Electric fireplace energy consumption dynamic optimization method and system

Abstract

The invention relates to the technical field of smart home and energy management, and discloses an electric fireplace energy consumption dynamic optimization method and system.The method comprises the steps that temperature data and weather data are obtained and preprocessed, and a first sequence is obtained; performing state updating processing on the first sequence and the like by using a Kalman filtering algorithm to obtain a second sequence; calculating the second sequence by combining a preset heat conduction parameter, a radiation value and a preset surface area parameter to obtain first heat loss; determining a correction coefficient by using the historical heat loss sequence to perform weighting operation on the first heat loss to obtain second heat loss; constructing an equilibrium equation based on the second heat loss and the temperature measured value for deduction calculation to obtain a limit temperature; switching logic is obtained according to matching of the temperature difference value and the operation state, a preset future time window is segmented to obtain a control interval, and a control scheme is obtained through distribution calculation; and performing evaluation calculation according to the control scheme and the actually measured temperature difference to obtain calibration data, and updating a preset heat conduction parameter. And flexible consumption reduction and stable room temperature regulation and control of the electric fireplace are realized.

Description

Technical Field

[0001] This invention relates to the field of smart home and energy management technology, and in particular to a method and system for dynamic optimization of energy consumption of electric fireplaces. Background Technology

[0002] Currently, with the popularization of smart home technology, energy-saving control of high-power heating equipment such as electric fireplaces has gradually become an important part of residential energy management. Electric fireplaces not only serve the function of localized physical heating, but their operation and control strategies are also directly related to the overall building's energy consumption level and the user's thermal comfort experience.

[0003] In existing technologies, a common approach is to rely on basic IoT data acquisition terminals to obtain indoor temperature in real time, employing start-stop control based on a fixed temperature threshold or a simple PID feedback regulation strategy. Specifically, when the sensor detects that the indoor temperature is below the set comfort lower limit, the controller instructs the electric fireplace to turn on at full power; when the temperature reaches the set target, the power is directly cut off or switched to the lowest standby power. However, this traditional control logic treats the room as a closed black box with constant thermal characteristics, severely neglecting the dynamic changes in the thermodynamic properties of the building environment. In reality, the real-time heat loss of a room is subject to the combined interference of various external factors such as sudden changes in outdoor temperature, fluctuations in meteorological conditions, and nonlinear solar radiation. Because existing methods rely solely on a single a posteriori temperature feedback, they lack accurate quantitative assessment of the heat flow conducted through the building foundation and the dynamic heat loss from the environment, and also fail to establish a thermodynamic evolution model that includes residual heat boundary conditions. This results in the controller being completely unable to anticipate the extreme temperature trend after the equipment adjusts its power. Often, it only issues a shutdown command when the temperature reaches the target level. At this point, the equipment's enormous thermal inertia and residual heat continue to dissipate, causing severe temperature overshoot (overheating). During the heat preservation phase, the inability to predict the rate of heat loss from the environment leads to frequent start-ups and shutdowns of the equipment. This blind and lagging response mechanism greatly negates the energy efficiency advantages of the hardware itself.

[0004] Therefore, existing technologies suffer from serious energy waste in electric fireplaces and poor stability in indoor temperature control. Summary of the Invention

[0005] This invention provides a method and system for dynamic optimization of energy consumption in electric fireplaces, in order to solve the problems of serious energy waste and poor stability of indoor temperature control in existing electric fireplaces.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a method for dynamic optimization of energy consumption in an electric fireplace, comprising:

[0007] Temperature and weather data are acquired and preprocessed to obtain the first sequence;

[0008] The first sequence, the preset transition matrix, and the preset observation matrix are processed using the Kalman filter algorithm to obtain the second sequence.

[0009] The radiation value is obtained, and a mapping operation is performed on the second sequence using a preset heat conduction parameter to obtain the conduction heat flow. The conduction heat flow is compensated according to the radiation value to obtain the comprehensive heat flow. The comprehensive heat flow is multiplied by a preset surface area parameter to obtain the first heat loss.

[0010] A historical heat loss sequence is obtained. If the first heat loss is greater than a preset heat loss threshold, the historical heat loss sequence is decomposed using the STL algorithm to obtain a trend component and a fluctuation component. A correction coefficient is determined based on the trend component and the fluctuation component. The first heat loss is weighted using the correction coefficient to obtain the second heat loss.

[0011] Obtain the measured temperature value, construct an equilibrium equation based on the second heat loss and the measured temperature value, and perform derivation calculations on the equilibrium equation using a numerical solution algorithm to obtain the limiting temperature;

[0012] The system obtains the operating status. If the extreme temperature is less than the preset target temperature, it calculates the temperature difference between the preset target temperature and the extreme temperature. It obtains the switching logic by matching the temperature difference with the operating status. It performs a segmentation operation on the preset future time window according to the switching logic and the preset time step to obtain the control interval. It performs an allocation calculation on the control interval to obtain the control scheme.

[0013] Obtain the measured temperature difference, perform evaluation calculations based on the control scheme and the measured temperature difference to obtain calibration data, and update the preset heat conduction parameters using the calibration data.

[0014] Secondly, the present invention provides a dynamic energy consumption optimization system for electric fireplaces, comprising:

[0015] The sequence processing module is used to acquire temperature data and weather data, and perform preprocessing to obtain the first sequence;

[0016] The state update module is used to perform state update processing on the first sequence, the preset transition matrix, and the preset observation matrix using the Kalman filter algorithm to obtain the second sequence;

[0017] The heat loss calculation module is used to obtain radiation values, perform mapping operations on the second sequence using preset heat conduction parameters to obtain conductive heat flow, perform compensation calculations on the conductive heat flow based on the radiation values ​​to obtain comprehensive heat flow, and perform product operations on the comprehensive heat flow in combination with preset surface area parameters to obtain the first heat loss;

[0018] The dynamic correction module is used to obtain historical heat loss sequences. If the first heat loss is greater than a preset heat loss threshold, the STL algorithm is used to decompose the historical heat loss sequence to obtain trend components and fluctuation components. A correction coefficient is determined based on the trend components and fluctuation components. The correction coefficient is used to perform a weighted operation on the first heat loss to obtain the second heat loss.

[0019] The evolution deduction module is used to obtain the measured temperature value, construct an equilibrium equation based on the second heat loss and the measured temperature value, and perform deduction calculations on the equilibrium equation using a numerical solution algorithm to obtain the limiting temperature.

[0020] The scheme generation module is used to obtain the operating status. If the extreme temperature is less than the preset target temperature, the module calculates the temperature difference between the preset target temperature and the extreme temperature. The module obtains the switching logic by matching the temperature difference with the operating status. The module performs a segmentation operation on the preset future time window according to the switching logic and the preset time step to obtain the control interval. The module performs an allocation calculation on the control interval to obtain the control scheme.

[0021] The feedback calibration module is used to acquire the measured temperature difference and perform evaluation calculations based on the control scheme and the measured temperature difference. Calibration data is obtained, and the preset heat conduction parameters are updated using this calibration data.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] (1) This invention obtains a second sequence by filtering and aligning temperature and weather data, performing state updates using Kalman filtering, and then performing mapping and compensation calculations using preset heat conduction parameters and radiation values. Finally, a product operation is performed to obtain the first heat loss. This process deeply incorporates multi-source heterogeneous environmental data (indoor and outdoor temperature difference and solar radiation) into the thermodynamic evaluation system, effectively filtering out high-frequency noise and measurement errors in the sensor acquisition process using Kalman filtering. At the same time, by superimposing heat conduction parameters with clear physical meaning and real-time radiation values, the interference of nonlinear external environmental factors on the room's background thermal characteristics is accurately removed, restoring the building's true combined conduction and radiation heat flow. This scheme effectively overcomes the data distortion and environmental perception lag problems caused by relying solely on single temperature feedback in existing technologies, significantly improving the physical fidelity and calculation accuracy of real-time identification of building foundation heat loss.

[0024] (2) This invention, when the first heat loss exceeds a preset threshold, uses the STL algorithm to decompose the historical heat loss sequence to obtain trend and fluctuation components, thereby determining correction coefficients. These correction coefficients are then used to perform a weighted calculation on the first heat loss to obtain the second heat loss. This process abandons the traditional approach of treating historical data as static constants, and innovatively employs time series decomposition technology to mathematically decouple long-term climate variations (such as the trend component of seasonal cooling) hidden in historical heat loss from diurnal or sudden weather disturbances (such as the fluctuation component of gusts). By constructing adaptive correction coefficients to weight and correct the instantaneously calculated first heat loss, it can effectively filter out accidental calculation errors caused by extreme sudden conditions and compensate for the building's own heat storage attenuation law to the current state. This scheme ensures that the final output heat loss assessment value balances real-time agility and the robustness of historical data, greatly enhancing the system's anti-interference capability and dynamic adaptive capability in the face of complex and variable weather conditions.

[0025] (3) This invention constructs a balance equation based on the second heat loss and measured temperature, uses a numerical solution algorithm to perform deductive calculations to obtain the limit temperature, and generates a control scheme based on the temperature difference and the switching logic matching the operating state. Finally, it uses the calibration data of the estimated and measured temperature difference to update the preset heat transfer parameters. This process constructs a thermodynamic evolution deduction model with temperature as the initial state and high-fidelity dynamic heat loss as the boundary condition, replacing blind empirical prediction. By predicting the steady-state limit temperature of the electric fireplace under future evolution, the system can proactively formulate a segmented power injection scheme before the actual room temperature exceeds the limit. At the same time, the innovative error feedback calibration mechanism enables the underlying heat transfer parameters to evolve autonomously with the aging of the physical environment. This scheme completely eliminates the overheating and frequent start-stop phenomena caused by traditional passive temperature control, achieving proactive and flexible energy saving and consumption reduction of the electric fireplace while ensuring high stability of indoor temperature control and the ultimate thermal comfort experience for users. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the dynamic energy consumption optimization method for electric fireplaces provided in the first embodiment of the present invention;

[0027] Figure 2 This is a schematic diagram of the structure of the electric fireplace energy consumption dynamic optimization system provided in the second embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Reference Figure 1 The first embodiment of the present invention provides a method for dynamic optimization of energy consumption of electric fireplaces, including the following steps:

[0030] S11: Acquire temperature and weather data, perform preprocessing, and obtain the first sequence;

[0031] S12, the Kalman filter algorithm is used to perform state update processing on the first sequence, the preset transition matrix and the preset observation matrix to obtain the second sequence;

[0032] S13, obtain radiation values, perform mapping operation on the second sequence using preset heat conduction parameters to obtain conduction heat flow, perform compensation calculation on the conduction heat flow according to the radiation values ​​to obtain comprehensive heat flow, and perform product operation on the comprehensive heat flow in combination with preset surface area parameters to obtain the first heat loss;

[0033] S14, obtain the historical heat loss sequence. If the first heat loss is greater than the preset heat loss threshold, use the STL algorithm to decompose the historical heat loss sequence to obtain the trend component and the fluctuation component. Determine the correction coefficient based on the trend component and the fluctuation component. Use the correction coefficient to perform a weighted operation on the first heat loss to obtain the second heat loss.

[0034] S15, obtain the measured temperature value, construct a balance equation based on the second heat loss and the measured temperature value, and use a numerical solution algorithm to perform deduction calculation on the balance equation to obtain the limit temperature;

[0035] S16, obtain the operating status. If the extreme temperature is less than the preset target temperature, calculate the temperature difference between the preset target temperature and the extreme temperature. Based on the temperature difference and the operating status, obtain the switching logic. Based on the switching logic and the preset time step, perform a segmentation operation on the preset future time window to obtain the control interval. Perform an allocation calculation on the control interval to obtain the control scheme.

[0036] S17, obtain the measured temperature difference, perform evaluation calculations based on the control scheme and the measured temperature difference to obtain calibration data, and update the preset heat conduction parameters using the calibration data.

[0037] In step S11, temperature data and weather data are acquired and preprocessed to obtain a first sequence, including:

[0038] Obtain temperature and weather data;

[0039] The temperature data was subjected to anomaly removal using the 3-Sigma criterion to obtain the baseline temperature.

[0040] The base temperature is smoothed using a moving average algorithm to obtain a smoothed temperature.

[0041] Extract the timestamp labels from the weather data;

[0042] Alignment is performed between the smoothed temperature and the weather data based on the timestamp label to obtain a first sequence.

[0043] In one implementation, temperature data and weather data are acquired. It should be noted that, through a physically deployed temperature sensor network, analog electrical signals are periodically collected according to a preset temperature sampling frequency. Analog-to-digital conversion is then performed to output a discrete numerical sequence, thus obtaining the temperature data. In this embodiment, structured messages are retrieved from a pre-established meteorological server via a network communication interface according to a preset weather update cycle, and the corresponding meteorological characteristic values ​​are parsed to obtain the weather data. It is worth noting that the preset temperature sampling frequency is determined by statistically analyzing the maximum cutoff frequency at which the thermodynamic state of the target room undergoes significant changes in historical natural years, multiplying this maximum cutoff frequency by a constant two, to obtain the preset temperature sampling frequency.

[0044] In one implementation, the 3-Sigma criterion is used to perform anomaly removal processing on the temperature data to obtain the baseline temperature. Specifically, firstly, an observation time window with a preset length is constructed, and data segments are extracted by sliding within the temperature data sequence. Then, the arithmetic mean and standard deviation of all temperature values ​​within the observation time window are calculated. The arithmetic mean is added to the product of a constant 3 and the standard deviation, and an addition operation is performed to obtain a dynamic upper limit threshold; the arithmetic mean is subtracted from the product of a constant 3 and the standard deviation, and a subtraction operation is performed to obtain a dynamic lower limit threshold. In this embodiment, each temperature data point within the observation time window is traversed. If the value of a temperature data point is greater than the dynamic upper limit threshold or less than the dynamic lower limit threshold, the temperature data point is determined to be an anomaly and deleted. For the data time sequence gaps caused by deletion, the valid temperature value of the previous time node is extracted and a forward filling operation is performed to output the continuous baseline temperature. It is worth noting that the length parameter in the observation time window with the preset length is determined by performing Fourier spectrum analysis on historical abnormal sensor logs, extracting the average duration period of the impulse interference, and then determining the length of the observation time window as five times the average duration period.

[0045] For example, a smoothing process is performed on the base temperature using a moving average algorithm to obtain a smoothed temperature. Specifically, a preset smoothing step size is set on the time axis. Within each step size interval, the weighted arithmetic mean of the multiple base temperatures is calculated, and this weighted arithmetic mean is mapped to the smoothed temperature at the center time node of that interval. It is worth noting that the preset smoothing step size is determined by calculating the autocorrelation function model of the historically collected base temperature sequence and extracting the time lag corresponding to the first decrease of the autocorrelation coefficient to the reciprocal of the natural constant. Since this time lag statistically accurately characterizes the decorrelation feature inflection point of the memory property of the time series signal, this embodiment determines it as the preset smoothing step size, thereby filtering out high-frequency thermal noise disturbances.

[0046] In one implementation, when extracting the timestamp tags from the weather data, the header fields of the structured message of the weather data are parsed to extract the absolute time information of the data generation time, and then converted into the timestamp tags in a unified format. In this embodiment, the smoothed temperature and the weather data are aligned based on the timestamp tags to obtain a first sequence.

[0047] It should be noted that, due to the frequency difference between the high-frequency acquisition attribute of temperature data and the low-frequency update attribute of weather data, the time series node of the smoothed temperature is used as a high-frequency reference time axis to map the relative position of the timestamp label of the weather data on the high-frequency reference time axis. For high-frequency time nodes between two adjacent timestamp labels, a linear interpolation algorithm is used to obtain the weather feature values ​​corresponding to these two timestamp labels. A linear weighted calculation is performed based on the time ratio distance between the current high-frequency time node and the two end timestamp labels to output the supplementary weather value for the corresponding high-frequency time node. Finally, in this embodiment, the smoothed temperature and the supplementary weather value under the same high-frequency time node are subjected to a dimension concatenation operation to generate the first sequence containing a multi-dimensional feature vector.

[0048] In step S12, the Kalman filter algorithm is used to perform state update processing on the first sequence, the preset transition matrix, and the preset observation matrix to obtain the second sequence, including:

[0049] The predicted state is obtained by performing prior estimation calculation on the first sequence using a preset transition matrix;

[0050] The Kalman gain is calculated by combining the preset observation matrix with the predicted state;

[0051] The predicted state is updated using the Kalman gain to obtain the second sequence.

[0052] In one implementation, a priori estimation calculation is performed on the first sequence using a preset transition matrix to obtain the predicted state. It should be noted that the preset transition matrix is ​​a pre-constructed block augmented matrix. The main diagonal block region of this block augmented matrix contains autoregressive coefficients characterizing the indoor temperature decay law, while the off-diagonal block region contains cross-mapping coefficients characterizing the influence of weather features such as outdoor temperature and wind speed on indoor temperature. The preset transition matrix is ​​determined by extracting the historical multidimensional state dataset of the target room, fitting an augmented state space model using a subspace system identification algorithm, extracting the constant coefficient matrix of the augmented state space model, and determining this constant coefficient matrix as the preset transition matrix. Finally, matrix multiplication is performed between the preset transition matrix and the multidimensional feature vector of the previous time node in the first sequence to output the predicted state of the current time node.

[0053] In one implementation, the Kalman gain is calculated by combining a preset observation matrix with the predicted state. It is worth noting that, since the first sequence contains multidimensional features of temperature and weather, while the physical temperature sensor can only observe temperature, mathematical dimensionality reduction matching is performed on the observation dimensions. Specifically, the proportional mapping relationship between the output electrical signal of the temperature sensor in the standard constant temperature chamber and the standard physical temperature value is collected, and this proportional mapping relationship is converted into diagonal coefficients. Subsequently, based on the multidimensional feature structure of the first sequence, a row vector mapping matrix is ​​constructed that has the diagonal coefficients only in the dimension corresponding to the indoor temperature value, and zero values ​​in all other weather feature dimensions. This row vector mapping matrix is ​​determined as the preset observation matrix. In this embodiment, the preset error covariance matrix of the previous time node is extracted, the preset transition matrix and the preset error covariance matrix are multiplied, and a preset process noise covariance matrix is ​​added to obtain the updated preset error covariance matrix. Substitute the updated preset error covariance matrix, the preset observation matrix, and the preset measurement noise covariance matrix into the Kalman filter gain equation, and perform matrix inversion and matrix multiplication operations in sequence to obtain the Kalman gain.

[0054] For example, the Kalman gain is used to perform an update calculation on the predicted state to obtain a second sequence. Specifically, the actual observed one-dimensional temperature value at the current time node is extracted from the first sequence. A matrix multiplication operation is performed between the preset observation matrix and the predicted state to obtain the predicted observation value. The predicted observation value is then subtracted from the actual observed one-dimensional temperature value to obtain the residual value. A matrix multiplication operation is performed between the Kalman gain and the residual value, and the result is then added to the predicted state to output the second sequence containing multi-dimensional calibration data after state update.

[0055] In step S13, radiation values ​​are obtained, and a mapping operation is performed on the second sequence using preset heat conduction parameters to obtain conductive heat flow. Compensation calculations are then performed on the conductive heat flow based on the radiation values ​​to obtain a comprehensive heat flow. Finally, a product operation is performed on the comprehensive heat flow using preset surface area parameters to obtain a first heat loss, including:

[0056] Obtain radiation values;

[0057] Extract the indoor and outdoor temperature values ​​from the second sequence;

[0058] Calculate the characteristic difference between the indoor temperature value and the outdoor temperature value;

[0059] The difference features are weighted and mapped using preset heat conduction parameters to obtain the conductive heat flow;

[0060] The combined heat flux is obtained by performing a linear superposition calculation on the conducted heat flux using the radiation values.

[0061] The first heat loss is obtained by multiplying the combined heat flow with the preset surface area parameters.

[0062] In one implementation, a total radiation sensor deployed in the building's external environment collects analog voltage signals in real time. After performing analog-to-digital conversion, the digital signal is multiplied by a preset photothermal conversion coefficient to output the radiation value. The preset photothermal conversion coefficient is determined based on the sensitivity calibration function table in the sensor's manufacturer's specifications. When extracting indoor and outdoor temperature values ​​from the second sequence, the indoor temperature value representing the indoor ambient temperature and the outdoor temperature value representing the outdoor ambient meteorological temperature are analyzed and separated by traversing the multidimensional feature data structure of the second sequence.

[0063] In one implementation, the difference feature is determined by subtracting the indoor temperature value from the outdoor temperature value. It is worth noting that a weighted mapping calculation is performed on the difference feature using a preset thermal conductivity parameter to obtain the conductive heat flow. The preset thermal conductivity parameter is determined by obtaining the building material type and wall thickness physical measurement values ​​of the target room's external envelope structure, matching and retrieving the basic thermal conductivity of the building material type in a building standard thermal database, dividing the basic thermal conductivity by the physical measurement value of the wall thickness, and performing a division calculation to obtain the preset thermal conductivity parameter. The preset thermal conductivity parameter is used as a weighting coefficient and a scalar multiplication operation is performed with the difference feature to calculate and output the conductive heat flow characterizing the heat transfer rate per unit area.

[0064] For example, the conducted heat flow is calculated by linear superposition using the radiation values ​​to obtain the comprehensive heat flow. It is important to note that, to ensure the rigor of the thermodynamic calculations in terms of physical principles, this embodiment pre-establishes a sign rule for heat flow, specifying that values ​​indicating heat flowing out of the indoor environment are positive, and values ​​indicating heat flowing into the indoor environment are negative. Based on this sign rule, when the indoor temperature is higher than the outdoor temperature, the conducted heat flow is positive; while the radiation values ​​input from an external solar heat source are converted to negative values. The conducted heat flow with its positive and negative signs is then algebraically added to the radiation values, and the combined output is the comprehensive heat flow representing the total net heat exchange per unit area.

[0065] It is worth noting that a preset surface area parameter is obtained, and the comprehensive heat flow is multiplied by the preset surface area parameter to obtain the first heat loss. Specifically, a 3D laser scanner is used to collect the 3D point cloud size data of all external envelope structures of the target room that are physically in contact with the outdoor environment, and the total area of ​​all physical contact surfaces is calculated. This total area is determined as the preset surface area parameter. The comprehensive heat flow is then multiplied by the preset surface area parameter to output the first heat loss, which represents the total power of the overall heat loss of the room at the current time point.

[0066] In step S14, a historical heat loss sequence is obtained. If the first heat loss is greater than a preset heat loss threshold, the historical heat loss sequence is decomposed using the STL algorithm to obtain a trend component and a fluctuation component. A correction coefficient is determined based on the trend component and the fluctuation component. The first heat loss is weighted using the correction coefficient to obtain the second heat loss.

[0067] It should be noted that in this embodiment, the calculation results of all continuous historical heat loss characteristics within a preset time window from the current time point are extracted from local memory and spliced ​​together in chronological order to construct the historical heat loss sequence. It is particularly important to note that the length parameter of the preset time window is determined based on the thermal inertia time constant of the building envelope, specifically extracted to be three times the calibrated thermal inertia time constant. It is also worth noting that after three time constants, the step decay rate of the transient response of a first-order thermodynamic system can reach 95%. This time span is theoretically sufficient to completely cover the entire decay cycle of a single sudden meteorological disturbance, thus ensuring that the historical sequence contains complete fluctuation characteristics.

[0068] In one implementation, if the first heat loss is greater than a preset heat loss threshold, the STL algorithm is used to decompose the historical heat loss sequence. It should be noted that the process of determining the preset heat loss threshold involves collecting background heat loss data samples of the target room under passive equipment-on conditions throughout a historical natural year. Based on these background heat loss data samples, a probability density function curve of kernel density estimation is constructed. The discrete value at the point where the cumulative probability of the probability density function curve reaches 95% is calculated, and this discrete value is determined as the preset heat loss threshold. The STL algorithm performs periodic term smoothing and removal operations through the inner loop and calculates polynomial fitting of low-frequency data points through the outer loop using a local weighted regression scatter smoothing method. This separates the trend component reflecting long-term climate variability and the fluctuation component reflecting short-term gusts or pressure disturbances from the historical heat loss sequence. It is worth noting that if the first heat loss is not greater than (i.e., less than or equal to) the preset heat loss threshold, the current weather is determined to be extremely stable and without sudden disturbances. In this case, the system directly assigns the value of the first heat loss and determines it as the second heat loss for output.

[0069] Specifically, a correction coefficient is determined based on the trend component and the fluctuation component, and the first heat loss is weighted using the correction coefficient to obtain the second heat loss, including:

[0070] Calculate the characteristic ratio of the trend component to the fluctuation component at the current time point;

[0071] A matching operation is performed in a preset allocation table based on the feature ratio to obtain the correction coefficient;

[0072] The first heat loss is weighted and calculated using the correction coefficient to obtain the second heat loss.

[0073] In one implementation, the characteristic ratio of the trend component to the fluctuation component at the current time point is calculated. It is worth noting that, to completely avoid the risk of system crash due to division-by-zero overflow caused by extremely stable weather conditions (i.e., the fluctuation component approaches zero), a fault-tolerant smoothing calculation mechanism is introduced. Specifically, the discrete value of the trend component at the current time point is extracted as the dividend; the discrete value of the fluctuation component at the same time point is extracted, and a preset smoothing minimum constant is added to this discrete value; the result of the sum is used as the divisor. An arithmetic division operation is performed by dividing the dividend by the divisor, and the result is determined as the characteristic ratio.

[0074] For example, a matching operation is performed on a preset allocation table based on the feature ratio to obtain the correction coefficient. The process of constructing the preset allocation table is as follows: a thermodynamic digital twin environment of the target building is built in an offline simulation computing node; a historical meteorological feature ratio sample dataset is input; the root mean square error between the simulated heat loss and the actual measured heat loss is minimized as a performance-driven constraint; a grid search algorithm is used to traverse and find the optimal correction coefficient value corresponding to each ratio interval; a one-dimensional key-value pair mapping relationship database is established and stored as the preset allocation table; the matching key-value node with the smallest Euclidean distance to the feature ratio at the current time is found in the preset allocation table; the constant value variable corresponding to the matching key-value node is extracted to obtain the correction coefficient.

[0075] In one implementation, the first heat loss is weighted and calculated using the correction coefficient to obtain the second heat loss. Specifically, the correction coefficient is used as a scalar multiplier and multiplied by the first heat loss calculated in the previous step to output the second heat loss, which incorporates corrections based on historical time-series memory and transient environmental attenuation characteristics.

[0076] In step S15, the measured temperature value is obtained, and an equilibrium equation is constructed based on the second heat loss and the measured temperature value. A numerical solution algorithm is then used to perform derivation calculations on the equilibrium equation to obtain the limiting temperature, including:

[0077] Obtain the measured temperature value;

[0078] Using the measured temperature value as the initial state and the second heat loss as the boundary condition, an equilibrium equation is constructed.

[0079] The evolutionary deduction calculation of the equilibrium equation is performed based on the numerical solution algorithm to obtain the change curve;

[0080] Extract the steady-state asymptote of the aforementioned change curve;

[0081] The steady-state asymptote is defined as the limiting temperature.

[0082] In one implementation, the measured temperature value is obtained. It should be noted that the analog electrical signal at the current moment is collected in real time by the temperature sensor module built into the indoor environment main control node. After performing an analog-to-digital conversion operation, a discrete value is output, and this discrete value is determined as the measured temperature value.

[0083] In one implementation, the measured temperature value is used as the initial state, and the second heat loss is used as the boundary condition to construct an equilibrium equation. It is worth noting that the equilibrium equation is constructed based on the property of a differential constant according to the first law of thermodynamics. Specifically, the construction process involves multiplying three physical parameters—preset air specific heat capacity, preset ambient air density, and preset room volume—to obtain the total indoor heat capacity constant. Then, a numerical equation is established, where the product of the first-order differential term of the temperature variable with respect to the time parameter and the total indoor heat capacity constant is equal to the arithmetic difference between the total power of the residual injected heat source of the current internal equipment and the second heat loss, which represents the dissipation boundary characteristic. The specific calculation formula for the equilibrium equation is as follows:

[0084]

[0085] in, This indicates the preset specific heat capacity of air. This represents the preset room temperature air density. This represents the preset room space volume. The product of these terms represents the total indoor heat capacity constant. This represents the temperature variable. This refers to the time parameter. This represents the first-order differential term of the temperature variable as a function of the time parameter; This indicates the total power of the residual injected heat source in the current internal equipment; This represents the second heat loss as a characteristic of the dissipation boundary.

[0086] Finally, the measured temperature value is substituted into the equation as the initial numerical solution at time zero, and the initial condition formula is as follows:

[0087]

[0088] in, This indicates at time zero (i.e. The initial temperature solution at time ( ). This represents the obtained measured temperature value. At this point, the complete construction of the equilibrium equation is finished.

[0089] For example, the equilibrium equation is subjected to evolutionary deduction calculations based on a numerical solution algorithm to obtain the change curve. It should be noted that the numerical solution algorithm is a fourth-order Runge-Kutta numerical approximation algorithm. Specifically, a preset deduction time step is obtained. Starting from the time zero point corresponding to the initial state, within each preset deduction time step interval, the slope values ​​of the differential equation at four different trial points are calculated. A weighted average calculation is performed on the slope values ​​of the differential equation at the four different trial points. The weighted average result is added to the temperature value at the current time node, and the estimated temperature value for the next time node is iteratively output. The preset deduction time step value is obtained by extracting the cutoff frequency of the building's thermal dynamic response and taking half of the reciprocal of that cutoff frequency. All iterative time nodes and the estimated temperature values ​​obtained from the deduction calculation are matrix-concatenated and recorded according to the time sequence dimension to generate the continuous change curve.

[0090] It is worth noting that this embodiment extracts the steady-state asymptote of the change curve. Specifically, it calculates the absolute value of the first-order backward difference of the estimated temperature values ​​between two adjacent iteration time nodes on the change curve. If, within a predetermined number of consecutive time nodes, the absolute value of the first-order backward difference is less than a predetermined convergence tolerance, then the estimated temperature value corresponding to the current last time node is extracted, and this estimated temperature value is determined as the steady-state asymptote.

[0091] It should be noted that the preset convergence tolerance value is determined based on the quantization noise floor error amplitude specified in the temperature sensor data specifications. The process of obtaining the preset quantity is as follows: obtain the maximum time overhead limit value reserved by the system processor for this computing thread, obtain the average execution time value of a single Runge-Kutta iteration, divide the maximum time overhead limit value by the average execution time value, perform an arithmetic division operation and round down, and determine the rounded result as the preset quantity. Through this explicit division logic, it is ensured that the system can find the most accurate steady-state temperature within the hard boundary of computing power resources. In this embodiment, the extracted steady-state asymptote is directly assigned a value, which is determined as the limit temperature that the room can maintain entirely by the system's waste heat under the current boundary conditions.

[0092] In step S16, the operating status is obtained. If the extreme temperature is less than the preset target temperature, the temperature difference between the preset target temperature and the extreme temperature is calculated. The switching logic is obtained by matching the temperature difference with the operating status. The preset future time window is segmented according to the switching logic and the preset time step to obtain the control interval. The control interval is allocated to obtain the control scheme.

[0093] Specifically, the process involves acquiring the operating status; if the extreme temperature is less than a preset target temperature, calculating the temperature difference between the preset target temperature and the extreme temperature; matching the temperature difference with the operating status to obtain switching logic; and performing a segmentation operation on a preset future time window based on the switching logic and a preset time step to obtain a control interval, including:

[0094] Get running status;

[0095] If the extreme temperature is less than the preset target temperature, then the temperature difference between the preset target temperature and the extreme temperature is calculated.

[0096] Extract the current power level from the operating status;

[0097] Extract rule entries that match the temperature difference value and the current power level from the preset strategy mapping table;

[0098] The rule entries are defined as switching logic;

[0099] The preset future time window is segmented according to the preset time step to obtain the basic interval;

[0100] The basic interval is merged according to the switching logic to obtain the control interval.

[0101] It should be noted that in this embodiment, the controller local area network bus interface inside the electric fireplace reads the digital frame containing the currently activated relay number from the underlying hardware feedback in real time, and parses the digital frame to generate the operating state. Then, the extreme temperature output in the previous step is compared with the preset target temperature. If the extreme temperature is not less than (i.e., greater than or equal to) the preset target temperature, it is determined that the current system waste heat has fully met the environmental heat demand. The system directly generates a time series instruction to maintain the zero-power standby state, and reuses the numerical solution algorithm to deduce the temperature decay trajectory under the zero-power state. The combination of the two is directly determined and output as the control scheme.

[0102] In one implementation, if the extreme temperature is less than a preset target temperature, the preset target temperature is used as the minuend, and the extreme temperature is used as the subtrahend; arithmetic subtraction is then performed to obtain the temperature difference. It is worth noting that a preset Celsius constant is received from an external input via the human-machine interface touchscreen panel on the surface of the electric fireplace device. This preset Celsius constant is written into memory and determined as the preset target temperature. This embodiment extracts the current power level in the operating state and retrieves rule entries matching the temperature difference with the current power level from a preset strategy mapping table.

[0103] It should be noted that the preset policy mapping table is generated offline based on a reinforcement learning Markov decision process in a discrete action space. The generation process involves constructing a reinforcement learning simulation environment, performing interval discretization on continuous temperature differences using a preset temperature segmentation step size, and using the discretized temperature difference intervals and the discrete power levels of the electric fireplace as joint state variables; using time window merging rules as action variables. A reward / penalty function is constructed, consisting of the algebraic sum of a temperature tracking reward term and an energy consumption penalty term, where the value of the temperature tracking reward term is negatively correlated with the absolute value of the steady-state temperature error, and the value of the energy consumption penalty term is negatively correlated with the value of the total operating power of the electric fireplace. Based on the above simulation environment and reward / penalty function, a value iteration algorithm is used to traverse the state space to find the optimal action output that maximizes the accumulated reward / penalty value for each state node, and the converged state-action full mapping relationship data is solidified and stored as the preset policy mapping table. Using the current temperature difference and the current power level as a joint index key, a hash lookup operation is performed in the preset strategy mapping table to extract the corresponding macroscopic time scheduling strategy instruction, i.e., the rule entry, and the rule entry is determined as the switching logic.

[0104] For example, a preset future time window is segmented according to a preset time step to obtain a basic interval. It is worth noting that the preset future time window is set based on the maximum time threshold of the instruction issuance cycle of the electric fireplace main control chip; the preset time step is extracted from the smallest timing unit of the system clock. In this embodiment, the total duration of the preset future time window is divided by the preset time step, uniformly dividing the time axis into multiple consecutive basic intervals. Subsequently, according to the time granularity merging instruction indicated in the switching logic, the beginning and end of the time dimension of adjacent basic intervals are spliced ​​and merged, outputting the merged control interval.

[0105] The control scheme is obtained by performing allocation calculations on the control interval, including:

[0106] Based on the equilibrium equation and the control interval, the predicted temperature distribution within the control interval is deduced.

[0107] If the temperature values ​​in the estimated temperature distribution are less than the preset target temperature, then calculate the amount of heat required to compensate for the temperature difference.

[0108] According to the injected heat generation control sequence;

[0109] An optimization operation is performed on the control sequence based on a preset sorting rule to obtain a control scheme.

[0110] In one implementation, based on the equilibrium equation and control interval established in the preceding steps, the predicted temperature distribution within the control interval is deduced. Specifically, using the predicted temperature value at the end of the previous time interval as the initial state, within the time span included in the control interval, the aforementioned numerical solution algorithm is reused to perform continuous evolution state calculations on the equilibrium equation. All calculated time nodes and their corresponding temperature values ​​are sequentially combined to construct the predicted temperature distribution representing the temperature change trend during that period.

[0111] In one implementation, the estimated temperature distribution is traversed. If the temperature value in the estimated temperature distribution is less than the preset target temperature, the heat injection required to compensate for the temperature difference is calculated. It should be noted that the temperature deviation value is obtained by subtracting the current temperature value from the preset target temperature. The injected heat is then calculated by multiplying the temperature deviation value, the aforementioned preset air specific heat capacity parameter, and the preset room mass parameter. It is worth noting that the preset room mass parameter is obtained by using a laser rangefinder to collect the three-dimensional outline physical dimensions of the room to calculate the internal space volume value, and then multiplying this internal space volume value by a preset room temperature air density constant.

[0112] For example, a control sequence is generated based on the injected heat. Specifically, in this embodiment, the injected heat is used as the divisor, and the fixed output heat value of each fixed physical power level of the electric fireplace within a single step is used as the divisor, and an integer programming modulo operation is performed. Through this operation, all power level combinations that can meet the total injected heat compensation requirement on the time axis are enumerated, and each enumerated power level combination is used as a candidate control sequence.

[0113] It is worth noting that the control sequence is optimized based on a preset sorting rule to obtain the control scheme. Specifically, the preset sorting rule is constructed by assigning the power level with the highest electrothermal conversion efficiency the smallest energy expenditure penalty weight, and the power level with the lowest electrothermal conversion efficiency the largest energy expenditure penalty weight, based on the factory-set electrothermal conversion efficiency calibration parameters of each heating component in the electric fireplace. According to the preset sorting rule, the penalty weights of each power level in each candidate control sequence are cumulatively calculated, and the total expenditure weight score of each control sequence is output. The candidate control sequence with the smallest total expenditure weight score is selected as the final control scheme, thereby optimizing the underlying physical energy consumption while ensuring accurate heat replenishment.

[0114] In step S17, the measured temperature difference is obtained, and an evaluation calculation is performed based on the control scheme and the measured temperature difference to obtain calibration data. The preset heat conduction parameters are then updated using the calibration data, including:

[0115] Obtain the measured temperature difference;

[0116] Extract the estimated temperature difference from the control scheme;

[0117] Calculate the absolute error between the estimated temperature difference and the measured temperature difference;

[0118] If the absolute error is greater than the preset error tolerance, the compensation parameters are calculated using the gradient descent algorithm.

[0119] The compensation parameters are determined as calibration data;

[0120] The preset thermal conductivity parameters are updated using the calibration data.

[0121] In one implementation, the measured temperature difference is obtained, and the estimated temperature difference in the control scheme is extracted. Specifically, the measured temperature difference is obtained by reading the actual ambient temperature values ​​at the end and beginning of the control interval using a temperature sensor and performing arithmetic subtraction on these two values. Simultaneously, the predicted terminal temperature value and the predicted initial temperature value corresponding to the derivation calculation node in the control scheme are analyzed, and arithmetic subtraction is performed to separate the estimated temperature difference. Finally, the estimated temperature difference is subtracted from the measured temperature difference, and the result of the subtraction is converted to an absolute value to obtain the absolute error.

[0122] In one implementation, if the absolute error exceeds a preset error tolerance, a compensation parameter is calculated using a gradient descent algorithm. It is worth noting that the preset error tolerance is determined by extracting the maximum quantization physical error amplitude of the analog-to-digital converter chip at the current resolution, and simultaneously extracting the static thermal noise standard deviation value given in the temperature sensor's manufacturer's specifications. The maximum quantization physical error amplitude is then added to the product of a constant 3 and the static thermal noise standard deviation value, and the result of this addition operation is determined as the preset error tolerance.

[0123] For example, the specific implementation steps for calculating the compensation parameter using the gradient descent algorithm are as follows: A loss model calculation logic is constructed with the square of the absolute error as the dependent variable, and a numerical difference algorithm is used to obtain the local gradient direction value. Specifically, a perturbation constant (e.g., a value of 0.001) is applied to the preset heat conduction parameter value in the current memory. The difference in absolute error before and after the perturbation is obtained through forward running system deduction. This difference in absolute error is divided by the perturbation constant, and arithmetic division is performed to calculate the local numerical gradient. In this embodiment, the local numerical gradient is multiplied by a preset learning step size coefficient to calculate the correction increment, and the negative of the correction increment is determined as the compensation parameter. It should be noted that the preset learning step size coefficient is set by calculating the extreme value of the Lipshitz constant of the loss model function, and taking one-tenth of the reciprocal of the extreme value of the Lipshitz constant as a fixed value, i.e., the preset learning step size coefficient. Finally, the calculated compensation parameter is directly assigned a value to determine the calibration data.

[0124] It is worth noting that the calibration data is used to perform a weighted update operation on the preset thermal conductivity parameters. Specifically, in this embodiment, the preset thermal conductivity parameters of the previous control cycle recorded in memory are used as the addend, and the calibration data is used as the addend. An algebraic superposition operation with the properties of arithmetic addition is performed, and the result of the superposition operation is written to the storage area of ​​the controller, replacing and serving as the reference preset thermal conductivity parameters for the next control cycle.

[0125] In summary, this invention accurately identifies and quantifies the dynamic comprehensive heat loss of a target building under complex environments by constructing a low-level perception link that incorporates multi-source meteorological feature fusion and Kalman filter state estimation. By introducing the STL time series decomposition algorithm and thermodynamic balance equations, it abandons traditional static empirical prediction and achieves high-fidelity forward-looking projection of the evolution trajectory of the room's residual heat limit temperature. Furthermore, by combining reinforcement learning policy mapping and integer programming optimization, a flexible power allocation scheme that balances time scheduling and precise heat injection is formulated. Finally, through an error feedback mechanism based on numerical differential gradients, the system's underlying physical heat conduction parameters are endowed with the ability to autonomously evolve with changes in the building environment. This invention completely breaks through the passive and lagging black-box temperature control limitations of traditional high-power heating equipment, achieving forward-looking and precise heating of electric fireplaces under nonlinear meteorological conditions. While maximizing the reduction of ineffective energy loss and avoiding frequent equipment start-ups and shutdowns and temperature overshoot, it greatly improves the dynamic stability of indoor temperature control and the ultimate thermal comfort experience for users.

[0126] Reference Figure 2 The second embodiment of the present invention provides a dynamic energy consumption optimization system for electric fireplaces, comprising:

[0127] The sequence processing module is used to acquire temperature data and weather data, and perform preprocessing to obtain the first sequence;

[0128] The state update module is used to perform state update processing on the first sequence, the preset transition matrix, and the preset observation matrix using the Kalman filter algorithm to obtain the second sequence;

[0129] The heat loss calculation module is used to obtain radiation values, perform mapping operations on the second sequence using preset heat conduction parameters to obtain conductive heat flow, perform compensation calculations on the conductive heat flow based on the radiation values ​​to obtain comprehensive heat flow, and perform product operations on the comprehensive heat flow in combination with preset surface area parameters to obtain the first heat loss;

[0130] The dynamic correction module is used to obtain historical heat loss sequences. If the first heat loss is greater than a preset heat loss threshold, the STL algorithm is used to decompose the historical heat loss sequence to obtain trend components and fluctuation components. A correction coefficient is determined based on the trend components and fluctuation components. The correction coefficient is used to perform a weighted operation on the first heat loss to obtain the second heat loss.

[0131] The evolution deduction module is used to obtain the measured temperature value, construct an equilibrium equation based on the second heat loss and the measured temperature value, and perform deduction calculations on the equilibrium equation using a numerical solution algorithm to obtain the limiting temperature.

[0132] The scheme generation module is used to obtain the operating status. If the extreme temperature is less than the preset target temperature, the module calculates the temperature difference between the preset target temperature and the extreme temperature. The module obtains the switching logic by matching the temperature difference with the operating status. The module performs a segmentation operation on the preset future time window according to the switching logic and the preset time step to obtain the control interval. The module performs an allocation calculation on the control interval to obtain the control scheme.

[0133] The feedback calibration module is used to acquire the measured temperature difference, perform evaluation calculations based on the control scheme and the measured temperature difference to obtain calibration data, and update the preset heat conduction parameters using the calibration data.

[0134] It should be noted that the electric fireplace energy consumption dynamic optimization system provided in this embodiment of the invention is used to execute all the process steps of the electric fireplace energy consumption dynamic optimization method in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0135] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0136] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for dynamic optimization of energy consumption in an electric fireplace, characterized in that, include: Temperature and weather data are acquired and preprocessed to obtain the first sequence; The first sequence, the preset transition matrix, and the preset observation matrix are processed using the Kalman filter algorithm to obtain the second sequence. The radiation value is obtained, and a mapping operation is performed on the second sequence using a preset heat conduction parameter to obtain the conduction heat flow. The conduction heat flow is compensated according to the radiation value to obtain the comprehensive heat flow. The comprehensive heat flow is multiplied by a preset surface area parameter to obtain the first heat loss. A historical heat loss sequence is obtained. If the first heat loss is greater than a preset heat loss threshold, the historical heat loss sequence is decomposed using the STL algorithm to obtain a trend component and a fluctuation component. A correction coefficient is determined based on the trend component and the fluctuation component. The first heat loss is weighted using the correction coefficient to obtain the second heat loss. Obtain the measured temperature value, construct an equilibrium equation based on the second heat loss and the measured temperature value, and perform derivation calculations on the equilibrium equation using a numerical solution algorithm to obtain the limiting temperature; The system obtains the operating status. If the extreme temperature is less than the preset target temperature, it calculates the temperature difference between the preset target temperature and the extreme temperature. It obtains the switching logic by matching the temperature difference with the operating status. It performs a segmentation operation on the preset future time window according to the switching logic and the preset time step to obtain the control interval. It performs an allocation calculation on the control interval to obtain the control scheme. Obtain the measured temperature difference, perform evaluation calculations based on the control scheme and the measured temperature difference to obtain calibration data, and update the preset heat conduction parameters using the calibration data.

2. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The acquisition of temperature and weather data, followed by preprocessing to obtain a first sequence, includes: Obtain temperature and weather data; The temperature data was subjected to anomaly removal using the 3-Sigma criterion to obtain the baseline temperature. The base temperature is smoothed using a moving average algorithm to obtain a smoothed temperature. Extract the timestamp tags from the weather data; Alignment is performed between the smoothed temperature and the weather data based on the timestamp label to obtain a first sequence.

3. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The second sequence is obtained by performing state update processing on the first sequence, the preset transition matrix, and the preset observation matrix using the Kalman filter algorithm, including: The predicted state is obtained by performing prior estimation calculation on the first sequence using a preset transition matrix; The Kalman gain is calculated by combining the preset observation matrix with the predicted state; The predicted state is updated using the Kalman gain to obtain the second sequence.

4. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The process of obtaining radiation values ​​involves performing a mapping operation on the second sequence using preset heat conduction parameters to obtain conductive heat flow, performing compensation calculations on the conductive heat flow based on the radiation values ​​to obtain a comprehensive heat flow, and performing a product operation on the comprehensive heat flow using preset surface area parameters to obtain a first heat loss, including: Obtain radiation values; Extract the indoor and outdoor temperature values ​​from the second sequence; Calculate the characteristic difference between the indoor temperature value and the outdoor temperature value; The difference features are weighted and mapped using preset heat conduction parameters to obtain the conductive heat flow; The combined heat flux is obtained by performing a linear superposition calculation on the conducted heat flux using the radiation values. The first heat loss is obtained by multiplying the combined heat flow with the preset surface area parameters.

5. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The step of determining a correction coefficient based on the trend component and the fluctuation component, and then applying the correction coefficient to perform a weighted calculation on the first heat loss to obtain the second heat loss includes: Calculate the characteristic ratio of the trend component to the fluctuation component at the current time point; A matching operation is performed in a preset allocation table based on the feature ratio to obtain the correction coefficient; The first heat loss is weighted and calculated using the correction coefficient to obtain the second heat loss.

6. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The process of obtaining the measured temperature value, constructing an equilibrium equation based on the second heat loss and the measured temperature value, and performing derivation calculations on the equilibrium equation using a numerical solution algorithm to obtain the limiting temperature includes: Obtain the measured temperature value; Using the measured temperature value as the initial state and the second heat loss as the boundary condition, an equilibrium equation is constructed. The evolutionary deduction calculation of the equilibrium equation is performed based on the numerical solution algorithm to obtain the change curve; Extract the steady-state asymptote of the aforementioned change curve; The steady-state asymptote is defined as the limiting temperature.

7. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The process of acquiring the operating status involves, if the extreme temperature is less than the preset target temperature, calculating the temperature difference between the preset target temperature and the extreme temperature. A switching logic is then obtained by matching the temperature difference with the operating status. Based on the switching logic and a preset time step, a preset future time window is segmented to obtain a control interval, including: Get running status; If the extreme temperature is less than the preset target temperature, then the temperature difference between the preset target temperature and the extreme temperature is calculated. Extract the current power level from the operating status; Extract rule entries that match the temperature difference value and the current power level from the preset strategy mapping table; The rule entries are defined as switching logic; The preset future time window is segmented according to the preset time step to obtain the basic interval; The basic interval is merged according to the switching logic to obtain the control interval.

8. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The process of performing allocation calculations on the control interval to obtain a control scheme includes: Based on the equilibrium equation and the control interval, the predicted temperature distribution within the control interval is deduced. If the temperature values ​​in the estimated temperature distribution are less than the preset target temperature, then calculate the amount of heat required to compensate for the temperature difference. According to the injected heat generation control sequence; An optimization operation is performed on the control sequence based on a preset sorting rule to obtain a control scheme.

9. The method for dynamic optimization of energy consumption in an electric fireplace according to claim 1, characterized in that, The process of obtaining the measured temperature difference, performing evaluation calculations based on the control scheme and the measured temperature difference to obtain calibration data, and updating the preset heat conduction parameters using the calibration data includes: Obtain the measured temperature difference; Extract the estimated temperature difference from the control scheme; Calculate the absolute error between the estimated temperature difference and the measured temperature difference; If the absolute error is greater than the preset error tolerance, the compensation parameters are calculated using the gradient descent algorithm. The compensation parameters are determined as calibration data; The preset thermal conductivity parameters are then updated using the calibration data with a weighted average.

10. A dynamic energy consumption optimization system for an electric fireplace, characterized in that, include: The sequence processing module is used to acquire temperature data and weather data, and perform preprocessing to obtain the first sequence; The state update module is used to perform state update processing on the first sequence, the preset transition matrix, and the preset observation matrix using the Kalman filter algorithm to obtain the second sequence; The heat loss calculation module is used to obtain radiation values, perform mapping operations on the second sequence using preset heat conduction parameters to obtain conductive heat flow, perform compensation calculations on the conductive heat flow based on the radiation values ​​to obtain comprehensive heat flow, and perform product operations on the comprehensive heat flow in combination with preset surface area parameters to obtain the first heat loss; The dynamic correction module is used to obtain historical heat loss sequences. If the first heat loss is greater than a preset heat loss threshold, the STL algorithm is used to decompose the historical heat loss sequence to obtain trend components and fluctuation components. A correction coefficient is determined based on the trend components and fluctuation components. The correction coefficient is used to perform a weighted operation on the first heat loss to obtain the second heat loss. The evolution deduction module is used to obtain the measured temperature value, construct an equilibrium equation based on the second heat loss and the measured temperature value, and perform deduction calculations on the equilibrium equation using a numerical solution algorithm to obtain the limiting temperature. The scheme generation module is used to obtain the operating status. If the extreme temperature is less than the preset target temperature, the module calculates the temperature difference between the preset target temperature and the extreme temperature. The module obtains the switching logic by matching the temperature difference with the operating status. The module performs a segmentation operation on the preset future time window according to the switching logic and the preset time step to obtain the control interval. The module performs an allocation calculation on the control interval to obtain the control scheme. The feedback calibration module is used to acquire the measured temperature difference, perform evaluation calculations based on the control scheme and the measured temperature difference to obtain calibration data, and update the preset heat conduction parameters using the calibration data.

Images