Adaptive temperature control system based on support vector machine

Abstract

The invention relates to an adaptive temperature control system based on a support vector machine. The adaptive temperature control system comprises data environment acquisition and preprocessing, an algorithm model, an adaptive energy efficiency control strategy, temperature adjustment and feedback, dynamic updating and optimization and interface display. The data environment acquisition and preprocessing comprises the steps of acquiring environment data, such as indoor temperature, humidity, illumination and the like, collected by a sensor network, and performing data cleaning and normalization on the acquired data. The algorithm model comprises the steps of constructing a training data set, making a decision function and solving an optimization problem. The self-adaptive energy efficiency control strategy comprises the step of formulating an energy efficiency control strategy according to the predicted optimized temperature set value. Temperature regulation and feedback comprise real-time monitoring of data and changes, and prediction of the model. The dynamic updating and optimizing comprises the steps of regularly collecting data after actual temperature adjustment, evaluating a model and training a data set for updating. The interface display comprises displaying temperature, humidity and illumination values received from a serial port, storing the values in a database and displaying the values in a broken line graph form.

Description

Technical Field

[0001] The present invention involves integrating SVM prediction and adaptive energy efficiency strategy, real-time monitoring of environmental data, optimizing temperature setting, and displaying real-time control effect through an intuitive visual interface to realize an intelligent and efficient indoor temperature management system, that is, an adaptive temperature control system based on support vector machine. Background Art

[0002] In today's intelligent and energy-saving and environmentally friendly environment, temperature control is particularly important in living and working environments. However, traditional temperature control systems often have problems such as low energy efficiency and inaccurate adjustment. To meet these challenges, adaptive energy-efficient temperature control dynamic support vector machine technology came into being.

[0003] Adaptive energy-efficient temperature control technology combines the support vector machine (SVM) prediction model with intelligent control strategies to achieve more efficient and accurate temperature control. As a machine learning algorithm, SVM can learn the complex relationship between the environment and temperature based on historical temperature, humidity, light and other data, thereby predicting future temperature changes. The adaptive energy-efficient strategy can dynamically adjust the temperature setting according to different time periods and usage needs, maximizing energy savings and improving indoor comfort.

[0004] By combining the SVM model with the adaptive energy efficiency strategy, a more intelligent and accurate temperature control is achieved. Compared with the traditional constant temperature control system, it can adjust the temperature in real time according to environmental changes, reduce energy consumption and save energy expenses. At the same time, the system also provides a visual interface, users can monitor indoor temperature changes in real time, and adjust according to feedback information, which enhances user participation and satisfaction.

[0005] Therefore, designing an adaptive temperature control system based on support vector machine can provide a new way to achieve energy saving, environmental protection and intelligent temperature management, which has important practical value and promotion significance. Summary of the invention

[0006] In view of this, the present invention provides an adaptive temperature control system based on support vector machine, which can integrate the prediction results of the SVM model with the adaptive energy efficiency control strategy. By analyzing the temperature trend predicted by SVM, the system can intelligently provide the trend of temperature change to minimize energy consumption while maintaining comfort. In addition, the system also has powerful data visualization capabilities. Users can monitor indoor environmental data and temperature control effects in real time through an intuitive interface, thereby enhancing users' perception and participation in energy efficiency management. In order to achieve the above purpose, the present invention adopts the following technical solutions:

[0007] The present invention relates to an adaptive temperature control system based on a support vector machine, comprising: data environment acquisition and preprocessing, algorithm model, adaptive energy efficiency control strategy, temperature regulation and feedback, dynamic update and optimization, and interface display. Data environment acquisition and preprocessing include collecting environmental data such as indoor temperature, humidity, light, etc. collected by a sensor network, and performing data cleaning and normalization on the collected data. The algorithm model includes constructing a training data set, a decision function, and solving optimization problems. The adaptive energy efficiency control strategy includes formulating an adaptive energy efficiency control strategy based on a predicted optimized temperature setting value. Temperature regulation and feedback include real-time monitoring of data and data changes, predicting the model to obtain an optimized temperature setting value so that the indoor temperature gradually approaches the optimized value. Dynamic update and optimization include regularly collecting data after actual temperature adjustment, and evaluating model performance and energy efficiency performance. According to the evaluation results, the training data set may be updated, including adding new data and eliminating abnormal data. The interface display includes displaying the temperature, humidity, and light values ​​received from the serial port stored in a database and displayed in the form of a line graph.

[0008] Preferably, in the above-mentioned adaptive temperature control system based on support vector machine, the data received from the serial port is transmitted as system environment data via the MQTT protocol based on the Internet of Things communication technology.

[0009] Preferably, in the above-mentioned adaptive temperature control system based on support vector machine, the environmental data are indoor temperature T, humidity H, light intensity L, and time information t. The three-times standard deviation method is used to process outliers, delete the rows containing missing values, and filter noise using weighted moving average. Using formula Normalize the data. Use the formula Encode time information.

[0010] Preferably, in the above-mentioned adaptive temperature control system based on support vector machine, a decision function is constructed: f(T, H, L, t) = z1·T+z2·H+z3·L+z4·t+q, the weight vector z and the bias term q are found, and the error term is minimized:

[0011] Preferably, in the above-mentioned adaptive temperature control system based on support vector machine, the formula Q(T o )=σ·E(T o )+τ·C(T o )+∈·O(T o ), where σ, τ, ∈ are weight coefficients used to balance the objectives. E(T o ) is the energy consumption model E(T o )=f·T o +g where f and g are model parameters. C(T o) is the comfort model, Where h and k are model parameters. This is achieved by adjusting energy utilization to minimize energy consumption.

[0012] Preferably, in the above-mentioned adaptive temperature control system based on support vector machine, the formula Adjust the temperature setting value, air conditioning temperature, etc. to gradually bring the indoor temperature closer to the optimized value.

[0013] Preferably, in the above-mentioned adaptive temperature control system based on support vector machine, the formula

[0014]

[0015]

[0016]

[0017] Calculate the partial derivatives of the loss function J(z, q) with respect to z and q, continuously adjust the model parameters, and gradually optimize to adapt to environmental changes and data characteristics to achieve better intelligent temperature control effects.

[0018] Preferably, in the above-mentioned adaptive temperature control system based on support vector machine, the trained model is used for prediction:

[0019] T o =sign(f(T,H,L,t))

[0020] f(T,G,L,t)=z1·T+z2·H+z3·L+z4·t+q

[0021] T o =sign(z1·T+z2·H+z3·L+z4·t+q)

[0022] Preferably, in the above-mentioned support vector machine-based adaptive temperature control system, the numerical values ​​of temperature, humidity and light are displayed and stored in a database so that they can be searched and displayed, and a line graph is used for visual management.

[0023] Preferably, in the above-mentioned support vector machine-based adaptive temperature control system, the system interface is displayed through an APP.

[0024] It can be seen from the above technical solution that compared with the prior art, the present invention provides an adaptive temperature control system based on support vector machine, which receives data through the serial port and transmits the data based on the MQTT protocol, and integrates the SVM model and the adaptive energy efficiency control strategy to visualize the data. The system achieves the dual goals of efficient energy management and user participation, and provides strong support for the realization of intelligent, comfortable and sustainable indoor environment management. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below.

[0026] Figure 1 The accompanying drawing is a general framework diagram of the present invention;

[0027] Figure 2 The accompanying drawing is a flow chart of the use of the Android APP of the present invention;

[0028] Figure 3 The accompanying drawing is a partial interface of the Android APP of the present invention. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0030] The embodiment of the present invention discloses an adaptive temperature control system based on a support vector machine, which can effectively integrate the SVM model with the adaptive energy efficiency strategy, and display and search the data. Since the user can monitor the indoor environment data and the temperature control effect in real time through an intuitive interface, the user's perception and participation in energy efficiency management is enhanced.

[0031] The present invention relates to an adaptive temperature control system based on a support vector machine, comprising: data environment acquisition and preprocessing, algorithm model, adaptive energy efficiency control strategy, temperature regulation and feedback, dynamic update and optimization, and interface display. Data environment acquisition and preprocessing include collecting environmental data such as indoor temperature, humidity, light, etc. collected by a sensor network, and performing data cleaning and normalization on the collected data. The algorithm model includes constructing a training data set, a decision function, and solving optimization problems. The adaptive energy efficiency control strategy includes formulating an adaptive energy efficiency control strategy based on a predicted optimized temperature setting value. Temperature regulation and feedback include real-time monitoring of data and data changes, predicting the model to obtain an optimized temperature setting value so that the indoor temperature gradually approaches the optimized value. Dynamic update and optimization include regularly collecting data after actual temperature adjustment, and evaluating model performance and energy efficiency performance. According to the evaluation results, the training data set may be updated, including adding new data and eliminating abnormal data. The interface display includes displaying the temperature, humidity, and light values ​​received from the serial port stored in a database and displayed in the form of a line graph.

[0032] In order to further optimize the above technical solution, the data received from the serial port is transmitted as system environment data through the MQTT protocol based on the Internet of Things communication technology.

[0033] In order to further optimize the above technical solution, the environmental data are indoor temperature T, humidity H, light intensity L, and time information t. The 3-times standard deviation method is used to process outliers, delete rows containing missing values, and use weighted moving average to filter noise. Using the formula Normalize the data. Use the formula Encode time information.

[0034] In order to further optimize the above technical solution, construct a decision function: f(T, H, L, t) = z1·T+z2·H+z3·L+z4·t+q, find the weight vector z and the bias term q, and minimize the error term:

[0035] In order to further optimize the above technical solution, the formula Q(T o )=σ·E(T o )+τ·C(T o )+∈·O(T o ), where σ, τ, ∈ are weight coefficients used to balance the objectives. E(T o ) is the energy consumption model E(T o )=f·T o +g where f and g are model parameters. C(T o ) is the comfort model, Where h and k are model parameters. This is achieved by adjusting energy utilization to minimize energy consumption.

[0036] In order to further optimize the above technical solution, the formula Adjust the temperature setting value, air conditioning temperature, etc. to gradually bring the indoor temperature closer to the optimized value.

[0037] In order to further optimize the above technical solution, the formula

[0038]

[0039]

[0040]

[0041] Calculate the partial derivatives of the loss function J(z, q) with respect to z and q, continuously adjust the model parameters, and gradually optimize to adapt to environmental changes and data characteristics to achieve better intelligent temperature control effects.

[0042] In order to further optimize the above technical solution, use the trained model for prediction:

[0043] T o =sign(f(T,H,L,t))

[0044] f(T, H, L, t)=z1·T+z2·H+z3·L+z4·t+q

[0045] T o =sign(z1·T+z2·H+z3·L+z4·t+q)

[0046] In order to further optimize the above technical solution, the numerical values ​​of temperature, humidity and light are displayed and stored in a database so that they can be searched and displayed, and a line chart can be used for visual management.

[0047] In order to further optimize the above technical solutions, the system interface is displayed through the APP.

[0048] Based on the process of data collection, SVM model prediction, adaptive control, temperature regulation, and dynamic optimization, the system achieves the goal of intelligent environmental control, improves energy efficiency and enhances user experience. Figure 1 shown.

[0049] Mobile APP is for users to use, and its usage flow chart is as follows Figure 2When the user enters the system and selects the data module, the user can choose to view the temperature / humidity / light values; in the viewing module, the user can view the latest 20 values ​​of temperature / humidity / light in the database; in the line chart module, the user can view line charts of different colors for temperature / humidity / light.

[0050] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. The above description is only a preferred embodiment of the present invention, and any changes and modifications made within the scope defined by the invention claims belong to the protection scope of the present invention.

Claims

1. The present invention relates to an adaptive temperature control system based on a support vector machine, comprising: Data environment collection and preprocessing, algorithm model, adaptive energy efficiency control strategy, temperature adjustment and feedback, dynamic update and optimization, interface display. Data environment collection and preprocessing includes collecting environmental data collected by the sensor network, such as indoor temperature, humidity, light, etc., and cleaning and normalizing the collected data. The algorithm model includes building a training data set, decision function, and solving optimization problems. The adaptive energy efficiency control strategy includes formulating an adaptive energy efficiency control strategy based on the predicted optimal temperature setting value. Temperature adjustment and feedback include real-time monitoring of data and data changes, and predicting the model to obtain optimized temperature setting values ​​so that the indoor temperature gradually approaches the optimized value. Dynamic update and optimization include regular collection of data after actual temperature adjustment, and evaluation of model performance and energy efficiency. Based on the evaluation results, the training data set may be updated, including adding new data and removing abnormal data. The interface display includes displaying the temperature, humidity, and light values ​​received from the serial port, stored in the database, and presented in the form of a line graph.

2. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: The data received from the serial port is transmitted as system environment data through the MQTT protocol based on the Internet of Things communication technology.

3. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: Data environment collection includes the collected environmental data indoor temperature T, humidity H, light intensity L, and time information t. Preprocessing: First, data cleaning is performed to deal with outliers using the 3-times standard deviation method, delete rows containing missing values, and filter noise using weighted moving average. Secondly, different sensors have different data ranges. In order to put them on a unified scale, the data is normalized. T, H, L, and t are scaled to the range of 0 to 1 using the following formula, and T, H, L, and t are normalized to T′, H′, L′, and t′, where min(T), min(H), min(L), and min(t) represent the minimum data values, and max(T), max(H), max(L), and max(t) represent the maximum data values. Finally, the time information is encoded using the formula: where t i is the time information, and m is the time range.

4. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: The training of the algorithm model is divided into: Building a training dataset Where T i , H i , L i , t i are input features, is the corresponding optimized temperature setting value. Predict the optimal temperature setting value T o , construct the decision function: f(T, H, L, t) = z1·T+z2·H+z3·L+z4·t+q, where z1, z2, z3, z4 are the components of the weight vector, and q is the bias term. Solve the algorithm optimization problem to find the weight vector z and the bias term q, and minimize the error term: where ||z|| 2 Represents the L2 norm of the weight vector z, the specific form is: C is the regularization parameter used to balance the margin and misclassified points. M is the number of training samples.

5. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: The optimized temperature setting value T predicted by the algorithm model o , formulate an adaptive energy efficiency control strategy. This is achieved by adjusting energy utilization to minimize energy consumption. First, define an energy efficiency objective function Q: Q(T o )=σ·E(T o )+τ·C(T o )+∈·O(T o ), where σ, τ, ∈ are weight coefficients used to balance the objectives. E(T o ) is the energy consumption model, energy consumption and T o Establish a linear energy consumption model to estimate energy consumption: E(T o )=f·T o +g where f and g are model parameters. C(T o ) is the comfort model, and the comfort is inversely proportional to the temperature difference: Where h and k are model parameters. O(T o ) is a model of other factors, which takes into account the influence of other aspects such as environmental requirements.

6. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: In temperature regulation and feedback, real-time monitoring of environmental data T, H, L, t, including changes in indoor temperature, is performed. The trained algorithm model is used for prediction to obtain the optimized temperature setting value T. o According to T o And the current actual temperature, by adjusting the temperature setting value, air conditioning temperature, etc., the indoor temperature can be gradually approached to the optimized value. This can be expressed using the following formula: Where ΔT a Indicates the temperature adjustment amount according to the system prompt, T o represents the optimized temperature setting value predicted by the algorithm model, T c Indicates the current actual temperature.

7. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: The algorithm model is dynamically updated and optimized according to the actual situation. The weight vector z and the bias term q are continuously updated using the gradient descent method to achieve dynamic control strategy optimization. First, the partial derivatives of the loss function J(z, q) with respect to z and q are calculated, and the parameters are adjusted in the opposite direction according to the gradient. The partial derivative with respect to z is: Partial derivative with respect to q: Use the gradient descent method to update the parameters, define the learning rate θ, and update the rule: Continuously adjust model parameters and gradually optimize to adapt to environmental changes and data characteristics to achieve better intelligent temperature control effects.

8. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: In the control, according to the current environmental data T, H, L, t, the trained model is used for prediction: T o =sign(f(T、H、L、t)) f(T, H, L, t)=z1·T+z2·H+z3·L+z4·t+q T o =sign(z1·T+z2·H+z3·L+z4·t+q) Finally, a symbolic function is used to map this predicted value to the actual temperature adjustment direction, thereby realizing the prediction and adjustment of intelligent temperature control.

9. The adaptive temperature control system based on support vector machine according to claim 1, characterized in that: The interface is displayed as an APP, showing the received temperature, humidity, and light, and storing the data in the database for display and search. The line chart data comes from the database to realize the visualization of the system and prompt the user with the current optimal temperature.

Images