelligently select the appropriate temperature and humidity control strategy to adapt to different environmental requirements and user preferences. The adaptive adjustment mechanism not only improves the user experience of the heater, but also achieves high efficiency and energy saving by dynamically adjusting the heating power and wind speed, while ensuring that the indoor temperature and humidity are within a comfortable range to meet the personalized needs of users.

[0034] By dividing a day into multiple intervals and matching the best prediction model, the heater can accurately predict the future trend of temperature and humidity. Based on the partition strategy, the system can optimize the prediction model for the temperature and humidity characteristics of different time periods, improving the prediction accuracy and applicability. At the same time, using the environmental data at the start time point, combined with the predicted sequence of temperature and humidity, the heater can formulate a heating plan more scientifically, optimize the control effect, and improve the overall system intelligence level and efficiency.

[0035] It is particularly noteworthy that the present invention can provide a personalized temperature adjustment strategy for heater users.

[0036] In this embodiment, the preset first segmentation strategy includes the following steps: A historical database is established, which includes historical ambient temperatures collected by different heaters. Each historical ambient temperature corresponds to user information. The historical ambient temperatures corresponding to the same user information are integrated into multiple initial temperature sequences within a preset period. The initial temperature sequences are corrected to obtain an ambient temperature sequence.

[0037] Among them, correcting the initial temperature sequence includes the following steps: Each initial temperature sequence is divided into a first sequence and a second sequence. The first sequence is a temperature change sequence after the heater is turned on, and the second sequence is a temperature change sequence after the heater is turned off.

[0038] A processing function is constructed based on the changing rules of the first and second sequences in each initial temperature sequence. Based on the processing function, the first sequence in each initial temperature sequence is corrected to a third sequence. The third sequence represents a virtual temperature sequence in which the heater is not turned on during the time period corresponding to the first sequence. The corrected initial temperature sequence is defined as the ambient temperature sequence.

[0039] Temperature data is collected through sensors in heaters of different users and different models and uploaded to the server as historical ambient temperature data. These historical ambient temperatures include the ambient temperatures when the heater is turned on and when it is not turned on. User information in the historical database includes user ID, regional information, room information, etc. The meaning of limiting by a preset period is: taking one day as a period, the historical ambient temperatures within the included time are integrated into an initial temperature sequence. As before, the initial temperature sequence includes the ambient temperatures when the heater is turned on and when it is not turned on, and the temperature prediction sequence to be obtained does not include the ambient temperature sequence under the interference of the heater. If the initial temperature sequence is directly used to establish a prediction model, the prediction effect of the prediction model will be poor.

[0040] To solve the above problem, the present embodiment further modifies the initial temperature sequence. Specifically, each initial temperature sequence is first split into a first sequence and a second sequence. The first sequence is a temperature change sequence after the heater is turned on, and the second sequence is a temperature change sequence after the heater is turned off. The processing function constructed by the present invention is a linear regression function. For example, by analyzing a large number of initial temperature sequences under the same user information, the following processing function is fitted through a first-order linear model: ,in, is the corrected temperature at time t in the first sequence, is the temperature before correction at time t, that is, the ambient temperature detected by the sensor, r is the correction coefficient, is the time from when the heater is turned on to time t. The third sequence can be obtained by correcting the first sequence using the processing function. For example, the first sequence before correction is [15.6, 16.0, 16.5], and the third sequence after correction is [15.6, 15.7, 15.8]. The correction process of the humidity sequence is similar. When the start and stop time points of the humidifier cannot be determined, the first target point and the second target point are located as the start and stop time points of the humidifier when the humidity sequence is corrected. In the humidity sequence, the rising speed of humidity after the first target point exceeds the first predetermined value, and the falling speed of humidity after the second target point exceeds the second predetermined value. The first target point is the start time point of the humidifier, and its judgment condition is: the current humidity change rate (ΔH / Δt) exceeds the first preset value (5%RH / hour) and lasts for more than 3 minutes; the second target point is the humidifier off time point, and the judgment condition is: the humidity drop rate exceeds the second preset value (3%RH / hour) and lasts for more than 5 minutes. The first and second preset values ​​are obtained through experimental calibration, and the value ranges are 3-8%RH / hour and 2-5%RH / hour respectively.

[0041] The user information in the historical database is clustered once to generate multiple user portraits. The ambient temperature sequence under the same user portrait is clustered again to generate multiple temperature change patterns, and the representative sequence of each temperature change pattern is determined.

[0042] Based on the representative sequence and the ambient temperature sequence under the temperature change mode corresponding to the representative sequence, multiple prediction models are established, the prediction time period with the best prediction effect of each prediction model within the preset period is determined, and the division method of the prediction time period within the preset period is set as the first segmentation strategy.

[0043] In this embodiment, the time domain features in each ambient temperature sequence are extracted, and the time domain features are used as clustering features. The ambient temperature sequence is secondary clustered using the K-mean clustering algorithm. The time domain features include average value, peak-to-peak value, standard deviation, method, maximum value, and minimum value. In the secondary clustering result, the ambient temperature sequence closest to the cluster center in the cluster is taken as the representative sequence.

[0044] When clustering user information, the following features can be used as clustering features: geographic location, house area, daily usage time, number of times the heater is turned on and off, etc., and then the above features are normalized, and the K-means algorithm is selected, and the number of clusters is determined using the elbow rule. In this embodiment, the following user portraits are obtained by clustering: User portrait 1, geographic location: Province A, room type: living room, working from 9:00 to 18:30 from Monday to Friday; User portrait 2, geographic location: Province B, room type: children's room, going to school from 7:00 to 17:30 from Monday to Friday.

[0045] Since each user information corresponds to multiple ambient temperature sequences, after clustering the user information, there will be multiple user information under one user profile, so there will be a large number of ambient temperature sequences. Secondary clustering is required to integrate the ambient temperature sequences into multiple temperature change patterns. The algorithm of secondary clustering is the same as that of primary clustering. When clustering, the time domain features of the ambient temperature sequence are extracted as clustering features. The time domain features include the average value, peak-to-peak value, standard deviation, method, maximum value, and minimum value of the sequence. After obtaining multiple temperature change patterns, an ambient temperature sequence is selected as the representative sequence. Specifically, the ambient temperature sequence closest to the cluster center is selected as the representative sequence.

[0046] Finally, multiple linear prediction models or nonlinear prediction models are established to divide each day into multiple intervals of appropriate size. The division method is the first segmentation strategy. Each interval after division will correspond to a specific prediction model. Assume that the model corresponding to interval 1 is the target model. Among all the prediction models, the target model has the highest prediction accuracy for interval 1. The generation method of the second segmentation strategy is the same as the first segmentation strategy, which will not be repeated here.

[0047] In this embodiment, determining the predicted time period includes the following steps: The extreme points in the representative sequence are located, the time points corresponding to the extreme points are defined as the initial segmentation points, the ambient temperature sequence in the temperature change mode corresponding to the representative sequence is defined as the target sequence, and all target sequences are divided into multiple initial intervals using the initial segmentation points.

[0048] The above steps are described below with an example. Assuming that there are 50 ambient temperature sequences under temperature change mode A, for a clear description, these 50 ambient temperature sequences are defined as target sequences, among which there is a representative sequence B. The extreme points in this embodiment are the maximum and minimum points in the representative sequence. After locating the extreme points, the time points corresponding to the extreme points are used as the initial segmentation points to divide the 50 ambient temperature sequences (target sequences) under temperature change mode A, as shown in FIG. Figure 2 The figure shows a temperature variation curve drawn according to one of the ambient temperature sequences, where the ambient temperature sequence is divided into initial intervals 1 to 5 by an initial segmentation point.

[0049] Taking the environmental information in the initial interval as the input feature and the temperature change sequence in the initial interval as the output feature, the corresponding prediction model is trained and generated, and the prediction accuracy of each prediction model for each initial interval is retained.

[0050] Specifically, the environmental information includes the indoor environmental temperature, outdoor weather temperature, solar radiation data, precipitation data, number of human activities, etc. collected before the start of the initial interval. The outdoor weather temperature, solar radiation data, and precipitation data can be obtained through the Internet, and the number of human activities can be obtained by linking the network camera in the home. If there is no network camera in the home, the input feature is set to 0 by default. The temperature change sequence in each initial interval is used as the output feature. After the prediction model training is completed, the temperature change sequence of multiple intervals after the prediction time point will be output. For example, the heater is turned on at 18:05, there is a first interval 1 from 18:00 to 19:00, and there is a first interval 2 from 19:00 to 20:00. The heater uses the environmental information before 18:05 and uses the prediction model 1 to predict the temperature change sequence from 18:05 to 19:00 in the first interval 1. When it reaches 19:00, the prediction model 2 is used to predict the temperature change sequence of the first interval 2.

[0051] The prediction model can be a linear model or a nonlinear model. The linear model is generated by least squares fitting. The nonlinear model can be an LSTM neural network, an SVM support vector machine model including different kernel functions, etc. When establishing the model, the temperature change sequence is divided into a training set and a validation set. The prediction accuracy of each prediction model is obtained based on the validation set. In the first test, it was found that the prediction accuracy of the initial intervals 3, 4, and 5 using prediction model 3 was the highest. Prediction model 3 is a linear function.

[0052] Adjust the position of the initial split point to change the size of the initial interval, re-establish the prediction model based on the changed initial interval, repeat this step to the preset number of times, screen out the prediction model with the highest prediction accuracy and the corresponding initial interval, and define the time period covered by the screened initial interval as the prediction time period.

[0053] Then, the size of the initial interval is adjusted. There are many specific adjustment processes for the initial interval, such as first moving the initial interval 1 to the right, and then moving the initial interval 2 to the right. Those skilled in the art can determine it based on experience, which will not be repeated here.

[0054] For example, if we move the initial interval 4 to the right, we will eventually find that after the initial interval 4 completely covers the initial interval 5, we get the initial interval 6, that is, the initial interval 6 is Figure 2 For the interval covered by the initial intervals 4 and 5, for the initial interval 6, it is found that the prediction accuracy of the SVM with the Gaussian kernel function is higher than the prediction accuracy of the previous prediction model 3 and all other prediction models. Then the initial interval 6 is taken as the final interval, the corresponding time period is the prediction time period, and the SVM model with the Gaussian kernel function is taken as the prediction model corresponding to the initial interval 6.

[0055] In this embodiment, selecting the first target strategy includes the following steps: According to the user information, the corresponding user profile is matched, and the representative sequences corresponding to various temperature change modes under the user profile are obtained. The actual temperature sequence is generated based on the historical room temperature. The similarity between the actual temperature sequence and each representative sequence is calculated, and the first segmentation strategy corresponding to the representative sequence with the highest similarity is used as the first target strategy.

[0056] After entering the user information, the user information is clustered with the user information in the previous database, and the user portrait is determined based on the clustering results. Or first match the most similar life pattern in the user information, then match the most similar regional information, and then match the most similar room information to determine the corresponding user portrait. After that, if the heater does not have enough historical room temperatures, the smart mode cannot be turned on. After collecting a sufficient number of historical room temperatures, such as collecting temperature data from 16:00 to 18:00, the actual temperature sequence will be generated according to the time period, and the actual temperature sequence from 16:00 to 18:00 will be compared with the ambient temperature sequence in the interval of 16:00 to 18:00 in each representative sequence to obtain the corresponding similarity. The similarity can be the Euclidean distance between the two sequences, or the Pearson correlation coefficient, etc. The representative sequence with the highest similarity will correspond to a temperature change pattern, and the temperature change pattern will correspond to a first segmentation strategy, and the corresponding first segmentation strategy will be used as the first target strategy.

[0057] In this embodiment, the heater includes a temperature priority mode and a humidity priority mode, and the heater includes multiple air outlets. When in the humidity priority mode and the current humidity is within the first range, the heater opens 50% of the air outlets. When the current humidity is within the second range, the heater opens 70% of the air outlets.

[0058] When the heater is in humidity priority mode, and the current humidity is within the first range (30% - 40%), in this case, the indoor humidity is low, so the heater only opens 50% of the air outlets. When the humidity is within the second range (40% -20%), in this case, the indoor humidity is moderate, and the heater only opens 70% of the air outlets to increase the indoor temperature as quickly as possible without reducing the humidity too much. In addition, this embodiment provides an automatic switching mode. If the heater is currently in temperature priority mode, but the humidity is already lower than 20%, the system will automatically switch to humidity priority mode. If the heater is currently in humidity priority mode, but the humidity is already higher than 80% and the target temperature has not been reached, the system will automatically switch to temperature priority mode.

[0059] In this embodiment, the room information includes room types, and each room type is set with a corresponding target temperature.

[0060] For example, room types include bedroom, living room, study, children's room, etc., and the corresponding target temperatures are set to 23℃, 22℃, 22.5℃, and 23.5℃ respectively.

[0061] It should be understood that the various technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the various technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0062] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A temperature adaptive adjustment method for a ring-shaped air outlet heater, characterized in that: Obtain user information of the heater user, including the user's life pattern, the area where the heater is located, and the room where the heater is located; Preset multiple first segmentation strategies and second segmentation strategies. After the heater is turned on, select the corresponding first segmentation strategy and second segmentation strategy based on user information and historical room temperature changes and historical room humidity changes collected by the heater, which are defined as the first target strategy and the second target strategy respectively; Using the first target strategy to divide the preset period into a plurality of first intervals, and using the second target strategy to divide the preset period into a plurality of second intervals; Locate the first interval and the second interval where the heater is turned on, generate a temperature prediction sequence for future intervals based on the environmental information before the turn-on time point and the prediction model corresponding to the first interval, and generate a humidity prediction sequence for future intervals based on the prediction model corresponding to the second interval; The heater generates a temperature control scheme based on the temperature prediction sequence, the humidity prediction sequence and the set target temperature to achieve adaptive adjustment of the heater temperature.

2. The method according to claim 1, characterized in that Presetting the first segmentation strategy includes the following steps: Establishing a historical database, the historical database includes historical ambient temperatures collected by different heaters, each of the historical ambient temperatures corresponds to user information, integrating the historical ambient temperatures corresponding to the same user information into multiple initial temperature sequences within the preset period, and correcting the initial temperature sequences to obtain an ambient temperature sequence; Performing a primary clustering on the user information in the historical database to generate multiple user portraits, performing a secondary clustering on the ambient temperature sequence under the same user portrait to generate multiple temperature change patterns, and determining a representative sequence for each temperature change pattern; Based on the representative sequence and the ambient temperature sequence under the temperature change mode corresponding to the representative sequence, multiple prediction models are established, and the prediction time period with the best prediction effect of each prediction model within the preset period is determined, and the division method of the prediction time period within the preset period is set as the first segmentation strategy.

3. The method according to claim 2, characterized in that Correcting the initial temperature sequence comprises the following steps: Divide each of the initial temperature sequences into a first sequence and a second sequence, wherein the first sequence is a temperature change sequence after the heater is turned on, and the second sequence is a temperature change sequence after the heater is turned off; A processing function is constructed based on the changing rules of the first sequence and the second sequence in each of the initial temperature sequences. Based on the processing function, the first sequence in each of the initial temperature sequences is corrected to a third sequence. The third sequence represents a virtual temperature sequence in which the heater is not turned on during the time period corresponding to the first sequence. The corrected initial temperature sequence is defined as the ambient temperature sequence.

4. The method according to claim 2, characterized in that: Determining the forecast time period comprises the following steps: Locate the extreme value point in the representative sequence, define the time point corresponding to the extreme value point as an initial segmentation point, define the ambient temperature sequence under the temperature change mode corresponding to the representative sequence as a target sequence, and use the initial segmentation point to divide all the target sequences into a plurality of initial intervals; Taking the environmental information in the initial interval as input features and the temperature change sequence in the initial interval as output features, training and generating the corresponding prediction model, and retaining the prediction accuracy of each prediction model for each initial interval; Adjust the position of the initial segmentation point to change the size of the initial interval, re-establish the prediction model based on the changed initial interval, repeat this step to a preset number of times, screen out the prediction model with the highest prediction accuracy and the corresponding initial interval, and define the time period covered by the screened initial interval as the prediction time period.

5. The method according to claim 2, characterized in that: Selecting the first target strategy includes the following steps: Match the corresponding user profile according to the user information, obtain the representative sequences corresponding to the various temperature change modes under the user profile, generate an actual temperature sequence based on the historical room temperature, calculate the similarity between the actual temperature sequence and each representative sequence, and use the first segmentation strategy corresponding to the representative sequence with the highest similarity as the first target strategy.

6. The method according to claim 2, characterized in that Performing secondary clustering on the ambient temperature sequence comprises the following steps: The time domain features in each of the ambient temperature sequences are extracted, the time domain features are used as clustering features, and the ambient temperature sequences are secondary clustered using a K-mean clustering algorithm.

7. The method according to claim 6, characterized in that The time domain features include average value, peak-to-peak value, standard deviation, means, maximum value, and minimum value.

8. The method according to claim 6, characterized in that In the secondary clustering result, the ambient temperature sequence closest to the cluster center in the cluster is taken as the representative sequence.

9. The method according to claim 1, characterized in that: The heater includes a temperature priority mode and a humidity priority mode, and the heater includes multiple air outlets. When in the humidity priority mode and the current humidity is within a first range, the heater opens 50% of the air outlets. When the current humidity is within a second range, the heater opens 70% of the air outlets.

10. The method according to claim 1, characterized in that The room information includes room types, and each room type is pre-set with the corresponding target temperature.

Images