Search

CN-121979338-A - AI self-adaptive algorithm-based water dispenser temperature control method

CN121979338ACN 121979338 ACN121979338 ACN 121979338ACN-121979338-A

Abstract

The invention discloses a water dispenser temperature control method based on an AI self-adaptive algorithm, which relates to the technical field of intelligent household appliance control and comprises the steps of acquiring an environment monitoring data stream and an equipment running state parameter stream in real time; the method comprises the steps of carrying out time sequence alignment and feature fusion on the target temperature and the equipment thermal inertia parameters to generate a user water behavior feature vector and an equipment working condition feature vector, inputting the feature vector into a pre-trained self-adaptive temperature control decision model, outputting a target heating temperature set value, a preheating starting advance and a step power adjustment strategy, calculating a real-time power compensation quantity according to the target temperature and the equipment thermal inertia parameters, dynamically correcting the step power strategy based on the compensation quantity to generate a power control instruction sequence, and sending the power control instruction sequence to a heating controller for execution. According to the invention, intelligent prediction and self-adaptive decision of water demand are realized by fusing multi-source data, and dynamic power compensation is performed by combining thermal inertia of equipment, so that the instant hot water demand of a user is accurately met, and meanwhile, the energy consumption is effectively reduced.

Inventors

  • DU TIANHAI
  • CAO SHICHEN

Assignees

  • 南京及时水智能科技有限公司
  • 南京脑一代智能科技有限公司

Dates

Publication Date
20260505
Application Date
20260208

Claims (10)

  1. 1. The temperature control method of the water dispenser based on the AI self-adaptive algorithm is characterized by comprising the following steps: the method comprises the steps of acquiring an environment monitoring data stream and an equipment operation state parameter stream output by a water dispenser sensor network in real time, wherein the environment monitoring data stream comprises an environment temperature, an environment humidity and a surrounding human body movement signal intensity; Performing time sequence alignment and feature fusion processing on the environment monitoring data stream and the equipment operation state parameter stream to generate a water behavior feature vector and an equipment working condition feature vector of a user; inputting the water behavior feature vector and the equipment working condition feature vector of the user into a pre-trained self-adaptive temperature control decision model for calculation, and outputting a target heating temperature set value, a preheating starting advance and a step power adjustment strategy; Calculating real-time power compensation quantity according to the target heating temperature set value and the thermal inertia parameters in the equipment working condition characteristic vector; And dynamically correcting the step power adjustment strategy based on the real-time power compensation quantity, generating a power control instruction sequence, and sending the power control instruction sequence to a heating controller of the water dispenser for execution.
  2. 2. The AI-adaptive algorithm-based water dispenser temperature control method of claim 1, wherein the performing time sequence alignment and feature fusion processing on the environmental monitoring data stream and the device operation state parameter stream to generate a user water behavior feature vector and a device working condition feature vector comprises: The water behavior feature vector of the user comprises predicted water consumption time, predicted water temperature preference and water consumption estimated value, and the equipment working condition feature vector comprises thermal inertia parameters, heat loss rate and heating efficiency attenuation coefficient; Establishing a unified time axis taking a system clock of the water dispenser as a reference, and mapping each data point in the environment monitoring data stream and each data point in the equipment running state parameter stream onto the unified time axis; According to the fluctuation mode of the surrounding human body movement signal intensity, identifying a behavior time point sequence of a user approaching the water dispenser; Matching and correlating the behavior time point sequence with the actual water receiving time in the water use history record of the current period, and extracting the environmental temperature before and after each water use event, the real-time temperature change curve of the heating cavity and the water use amount; Analyzing the data of a plurality of continuous water use events by using a time sequence analysis algorithm, extracting preference rules of users for water temperature in different periods of time under different environmental temperatures and average interval duration between two adjacent water use times, and generating predicted water use time, predicted water temperature preference and water use amount estimated values; And calculating the thermal inertia parameter, the heat loss rate and the heating efficiency attenuation coefficient through thermodynamic model fitting based on the heating power history record and the heating cavity real-time temperature change curve.
  3. 3. The method for controlling the temperature of the water dispenser based on the AI self-adaptive algorithm as set forth in claim 2, wherein the analyzing the data of the continuous plurality of water use events by using the time sequence analysis algorithm to extract the preference rule of the user for the water temperature in different periods under different environmental temperatures comprises: Constructing a hidden Markov model which takes the ambient temperature and the time point in the day as input and takes the actual water temperature finally selected by a user as output; training the hidden Markov model by utilizing the ambient temperature, the heating cavity real-time temperature change curve and water use time point data in the historical water use event; Inputting the environmental temperature and the time point at the current moment into a trained hidden Markov model, and outputting corresponding water temperature state probability distribution by the model; And selecting a water temperature value with highest probability in the water temperature probability distribution as a core value of the predicted water temperature preference, and calculating the variance of the probability distribution as the confidence coefficient of the predicted water temperature preference.
  4. 4. The AI-adaptive algorithm-based water dispenser temperature control method of claim 2, wherein the inputting the user water behavior feature vector and the device operating condition feature vector into a pre-trained adaptive temperature control decision model for calculation, outputting a target heating temperature set value, a warm-up start advance and a step power adjustment strategy, comprises: The step power adjustment strategy defines heating power curves adopted in different heating stages; inputting the predicted water consumption time and the predicted water temperature preference into a demand prediction sub-network of an adaptive temperature control decision model, wherein the demand prediction sub-network outputs an optimal water storage temperature range meeting the predicted water temperature preference; Inputting the thermal inertia parameters and the heat loss rate into a device dynamic sub-network of a decision model, and calculating a theoretical shortest time and a corresponding theoretical average power required for heating water from the current temperature to the lower limit of the optimal water storage temperature range by the device dynamic sub-network; Comparing the theoretical shortest time with the predicted water consumption time, and calculating the duration of heating needing to be started in advance, namely the preheating starting advance, so as to ensure that a user can obtain water meeting the preference water temperature in the predicted water consumption time; Combining the heating efficiency attenuation coefficient and a safety constraint condition, carrying out sectional optimization design on the theoretical average power to form the step power adjustment strategy consisting of high-power quick heating in an initial stage, smooth transition in a middle stage and low-power heat preservation in a final stage; And setting the central value of the optimal water storage temperature range as the target heating temperature set value.
  5. 5. The AI-adaptive algorithm-based water dispenser temperature control method of claim 4, wherein calculating the real-time power compensation amount based on the target heating temperature set point and the thermal inertia parameter in the device operating condition feature vector comprises: continuously monitoring the difference value between the real-time temperature of the heating cavity and the target heating temperature set value at preset time intervals, and recording the difference value as real-time temperature deviation; inputting the real-time temperature deviation and the thermal inertia parameter into a proportional-integral-derivative controller together; The proportional-integral-derivative controller adjusts control parameters according to thermal inertia parameters, wherein the larger the thermal inertia parameters are, the slower the system temperature response is, the weight of an integral term is increased to reduce steady-state errors; The proportional-integral-derivative controller calculates based on the adjusted control parameter and the real-time temperature deviation, and outputs the real-time power compensation quantity for dynamically eliminating the real-time temperature deviation.
  6. 6. The AI-adaptive algorithm-based water dispenser temperature control method of claim 5, wherein dynamically modifying the step power adjustment strategy based on the real-time power compensation amount to generate a power control command sequence comprises: acquiring a preset reference power value corresponding to the current heating stage in the step power adjustment strategy; Superposing the real-time power compensation quantity and the preset reference power value to obtain actual execution power at the current moment; Judging whether the actual execution power exceeds the upper limit or the lower limit of the safe operation of the heating element of the water dispenser; if the actual execution power exceeds the upper limit, limiting the actual execution power to a safe work upper limit value, and if the actual execution power is lower than the lower limit, limiting the actual execution power to a safe work lower limit value or turning off heating; and arranging the actual execution power values calculated and checked at each moment according to a time sequence to form the power control instruction sequence for directly driving the power adjusting unit of the heating controller.
  7. 7. The AI-adaptive algorithm-based water dispenser temperature control method of claim 1, further comprising an online learning update step of an adaptive temperature control decision model: After the power control instruction sequence is executed and one heating-water taking cycle is completed, collecting actual execution data of the heating-water taking cycle, wherein the actual execution data comprises a complete heating cavity real-time temperature change curve, actual total power consumption and final actual water receiving temperature of a user; Comparing the actual execution data with the predicted water temperature preference and the water consumption estimated value in the water consumption behavior feature vector of the user to generate model prediction error feedback data; the model prediction error feedback data, the corresponding environment monitoring data stream, the corresponding equipment running state parameter stream and the decision parameters output by the previous model are formed into a training sample; And performing incremental training on the pre-trained self-adaptive temperature control decision model by using the training sample so as to finely adjust model parameters, so that the subsequent prediction and decision of the model can be better adapted to the actual working condition of a specific water dispenser and the actual water consumption habit of a specific user.
  8. 8. The AI-adaptive algorithm-based water dispenser temperature control method of claim 7, wherein the incrementally training the pre-trained adaptive temperature control decision model using the training samples comprises: Taking the environment monitoring data stream and the equipment running state parameter stream in the training sample as input, and taking the model prediction error feedback data as a supervision signal; calculating a loss function gradient between the current output of the self-adaptive temperature control decision model and the ideal output based on actual execution data feedback by adopting a back propagation algorithm; updating the connection weights of all layers of the neural network in the model along the gradient descending direction of the loss function according to the preset increment learning rate; after the incremental learning of the predetermined number of training samples is completed, the performance of the updated model is evaluated using the cross-validation set, and if the performance improvement is below a threshold or a drop occurs, the model is rolled back to the model version before the update.
  9. 9. The water dispenser temperature control method based on the AI adaptive algorithm as claimed in claim 1, wherein the pre-trained adaptive temperature control decision model is obtained by: constructing a neural network model, wherein an input layer corresponds to the water behavior feature vector of the user and the equipment working condition feature vector, and an output layer corresponds to the target heating temperature set value, the preheating starting advance and the step power adjustment strategy parameter; Obtaining a plurality of groups of training samples, wherein each group of training samples comprises a historical environment monitoring data stream, a device running state parameter stream, a historical feature vector extracted from the historical environment monitoring data stream and the device running state parameter stream, and an optimal decision parameter which is optimized and verified in the historical scene; taking the difference between the decision parameters output by the minimized model and the optimal decision parameters as a loss function, and performing supervision training on the neural network model by using the plurality of groups of training samples; When the prediction precision of the model on the verification set reaches a preset threshold, training is completed, and the pre-trained self-adaptive temperature control decision model is obtained.
  10. 10. The AI-adaptive algorithm-based water dispenser temperature control method of claim 6, wherein the generation of the power control command sequence further takes into account an energy efficiency optimization objective: Establishing a function taking the total heating energy consumption as an optimization target, wherein the variable of the function is the power value and duration of each stage in the step power adjustment strategy; Under the constraint condition that water is heated to the target heating temperature set value and the preheating starting advance is guaranteed, a group of theoretical optimal power distribution schemes are obtained by solving the extremum of the function; And when the actual execution power is generated, taking the theoretical optimal power distribution scheme as a reference standard, so that the whole energy approaches to the theoretical optimal value while dynamically compensating the temperature deviation.

Description

AI self-adaptive algorithm-based water dispenser temperature control method Technical Field The invention belongs to the technical field of intelligent household appliance control, and particularly relates to a water dispenser temperature control method based on an AI self-adaptive algorithm. Background The current temperature control schemes of the water dispenser in the market mostly adopt switch control based on preset temperature or a simple timing heating strategy. The method mainly relies on feedback of a temperature sensor in a heating cavity, heating is started when the detected water temperature is lower than a set threshold value, and heating is stopped after the detected water temperature reaches the threshold value. Some of the improvements incorporate a timing function to maintain the heating state for a preset period of time to accommodate a substantially regular water usage time. These technical schemes constitute the mainstream practice of current water dispenser temperature control. The prior art solutions have drawbacks. The control logic is relatively isolated and static, only focuses on the instant water temperature state of the equipment, and completely ignores the dynamic influence of external environment change and the actual water use behavior mode of the user. The existing control method can not sense the factors, so that energy waste frequently occurs, such as continuous ineffective heat preservation in an unmanned period or a low-temperature environment, or enough hot water can not be prepared in advance before a water use peak comes, and user experience is affected. Meanwhile, the conventional method has insufficient consideration on the thermal inertia characteristic of the equipment, a power output mode is fixed, quick and accurate temperature control response is difficult to realize in the reheating process after frequent water taking, and temperature overshoot or heating delay is easy to cause. What is needed is an intelligent temperature control method that can integrate perceived environment and user behavior and adaptively and dynamically adjust control strategies. The method needs to solve the problems of how to effectively extract characteristics from multi-source heterogeneous time sequence data and predict requirements, and how to finely regulate and control power output according to real-time thermodynamic characteristics of equipment, so that the optimization of the whole energy consumption is realized while the real-time hot water requirements are ensured. Disclosure of Invention The present invention aims to solve at least one of the technical problems existing in the prior art; therefore, the invention provides a water dispenser temperature control method based on an AI self-adaptive algorithm, which comprises the following steps: the method comprises the steps of acquiring an environment monitoring data stream and an equipment operation state parameter stream output by a water dispenser sensor network in real time, wherein the environment monitoring data stream comprises an environment temperature, an environment humidity and a surrounding human body movement signal intensity; Performing time sequence alignment and feature fusion processing on the environment monitoring data stream and the equipment operation state parameter stream to generate a water behavior feature vector and an equipment working condition feature vector of a user; inputting the water behavior feature vector and the equipment working condition feature vector of the user into a pre-trained self-adaptive temperature control decision model for calculation, and outputting a target heating temperature set value, a preheating starting advance and a step power adjustment strategy; Calculating real-time power compensation quantity according to the target heating temperature set value and the thermal inertia parameters in the equipment working condition characteristic vector; And dynamically correcting the step power adjustment strategy based on the real-time power compensation quantity, generating a power control instruction sequence, and sending the power control instruction sequence to a heating controller of the water dispenser for execution. Further, the performing time sequence alignment and feature fusion processing on the environmental monitoring data stream and the equipment operation state parameter stream to generate a water behavior feature vector and an equipment working condition feature vector for a user, including: The water behavior feature vector of the user comprises predicted water consumption time, predicted water temperature preference and water consumption estimated value, and the equipment working condition feature vector comprises thermal inertia parameters, heat loss rate and heating efficiency attenuation coefficient; Establishing a unified time axis taking a system clock of the water dispenser as a reference, and mapping each data point in the environment monitoring data str