Zero-carbon energy storage building comprehensive illumination energy load optimization scheduling method

Abstract

The invention discloses a zero-carbon energy storage building comprehensive illumination energy load optimization scheduling method, and relates to the technical field of power grid coupling scheduling, and the method comprises the steps: recognizing a reflected light path through a light path random sampling simulation method based on a building photo-thermal environment feature matrix, and constructing a building surface light intensity distribution thermodynamic diagram; according to the building surface light intensity distribution thermodynamic diagram, the deflection angle of the building reflector array is dynamically adjusted, and a building light guide performance parameter set is generated through a dynamic light field efficiency quantification method; according to the building light guide performance parameter set, a hybrid MPPT algorithm is adopted to obtain a photovoltaic operation state data set; building energy hierarchical scheduling is carried out based on the photovoltaic operation state data set, and an energy scheduling instruction set and a cross-domain energy scheduling collaborative log are obtained; according to the method, the photovoltaic operation state data set and the power grid hierarchical framework are fused through the building energy hierarchical scheduling model, cross-space-time collaborative optimization of the illumination energy and the power grid load is realized, and the target of zero-carbon building light-storage-network full-link efficient scheduling is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid coupling scheduling, and in particular to a method for optimizing the scheduling of integrated lighting energy loads in zero-carbon energy storage buildings. Background Art

[0002] The integrated lighting energy load optimization scheduling technology for zero-carbon energy storage buildings occupies a vital position in today's zero-carbon building energy scheduling field. Traditional methods generally adopt a phased optimization strategy, collecting ambient light and heat data through fixed-parameter irradiance sensors, and combining statistical methods to predict photovoltaic output; then, a static scheduling strategy is generated based on the load history database, and a hierarchical control architecture is used to realize grid power distribution.

[0003] In the field of optimized scheduling of integrated lighting energy loads for zero-carbon energy storage buildings, traditional light path analysis often uses static BRDF models and single-shot ray tracing, resulting in significant lag in identifying areas of abnormal light intensity. When the solar altitude angle changes rapidly, it is difficult to update the reflection path in real time, resulting in large deviations in the calculation of the effective light-guiding area of ​​the photovoltaic panel. Furthermore, there is a lack of dynamic coupling mechanisms between the photovoltaic operating status and the grid scheduling level, especially in terms of the coordination of light-guiding performance parameters and confidence interval constraints. Traditional technologies require independent processing of light-to-thermal conversion efficiency and power allocation, and are unable to adjust transmission loss compensation strategies in real time based on thermal maps of light intensity distribution on the building surface. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method to solve the problems of delayed light path response of building envelope structures and insufficient coordination of energy levels.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, the present invention provides a method for optimizing and scheduling the comprehensive lighting energy load of a zero-carbon energy storage building, which includes collecting multi-dimensional light and thermal environment monitoring data, fusing them through principal component analysis dimensionality reduction and K-means clustering, and generating a building light and thermal environment feature matrix; based on the building light and thermal environment feature matrix, identifying the reflected light path through a light path random sampling simulation method, and constructing a heat map of the light intensity distribution on the building surface; dynamically adjusting the deflection angle of the building reflector array according to the light intensity distribution heat map on the building surface, and generating a building light guiding performance parameter set through a dynamic light field efficiency quantification method; based on the building light guiding performance parameter set, using a hybrid MPPT algorithm to obtain a photovoltaic operation status data set; based on the photovoltaic operation status data set, performing building energy hierarchical scheduling, obtaining an energy scheduling instruction set and a cross-domain energy scheduling collaborative log; and regularly updating the energy scheduling instruction set according to the collaborative operation log and historical environmental data.

[0007] As a preferred solution of the zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method of the present invention, wherein: the principal component analysis dimensionality reduction and K-means clustering are integrated to generate the building light and thermal environment feature matrix, the steps are as follows: The multi-dimensional photothermal environment monitoring data is Z-score standardized, the covariance matrix is ​​calculated, and the feature decomposition is performed through component analysis and dimensionality reduction to extract the photothermal coupling characteristics; Based on the light-thermal coupling characteristics, the K-means clustering algorithm is used to initialize the cluster centers, and the building light-thermal environment characteristic matrix is ​​generated through Euclidean distance measurement and iterative optimization.

[0008] As a preferred solution of the zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method of the present invention, wherein: the reflected light path is identified by the light path random sampling simulation method to construct the building surface light intensity distribution heat map, the steps are as follows: Based on the building light and thermal environment characteristic matrix and surface material property library, a level reflection characteristic mapping table is generated through dynamic parameter mapping rules; Based on the reflection characteristic mapping table, the Monte Carlo random sampling method is used to simulate the light incident angle distribution, and the BRDF response surface is dynamically generated in combination with the material property library; The reflection direction and energy attenuation are calculated based on the BRDF response surface, the hitting coordinates and residual energy of the light on the photovoltaic panel surface are recorded, and the reflection path is identified through path clustering analysis. At the same time, the residual energy value is accumulated and normalized to light intensity. The spatial discretization grid statistics and Gaussian filter reconstruction technology are used to generate a heat map of the light intensity distribution on the building surface.

[0009] As a preferred solution of the zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method of the present invention, wherein: the deflection angle of the building reflector array is dynamically adjusted according to the light intensity distribution heat map of the building surface, the steps are as follows: According to the heat map of light intensity distribution on the building surface, the abnormal area is located through the three-dimensional coordinate transformation matrix, and the effective light guiding area is obtained based on the collective difference between the total area and the abnormal area.

[0010] Based on the distribution of abnormal areas and combined with the initial installation parameters of the reflector array, the reflector deflection angle is obtained, and the gradient descent method is used to optimize the reflector deflection angle.

[0011] As a preferred solution of the zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method of the present invention, wherein: the building light guide performance parameter set is generated by the dynamic light field efficiency quantification method, the steps are as follows: The optimized reflector deflection angle is converted into a pulse signal through a stepper motor pulse equivalent, and the reflector array is driven to adjust the deflection angle of the building reflector array; The deflection angle error of the actual building reflector array is monitored synchronously through a photoelectric encoder, and the number of pulses is dynamically compensated using a PID algorithm to adjust the deflection angle of the building reflector array again; A high-speed spectrometer is used to collect the adjusted light intensity distribution data, and the photothermal conversion efficiency is calculated through the dynamic light field efficiency quantification method, and a set of building light guidance performance parameters is generated.

[0012] As a preferred solution of the zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method of the present invention, wherein: the hybrid MPPT algorithm is used to obtain the photovoltaic operation status data set, the steps are as follows: Dynamically assign weights to the perturbation-observation method and the conductance increment method, extract the future illumination trend time series characteristics through sliding window time series analysis, and construct a hybrid MPPT algorithm. The building light guiding performance parameter set is combined with the series-parallel topology of the photovoltaic array to generate the initial VI reference curve through least square fitting; Execute the hybrid MPPT algorithm, trigger mode switching based on the light intensity gradient threshold, output candidate optimal power point parameters through confidence voting, and calculate the deviation by comparing the initial VI reference curve in real time. At the same time, calibrate through Bayesian parameter inference to generate the optimal power point parameters; Based on the optimal power point parameters and the building light guiding performance parameter set, combined with the physical size of the photovoltaic array, the effective grid is marked, the effective grid is numerically integrated, and the local photoelectric conversion efficiency is obtained through integral calculation; The optimal power point parameters, local photoelectric conversion efficiency, uniformity index and effective light guiding area ratio output by the hybrid MPPT algorithm are encapsulated into a key-value pair mapping table through structured data to form a photovoltaic operation status data set.

[0013] As a preferred solution of the zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method of the present invention, wherein: the building energy hierarchical scheduling is performed based on the photovoltaic operation status data set, the steps are as follows: Combined with the photovoltaic operating status dataset, outliers are removed through Kalman filtering to generate a spatiotemporal calibration dataset; Based on the physical node topology of the distribution network, the power generation layer, transmission layer, distribution layer, and user layer are explicitly divided, and the grid node connection relationship with hierarchical labels is generated. Combined with the photovoltaic operation status dataset, a building energy hierarchical scheduling model is created; The spatiotemporal calibration dataset is input into the building energy hierarchical scheduling model, and the allocation algorithm is applied at each level to generate the hierarchical electricity allocation results.

[0014] As a preferred solution of the zero-carbon energy storage building comprehensive lighting energy load optimization scheduling method of the present invention, wherein: the steps of obtaining the energy scheduling instruction set and the cross-domain energy scheduling collaborative log are as follows: The hierarchical power distribution results are encapsulated into structured parameters, and an energy scheduling instruction set is constructed through the instruction parameter compilation engine to optimize the scheduling of lighting energy loads. Real-time monitoring of actual power transmission and solar energy load optimization scheduling delay and deviation. If the deviation is greater than the delay limit, the energy scheduling instruction set is triggered to retry and record an alarm, and a cross-domain energy scheduling collaboration log is generated simultaneously.

[0015] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the method for optimizing and scheduling the comprehensive lighting energy load of a zero-carbon energy storage building as described in the first aspect of the present invention is implemented.

[0016] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the method for optimizing and scheduling the integrated lighting energy load of a zero-carbon energy storage building as described in the first aspect of the present invention is implemented.

[0017] The beneficial effects of the present invention are as follows: through the heat map of light intensity distribution on the building surface, the energy distribution and path characteristics of indirect light on the surface of the photovoltaic panel are accurately quantified, achieving millimeter-level spatial accuracy in identifying local hotspots and shadow areas; combining the building energy hierarchical scheduling model to integrate the photovoltaic operation status data set and the grid hierarchy framework, an energy scheduling instruction set is generated to drive energy storage priority call, dynamic compensation for line losses and cross-domain power bidding, realizing the cross-temporal and spatial coordinated optimization of lighting energy and grid load, and achieving the goal of efficient scheduling of the entire link of zero-carbon building light-storage-grid. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0019] Figure 1 Flowchart of the optimized scheduling method for integrated lighting energy load of zero-carbon energy storage buildings.

[0020] Figure 2 Flowchart for hybrid MPPT algorithm implementation.

[0021] Figure 3 This is the flow chart of the optical path random sampling simulation method.

[0022] Figure 4 Flowchart of the building energy hierarchical scheduling model. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0026] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides a method for optimizing and scheduling the integrated lighting energy load of a zero-carbon energy storage building, comprising the following steps: S1. Collect multi-dimensional light and thermal environment monitoring data, fuse them through principal component analysis dimensionality reduction and K-means clustering, and generate the building light and thermal environment feature matrix; Multi-dimensional light and thermal environment monitoring data includes spectral irradiance, surface temperature, humidity and solar altitude angle; Noise filtering, missing value filling and time-space alignment are performed on multi-dimensional light and thermal environment monitoring data.

[0027] Furthermore, precise spectral irradiance measurement relies on a combination of a Hamamatsu C12880MA miniature spectrometer and an InGaAs detector. Visible light is sampled via a fiber-optic probe, while near-infrared wavelengths are captured by a cooled InGaAs module. Polarization angle scanning is achieved using a built-in motorized filter wheel and a Thorlabs WP25M-UB polarizer. NIST-traceable calibration using an Ocean Optics LS-1-CAL standard halogen lamp ensures sub-nanometer wavelength accuracy across the entire wavelength range. Dynamic range extension technology combined with an automatic gain algorithm can handle variations from sunny midday to rainy weather. Surface temperature is detected using a FLIR T865 uncooled infrared thermal imager. Humidity monitoring is achieved using a multi-point deployment of Honeywell HIH-4000 series capacitive sensors coupled with Pt100 platinum resistance temperature probes for temperature and humidity coupling. Correction of the solar altitude angle is based on an AS5048B photoelectric encoder and a four-quadrant diode array. The multi-dimensional photothermal environment monitoring data were Z-score standardized to eliminate dimensional differences, and the covariance matrix was calculated to reveal the correlation between parameters.

[0028] Calculate the covariance matrix , the expression is: ; in, The number of samples for light and heat environment monitoring data; is the centralized data matrix, is the transpose of the centered data matrix.

[0029] The covariance matrix is ​​decomposed by component analysis and dimensionality reduction to extract the light-thermal coupling characteristics; Furthermore, a dimensionality reduction method is used to analyze the input components of the eigenvalue sequence and eigenvector matrix, and a dimensionality reduction algorithm is applied to generate a reduced-dimensional feature matrix. Based on the light-thermal coupling feature screening rules, the interception position index is determined. The corresponding column vectors are intercepted from the 1st to the kth columns of the reduced-dimensional feature matrix. These k intercepted column vectors are concatenated in column order, and the light-thermal coupling features are generated by feature encapsulation of the coefficients of the first k columns in the eigenvector matrix.

[0030] It should be noted that the eigenvalue sequence refers to the numerical sequence formed by arranging the eigenvalues ​​obtained after the eigendecomposition of the covariance matrix in descending order, which represents the variance contribution ranking of the original light and thermal environment monitoring parameters in the orthogonal transformation direction. The eigenvector matrix is ​​a projection matrix composed of orthogonal unit vectors generated by the eigendecomposition of the covariance matrix, and each column represents a principal component direction.

[0031] Based on the light-thermal coupling characteristics, the K-means clustering algorithm is used to initialize the cluster center, and the building light-thermal environment characteristic matrix is ​​generated through Euclidean distance measurement and iterative optimization. Furthermore, the first center point is randomly selected from the building's thermal environment data. The next center point is then chosen according to a probability distribution based on the square of the minimum Euclidean distance from each building monitoring point to the selected center point, ensuring that the initial center point reflects the characteristics of a typical building environment zone. Entering the iterative optimization phase, the Euclidean distance from all building monitoring points to each center point is calculated, and each building environment data point is assigned to the nearest cluster center to form the building thermal environment zone. The cluster center is updated to the mean vector of the building monitoring points within the zone, and the algorithm is iterated until the maximum number of times (e.g., 300) converges. Upon completion, the coordinates of the building thermal zone centers are arranged in rows to form a building zone center matrix. Simultaneously, the thermal coupling feature vector of each building monitoring sample is combined with the zone label to form a building feature-zone association row vector. Finally, the feature-zone matrix and the zone center matrix are concatenated column by column to generate the building thermal environment feature matrix.

[0032] S2. Based on the building's light and thermal environment characteristic matrix, the reflected light path is identified through the light path random sampling simulation method to construct a heat map of the building's surface light intensity distribution; Based on the building light and thermal environment characteristic matrix and surface material property library, a reflection characteristic mapping table is generated through dynamic parameter mapping rules; It should be noted that the surface material property library is a structured database that stores the optical property parameters of the building surface, including physical property fields such as reflectivity, refractive index and anisotropic roughness, and associates the spatial coordinates of the photovoltaic panel coverage area through the material type index.

[0033] Furthermore, the building light and thermal environment characteristic matrix and the surface material property library are input into the dynamic parameter mapping rules. By parsing the solar altitude angle and spectral irradiance parameters in the building light and thermal environment characteristic matrix, the material type of the surface material property library is associated (for example, the photovoltaic panel area corresponds to polysilicon material), and the material reflectivity baseline value is called (for example, the polysilicon reflectivity is 0.25) and corrected in real time based on the environmental parameters: for example, when the solar altitude angle is >50°, the reflectivity is ×1.2 (for example, 0.25→0.30), and when the surface temperature is >40℃, the roughness is increased by 0.1μm (for example, 1.2μm→1.3μm), and finally a level reflection characteristic mapping table is generated.

[0034] Based on the reflection characteristic mapping table, the Monte Carlo random sampling method is used to simulate the light incident angle distribution, and the BRDF response surface is dynamically generated in combination with the material property library; Furthermore, the reflection characteristic mapping table is input into the Monte Carlo random sampling method. By setting the sampling points to simulate the distribution of light incidence angles (for example, uniform sampling of the zenith angle 0°-90°), and at the same time calling the material parameters in the material property library (for example, the glass refractive index is 1.5), the bidirectional reflectivity distribution function value corresponding to each incident angle is calculated in real time, and finally the BRDF response surface is dynamically generated.

[0035] Use GPU acceleration to launch millions of random rays. For each ray, the reflection direction and energy attenuation are calculated based on the BRDF response surface, and the hit coordinates and remaining energy of the ray on the photovoltaic panel surface are recorded. Furthermore, the GPU parallel computing architecture (such as NVIDIA CUDA core) is used to synchronously emit millions of random light rays. Each ray starts from a set light source position (such as a solar azimuth angle of 30°). During the ray tracing process, the reflection behavior is calculated in real time based on the dynamic bidirectional reflectance distribution function (BRDF) response surface of the photovoltaic panel surface. According to the incident angle, wavelength (such as the 400-1100nm band) and surface material properties (such as the single-crystal silicon roughness coefficient of 0.2), the reflection direction vector and energy attenuation rate are obtained through the BRDF model (such as the Cook-Torrance model). After the light interacts with the photovoltaic panel surface, the hit coordinates and residual energy values ​​are recorded. Finally, a point cloud dataset containing tens of millions of hit point coordinates and residual energy is output for constructing a heat map of the light intensity distribution on the building surface and analyzing the energy absorption efficiency.

[0036] Based on the reflection direction and hit coordinates, the reflection paths are identified through path cluster analysis; Furthermore, the reflection direction and hit coordinates of each ray are combined into a five-dimensional feature vector and input into the DBSCAN density clustering method. The Euclidean distance between the feature vectors is used to automatically classify rays with adjacent spatial positions and similar reflection directions. If there are at least 50 rays around a ray (for example, within a radius of 0.1 radians), the rays are aggregated into a reflection path cluster, and finally a reflection path set with cluster labels is output.

[0037] According to the grid-hitting light data, the remaining energy value is accumulated and normalized to light intensity. Through spatial discretization grid statistics and Gaussian filter reconstruction technology, a heat map of light intensity distribution on the building surface is generated; Furthermore, the grid-hitting light data (a set of light-hitting records divided into a preset grid size, such as a 20cm×20cm grid) is input into the energy accumulation process, and the residual energy values ​​of all the light rays in each grid unit are summed to obtain the grid cumulative energy. The cumulative energy is divided by the grid area and the total number of light rays, and multiplied by the unit time conversion coefficient (the inverse of the total duration of the light emission simulation) to generate a normalized light intensity value. A matrix is ​​constructed based on the light intensity values ​​of all grids, and the energy of the light-hitting points is discretized according to the grid. The light intensity of each grid is normalized and calculated. After smoothing the noise with Gaussian filtering, a heat map of the light intensity distribution on the building surface is generated through pseudo-color mapping.

[0038] It should be noted that the photovoltaic panel is divided into a two-dimensional grid based on the three-dimensional spatial coordinates of the photovoltaic panel surface and the preset grid size, and the difference between the minimum coordinate value and the maximum coordinate value in the X-axis direction of the grid is extracted as the length of the photovoltaic panel, and the difference between the minimum coordinate value and the maximum coordinate value in the Y-axis direction is extracted as the width of the photovoltaic panel, and the area of ​​a single grid is obtained based on the total number of grids.

[0039] S3. Dynamically adjust the deflection angle of the building reflector array based on the light intensity distribution heat map of the building surface, and generate a building light guide performance parameter set through a dynamic light field efficiency quantification method; According to the light intensity distribution heat map of the building surface, the abnormal area is located through the three-dimensional coordinate transformation matrix, and the effective light guide area is obtained based on the collective difference between the total area and the abnormal area; It should be noted that the safety threshold is the upper limit of light intensity set inside the photovoltaic module (for example, >1500 W / m²). Exceeding this value may cause thermal damage to the module or safety risks; abnormal areas include areas with excessive light intensity and shadow areas. The areas with excessive light intensity are areas extracted based on the thermal map of light intensity distribution on the building surface, and meet the conditions that the light intensity value exceeds the safety threshold and is accompanied by abnormal local temperature rise (for example, temperature difference >5°C); the minimum operating threshold is the lowest light intensity limit at which the photovoltaic module can maintain its power generation function (for example, <100 W / m²). When the light intensity is lower than this value and the temperature is lower than the adjacent area, the area is identified as a shadow area based on the three-dimensional spatial coordinate positioning.

[0040] Based on the distribution of abnormal areas and combined with the initial installation parameters of the reflector array, the reflector deflection angle is obtained, and the gradient descent method is used to optimize the reflector deflection angle.

[0041] Furthermore, starting with the initial angle, fine-tuning is performed step by step to determine the rate of change in the total area of ​​the abnormal region. Combined with the maximum temperature rise change and the reduction in the deviation between the actual output power and the ideal value, the optimization weights of the three indicators are comprehensively evaluated. The angle is then gradually adjusted in the direction of overall improvement until the adjustment effect stabilizes, ultimately outputting the optimized angle. For example, starting with the initial reflector installation angle of the photovoltaic curtain wall system on the south facade of a building, the deflection angle is gradually fine-tuned using a gradient descent method. After each angle adjustment, three key indicators are monitored: the rate of change in the total area of ​​the abnormal region on the building's daylighting surface reflects the improvement in light uniformity; the maximum temperature rise change of the photovoltaic curtain wall unit indicates the effectiveness of thermal stress relief; and the reduction in the deviation between the actual output power and the theoretical maximum value reflects the improvement in power generation efficiency. The algorithm comprehensively evaluates the weights of each indicator and iteratively adjusts the angle in the direction of optimal overall benefit. The learning rate is set to the typical engineering optimization step size. The optimization terminates when the rate of change in the abnormal region falls below a set threshold or the total number of iterations reaches an upper limit over multiple consecutive iterations, and the optimal deflection angle for the building's reflective system is output.

[0042] The optimized reflector deflection angle is converted into a pulse signal through the stepper motor pulse equivalent, and the reflector array is driven to adjust the deflection angle of the building reflector array. The actual building reflector array deflection angle error is monitored synchronously through the photoelectric encoder, and the PID algorithm is used to dynamically compensate the number of pulses to adjust the deflection angle of the building reflector array again.

[0043] Furthermore, the optimized reflector deflection angle is divided by the stepper motor pulse equivalent, the initial driving pulse number is calculated, and the stepper motor is driven to rotate the reflector array to initially adjust the deflection angle of the building reflector array; the actual building reflector array deflection angle is synchronously monitored in real time by a photoelectric encoder, and the error between the deflection angle and the target angle is calculated. Based on the error value, a PID algorithm is used to dynamically generate compensation pulses, and the actual building reflector array deflection angle error monitored by the photoelectric encoder is input into the PID controller. First, the proportional link (the linear amplification module of the current real-time error in the control system) is multiplied by the gain coefficient (the scaling factor in the proportional link) to generate an immediate correction component that changes synchronously with the current error; secondly, the historical error is accumulated over time through the integral link to generate a continuous correction component that eliminates the steady-state error; finally, the instantaneous rate of change of the error is captured through the differential link to generate a predictive correction component that suppresses overshoot; the output components of the three links are superimposed to generate a comprehensive compensation angle; the compensation angle is divided by the stepper motor pulse equivalent and converted into the number of compensation pulses, which drives the stepper motor to perform reverse rotation to dynamically offset the angle error.

[0044] A high-speed spectrometer is used to collect the adjusted light intensity distribution data, and the average light intensity, uniformity index, single grid area, proportion of effective light guiding area and spectral energy distribution are calculated. The photothermal conversion efficiency is calculated through the dynamic light field efficiency quantification method, and a set of building light guiding performance parameters is generated.

[0045] Furthermore, the product of the light intensity value of the grid in the effective light-guiding area and the area of ​​the single grid is accumulated, multiplied by the proportion of the effective light-guiding area and divided by the total number of grids; secondly, the spectral matching ablation coefficient is obtained: the measured spectral energy distribution is compared with the ideal spectrum of the photovoltaic module (such as the AM1.5 standard), combined with the module spectral response value (such as the single crystal silicon response peak of 900nm), and the energy matching ratio of the actual spectrum and the module response is integrated to obtain; then, the initial photothermal conversion efficiency is generated by combining the light absorption rate of the photovoltaic module (fixed value such as 0.92) and the illumination duration and the total incident light power; finally, the uniformity index correction loss is calculated based on the average value and standard deviation of the light intensity, and the energy loss caused by the uneven distribution is compensated with an empirical coefficient (such as 0.15). The initial efficiency value is corrected to obtain the final photothermal conversion efficiency, and a set of building light-guiding performance parameters including average light intensity, uniformity index and other parameters are simultaneously output.

[0046] S4. Based on the building light guiding performance parameter set, a hybrid MPPT algorithm is used to obtain a photovoltaic operation status data set; Based on the average light intensity and spectral energy distribution in the light guiding parameter set, the weights of the perturbation observation method and the conductance increment method are dynamically allocated. The time series characteristics of future lighting trends are extracted through sliding window time series analysis, and a hybrid MPPT algorithm with multiple parallel algorithms and adjustable weights is constructed. Furthermore, based on the real-time average light intensity data sequence in the light guiding parameter set (for example, the light intensity value of 10 consecutive points under a 300ms sampling period) and the synchronously collected spectral energy distribution (such as the energy value of each wavelength in the 400-1100nm band), the following time series features are extracted through sliding window timing analysis with a specific step size (such as 300ms) and window width (such as 10 sampling points): calculation of the light intensity change rate within the window (such as an upward gradient of 1.2% / s), detection of light intensity mutation events (such as transients marked by the difference between adjacent points >50W / m²), and quantification of spectral shift characteristics (such as a blue shift rate of 0.5% / s in the 500-600nm band); analysis of the building light guiding performance parameter set to obtain Based on historical light intensity and corresponding light event labels, the Adam optimizer minimizes the weighted cross-entropy loss to complete the supervised training of a two-layer LSTM time series model using spectral time series data. The time series feature vector is input into the pre-trained two-layer LSTM time series model, which outputs a probability distribution of light trends for the next 10 seconds. Weights are dynamically assigned based on the prediction results. If the probability of a sudden drop is high (e.g., >80%), the weight of the perturbation-observation method is increased (e.g., 85% to enhance mutation response capability) and the weight of the conductance increment method is decreased (e.g., 15% to suppress steady-state oscillation risk). This creates a hybrid architecture in which the two algorithms are executed in parallel according to weight coefficients to drive the pulse width modulation signal of the photovoltaic inverter to achieve dynamic operating condition adaptation.

[0047] The light intensity distribution data, uniformity index and effective light guiding area coordinates in the building light guiding performance parameter set are combined with the series-parallel topology of the photovoltaic array to generate an initial VI reference curve through least squares fitting. This curve serves as the search starting point and deviation calibration benchmark for the hybrid MPPT algorithm. It should be noted that the photovoltaic array series-parallel topology refers to the circuit structure formed by photovoltaic cell modules in the array through a specific electrical connection method. Multiple photovoltaic cell modules are connected in series end to end to form a component string with superimposed voltages. Several photovoltaic cell modules are then connected in series to a unified busbar to expand the output current (for example, 5 series in parallel increase the maximum current from 10A to 50A), forming a complete array circuit network composed of a series-parallel hierarchical relationship.

[0048] Execute the hybrid MPPT algorithm, trigger mode switching based on the light intensity gradient threshold, output candidate optimal power point parameters through confidence voting, and calculate the deviation by comparing the initial VI reference curve in real time. At the same time, calibrate through Bayesian parameter inference to generate the optimal power point parameters; It should be noted that the light intensity gradient threshold is defined by the core criterion of responding to sudden changes in illumination (for example, a significant change is determined when the absolute value of the gradient is >1.0% / s). Its real-time calculation method is the sliding average of the light intensity change rate of adjacent sampling points in the sliding window. The value range is the natural fluctuation boundary (lower limit ≥0.8% / s), the photovoltaic cell module safety boundary (upper limit ≤1.4% / s), and the hardware performance boundary (limit value ≤50% / s). When the real-time gradient exceeds the threshold, the mode switch is triggered immediately. In the stable mode (such as gradient ≤1.0% / s), the conductance increment method (INC) is maintained as the dominant method, focusing on steady-state accuracy. In the transient mode (such as gradient >1.0% / s), the perturbation observation method (PO) is switched as the dominant method, giving priority to tracking sudden changes.

[0049] Furthermore, the real-time deviation value (e.g., voltage deviation -30V) between the candidate optimal power point parameters output by the hybrid MPPT algorithm and the initial VI reference curve is input into the Bayesian inference model. Combined with the prior normal distribution of the photovoltaic cell module temperature, the aging attenuation index, and the uniform distribution of irradiance, the duty cycle calibration offset (e.g., -0.8%) corresponding to the maximum posterior probability is solved through the Monte Carlo Markov chain (MCMC) sampling parameter combination to generate the calibrated optimal power point parameters. Finally, the output drive signal makes the output voltage error converge from the real-time deviation value to a reasonable range (e.g., -1.5V).

[0050] Based on the optimal power point parameters, light intensity distribution data in the building light guiding performance parameter set, single grid area and effective light guiding area coordinates, combined with the physical size of the photovoltaic array, the effective grid is marked, the effective grid is numerically integrated, and the local photoelectric conversion efficiency is obtained through integral calculation; It should be noted that the physical dimensions of a photovoltaic array refer to the actual geometric dimensions of the photovoltaic panels in three-dimensional space (such as length, width, and mounting angle). These dimensions are obtained through measurement or design drawings and are used to locate the boundary reference of the grid coordinate system. An effective grid refers to a grid cell within the physical dimensions that can generate electricity normally, dynamically calibrated based on the building's light-guiding performance parameter set. The grid must meet the following requirements: the light intensity within the grid is within the operating threshold of the photovoltaic cell assembly (for example, 100-1500 W / m²); there must be no areas with excessive light intensity or shadows; and the local temperature rise must be less than 5°C (to avoid hot spot effects).

[0051] Furthermore, based on the physical size of the photovoltaic array, the single grid area is preset and divided into a grid matrix on the array surface. The light intensity distribution data of the building light guide performance parameter set and the coordinates of the effective light guide area are combined to eliminate the grids with excessive light intensity and shadow areas. The remaining effective grids (such as 70) must meet the requirements of light intensity values ​​within the working threshold of the photovoltaic cell module and local temperature rise <5°C. Then, each effective grid is numerically integrated: the grid light intensity value is multiplied by the single grid area to obtain the input light power, and then multiplied by the local conversion efficiency under the current optimal power point parameters. Finally, the power generation of all effective grids is accumulated and divided by the total light input to generate the local photoelectric conversion efficiency. The efficiency distribution heat map of each grid is output for array health diagnosis. Among them, the calculation formula for the local conversion efficiency of a single grid is: in, is the single grid local conversion efficiency, is the maximum power point voltage, is the maximum power point current, is the measured light intensity value of the grid, is the physical area of ​​a single grid.

[0052] The optimal power point parameters of the hybrid MPPT output, the local photoelectric conversion efficiency and uniformity index, and the proportion of the effective light-guiding area are encapsulated into a key-value pair mapping table through structured data to form a photovoltaic operation status data set.

[0053] S5. Based on the photovoltaic operation status dataset, perform building energy hierarchical scheduling, obtain energy scheduling instruction sets and cross-domain energy scheduling coordination logs; The total power supply measurement value of the measurement area power grid is received in real time through the electric energy remote terminal, and the power supply data is intercepted and accumulated through the start and end times of the photovoltaic data acquisition time window to obtain the power grid load demand data.

[0054] Furthermore, the second-level pulse signal of the total power supply of the power grid in the measurement area is received in real time through the electric energy remote terminal. Based on the start and end times of the photovoltaic data collection time window, the cumulative number of pulses in the period is intercepted, multiplied by the pulse equivalent to convert it into the actual power supply value, and then divided by the time window length (such as 4 hours) to obtain the average power. The average power in the time window is used as the load reference value, and the measured power fluctuation range (such as ±5%) and periodic change characteristics (such as a 10% increase from 12:00 to 14:00 on weekdays) in the period are superimposed to form a dynamic demand description with a confidence interval, thereby generating power grid load demand data.

[0055] The local photoelectric conversion efficiency, effective light-guiding area ratio, and uniformity index are extracted from the photovoltaic operation status dataset. Outliers are removed through Kalman filtering, and the grid load demand data is aligned by time window to generate a spatiotemporal calibration dataset. Furthermore, a Kalman filter was used for iterative optimization with a step size of 1 second to eliminate outliers caused by sudden changes in cloud cover (for example, the efficiency jump from 20% to 80% in the 5th second was corrected to 23%). At the same time, grid load demand data was captured according to a fixed time window (such as 1 minute) (such as the average power of 2,850kW±142kW from 08:30:00 to 08:31:00). The filtered photovoltaic parameters were precisely aligned to the starting time of the time window according to the timestamp (error <50ms), generating a calibration dataset containing spatiotemporal matching fields.

[0056] Based on the physical node topology of the distribution network, the power generation layer, transmission layer, distribution layer, and user layer are explicitly divided, and the grid node connection relationship with hierarchical labels is generated; It should be noted that the physical node topology of the distribution network refers to a node connection relationship diagram constructed based on the actual physical structure of the power grid (such as the longitude and latitude coordinates of the substation, the switching station, the distribution transformer, the spatial position relationship of the poles, and the cable connection path of the feeder branch). Each physical node contains voltage level, capacity, and spatial coordinate attributes, and the current transmission path is explicitly marked through topological connections (such as node A→switch K1→node B), forming a weighted connection network that can be mapped to a geographic information system.

[0057] The node type-voltage level association rule is used to directly map the hierarchical grid node connection relationship into the hierarchical framework of the building energy hierarchical scheduling model. It should be noted that the node type-voltage level association rule is a predefined grid hierarchical division standard mapping table, which stipulates the mandatory hierarchical affiliation corresponding to different node types and their voltage levels, forming a rigid mapping logic from node physical attributes to hierarchical labels.

[0058] Furthermore, according to the node type-voltage level association rule table, each node in the grid node connection relationship with the hierarchical label is traversed, and the hierarchical affiliation (for example, 500kV substation → transmission layer) is matched according to the node type (power station, substation, distribution cabinet and meter) and its voltage level. Then, based on the topological connection edge (such as cable A→B), the energy transfer relationship between the layers (such as power generation layer → transmission layer) is established, and finally a hierarchical power distribution model framework is generated.

[0059] The load forecast data is called from the historical load database of the power grid in the measurement area to generate dynamic confidence interval constraints for power at each level; It should be noted that the power dynamic confidence interval constraint is a safe operation boundary of the power grid generated based on the historical load fluctuation pattern and the current load forecast value. Its core is to quantify the prediction uncertainty through statistical analysis and set the allowable power fluctuation range for each level of the power grid.

[0060] Furthermore, load time series data for the target time period (such as the load value every 15 minutes in the past 30 days) is collected from the historical load database of the measurement area power grid. After grouping by level labels (generation, transmission, distribution and user level), statistical characteristics are calculated for the data sets at each level. Combined with the load forecasting model, the power trend value for the next 24 hours is output. The confidence scaling factor k (such as the default k=2) is superimposed to obtain the dynamic confidence interval, and the power dynamic confidence interval constraints for each level are generated.

[0061] Based on the hierarchical framework and power dynamic confidence interval constraints, combined with the spatial distribution characteristics of local photovoltaic conversion efficiency in the photovoltaic operation status dataset, a building energy hierarchical scheduling model is created. The spatiotemporal calibration dataset is input into the building energy hierarchical scheduling model, and the allocation algorithm is applied at each level (generation layer, transmission layer, distribution layer, user layer) to generate a hierarchical power allocation result. The steps are as follows: The local layer enforces priority allocation of energy storage power in high-efficiency areas through power dynamic confidence interval constraints; It should be noted that the high-efficiency area refers to a collection of photovoltaic power generation units in the photovoltaic array surface grid that simultaneously meet the requirements of local photoelectric conversion efficiency significantly higher than the average level, fast dynamic response capability and gentle temperature rise changes. Its real-time screening is based on the spatial distribution characteristics of light intensity, temperature rise monitoring data and dynamic response performance indicators in the photovoltaic operation status data set. In the local power distribution, forced temperature rise is prioritized to allocate the main dispatchable power to the energy storage units in this area. By avoiding abnormal light intensity areas and risk points, the core scheduling goals of greatly improving the absorption efficiency and significantly reducing the risk of hot spots are achieved.

[0062] The regional layer dynamically optimizes the transmission loss compensation amount through power dynamic confidence interval constraints and real-time transmission loss; It should be noted that the amount of electric loss compensation refers to the power increment value dynamically calculated to offset the energy transmission loss caused by resistance thermal effects, electromagnetic eddy currents and ambient temperature and humidity in the transmission line. Its real-time optimization process first obtains the theoretical loss rate based on the line impedance parameters and layered distribution current values ​​of the grid node topology (a hierarchical network model constructed by the physical entity connection relationship of the distribution network). Then, combined with the predicted load interval of the dynamic confidence interval of the transmission layer power, the gradient descent method is used to solve the compensation power that meets the minimum compensation requirements and does not exceed the limit. Finally, the compensation amount is superimposed on the target distribution power to achieve cross-layer energy balance.

[0063] The cross-domain layer generates bid power by combining power dynamic confidence interval constraints with grid load demand data and load fluctuation prediction; It should be noted that load fluctuation refers to the periodic changes in the standard deviation of power at each level in the power grid over time. Its dynamic characteristics are manifested in the separation of the change rates of node groups in different regions (for example, the sudden increase rate of load in industrial areas is significantly higher than that in residential areas), and the probability of sudden changes in risk in future time periods is quantified through prediction models (such as the probability of sudden drops and the extreme amplitude). Finally, the rigid boundaries of cross-domain bidding power are generated in combination with the constraints of dynamic power confidence intervals to ensure that trading decisions are compatible with both the time-varying laws of load and the safety margin of the power grid.

[0064] The hierarchical power distribution results are encapsulated into structured parameters, and an energy scheduling instruction set is constructed through an instruction parameter compilation engine to optimize the scheduling of lighting energy loads.

[0065] Furthermore, the layered power distribution results of the building's photovoltaic power generation layer, transmission layer, distribution layer, and user layer within the building, including the local layer energy storage power allocation value, the regional layer transmission loss compensation amount, and the cross-domain layer bid power, are encapsulated into a unified building energy scheduling structured parameter set. Subsequently, the building energy management system instruction parameter compilation engine performs syntax compliance verification to verify whether the power exceeds the limit, uses protocol coding conversion (such as compiling into IEC 61850-MMS control messages), injects time offsets to offset transmission delays for timing offset compensation, and generates an energy scheduling instruction set with a security signature (such as SHA-256).

[0066] During the optimization and scheduling of solar energy loads, the actual power transmission, delay, and deviation are monitored in real time. If the deviation exceeds the delay limit, the energy scheduling instruction set is triggered to retry and record an alarm, and a cross-domain energy scheduling collaboration log is generated simultaneously.

[0067] Furthermore, during the optimal scheduling of solar energy loads, the transmission delay (e.g., >500ms) and power deviation (e.g., >5%) between the actual power transmission and the scheduling instructions are monitored in real time. When it is detected that the deviation value exceeds the dynamic delay threshold (e.g., a power deviation of 6.2% caused by a transmission delay of 800ms >5% tolerance), the energy scheduling instruction set retry mechanism is triggered to reissue the instructions and compensate for the correction amount (e.g., -1.8% power). At the same time, the alarm event is recorded and a cross-domain energy scheduling collaboration log is generated synchronously.

[0068] This embodiment also provides a computer device, which is applicable to the case of a method for optimizing and scheduling the comprehensive lighting energy load of a zero-carbon energy storage building, and includes: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the method for optimizing and scheduling the comprehensive lighting energy load of a zero-carbon energy storage building proposed in the above embodiment.

[0069] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0070] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for optimizing the scheduling of the integrated lighting energy load of a zero-carbon energy storage building as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0071] In summary, the present invention achieves millimeter-level spatial accuracy in identifying local hotspots and shadow areas through: a heat map of light intensity distribution on the building surface, accurately quantifying the energy distribution and path characteristics of indirect light on the surface of photovoltaic panels; and combining the building energy hierarchical scheduling model to fuse the photovoltaic operation status dataset and the grid hierarchy framework, generating an energy scheduling instruction set to drive energy storage priority call, dynamic compensation for line losses, and cross-domain power bidding, thus achieving cross-temporal and spatial coordinated optimization of light energy and grid load, and achieving the goal of efficient scheduling of the entire photovoltaic-storage-grid link in zero-carbon buildings.

[0072] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for optimizing the scheduling of comprehensive lighting energy loads in zero-carbon energy storage buildings, characterized by: include, Collect multi-dimensional light and thermal environment monitoring data, fuse them through principal component analysis dimensionality reduction and K-means clustering, and generate the building light and thermal environment feature matrix; Based on the building's light and thermal environment characteristic matrix, the reflected light path is identified through the light path random sampling simulation method to construct a heat map of the building's surface light intensity distribution; According to the light intensity distribution heat map of the building surface, the deflection angle of the building reflector array is dynamically adjusted, and the building light guide performance parameter set is generated through the dynamic light field efficiency quantification method; Based on the building light guiding performance parameter set, a hybrid MPPT algorithm is used to obtain the photovoltaic operation status data set; Based on the photovoltaic operation status data set, building energy hierarchical scheduling is carried out to obtain energy scheduling instruction sets and cross-domain energy scheduling coordination logs; The energy scheduling instruction set is updated regularly based on collaborative operation logs and historical environmental data.

2. The method for optimizing and scheduling comprehensive lighting energy loads for zero-carbon energy storage buildings according to claim 1, characterized in that: The principal component analysis dimensionality reduction and K-means clustering are combined to generate the building light and thermal environment feature matrix. The steps are as follows: The multi-dimensional photothermal environment monitoring data is Z-score standardized, the covariance matrix is ​​calculated, and the feature decomposition is performed through component analysis and dimensionality reduction to extract the photothermal coupling characteristics; Based on the light-thermal coupling characteristics, the K-means clustering algorithm is used to initialize the cluster centers, and the building light-thermal environment characteristic matrix is ​​generated through Euclidean distance measurement and iterative optimization.

3. The method for optimizing and scheduling comprehensive lighting energy loads for zero-carbon energy storage buildings according to claim 1, characterized in that: The steps of identifying the reflected light path by the light path random sampling simulation method and constructing the light intensity distribution heat map of the building surface are as follows: Based on the building light and thermal environment characteristic matrix and surface material property library, a reflection characteristic mapping table is generated through dynamic parameter mapping rules; Based on the reflection characteristic mapping table, the Monte Carlo random sampling method is used to simulate the light incident angle distribution, and the BRDF response surface is dynamically generated in combination with the material property library; The reflection direction and energy attenuation are calculated based on the BRDF response surface, the hitting coordinates and residual energy of the light on the photovoltaic panel surface are recorded, and the reflection path is identified through path clustering analysis. At the same time, the residual energy value is accumulated and normalized to light intensity. The spatial discretization grid statistics and Gaussian filter reconstruction technology are used to generate a heat map of the light intensity distribution on the building surface.

4. The method for optimizing and scheduling comprehensive lighting energy loads for zero-carbon energy storage buildings according to claim 1, wherein: The deflection angle of the building reflector array is dynamically adjusted according to the thermal map of light intensity distribution on the building surface. The steps are as follows: According to the heat map of light intensity distribution on the building surface, the abnormal area is located through the three-dimensional coordinate transformation matrix, and the effective light guiding area is obtained based on the collective difference between the total area and the abnormal area. Based on the distribution of abnormal areas and combined with the initial installation parameters of the reflector array, the reflector deflection angle is obtained, and the gradient descent method is used to optimize the reflector deflection angle.

5. The method for optimizing and scheduling comprehensive lighting energy loads for zero-carbon energy storage buildings according to claim 4, characterized in that: The steps of generating the building light guiding performance parameter set by the dynamic light field efficiency quantification method are as follows: The optimized reflector deflection angle is converted into a pulse signal through a stepper motor pulse equivalent, and the reflector array is driven to adjust the deflection angle of the building reflector array; The deflection angle error of the actual building reflector array is monitored synchronously through a photoelectric encoder, and the number of pulses is dynamically compensated using a PID algorithm to adjust the deflection angle of the building reflector array again; A high-speed spectrometer is used to collect the adjusted light intensity distribution data, and the photothermal conversion efficiency is calculated through the dynamic light field efficiency quantification method, and a set of building light guidance performance parameters is generated.

6. The method for optimizing and scheduling comprehensive lighting energy loads for zero-carbon energy storage buildings according to claim 1, characterized in that: The hybrid MPPT algorithm is used to obtain the photovoltaic operation status data set, and the steps are as follows: Dynamically assign weights to the perturbation-observation method and the conductance increment method, extract the future illumination trend time series characteristics through sliding window time series analysis, and construct a hybrid MPPT algorithm. The building light guiding performance parameter set is combined with the series-parallel topology of the photovoltaic array to generate the initial VI reference curve through least square fitting; Execute the hybrid MPPT algorithm, trigger mode switching based on the light intensity gradient threshold, output candidate optimal power point parameters through confidence voting, and calculate the deviation by comparing the initial VI reference curve in real time. At the same time, calibrate through Bayesian parameter inference to generate the optimal power point parameters; Based on the optimal power point parameters and the building light guiding performance parameter set, combined with the physical size of the photovoltaic array, the effective grid is marked, the effective grid is numerically integrated, and the local photoelectric conversion efficiency is obtained through integral calculation; The optimal power point parameters, local photoelectric conversion efficiency, uniformity index and effective light guiding area ratio output by the hybrid MPPT algorithm are encapsulated into a key-value pair mapping table through structured data to form a photovoltaic operation status data set.

7. The method for optimizing and scheduling comprehensive lighting energy loads for zero-carbon energy storage buildings according to claim 1, characterized in that: The steps for performing building energy hierarchical scheduling based on the photovoltaic operation status data set are as follows: Combined with the photovoltaic operating status dataset, outliers are removed through Kalman filtering to generate a spatiotemporal calibration dataset; Based on the physical node topology of the distribution network, the power generation layer, transmission layer, distribution layer, and user layer are explicitly divided, and the grid node connection relationship with hierarchical labels is generated. Combined with the photovoltaic operation status dataset, a building energy hierarchical scheduling model is created; The spatiotemporal calibration dataset is input into the building energy hierarchical scheduling model, and the allocation algorithm is applied at each level to generate the hierarchical electricity allocation results.

8. The method for optimizing and scheduling comprehensive lighting energy loads for zero-carbon energy storage buildings according to claim 7, characterized in that: The steps for obtaining the energy scheduling instruction set and the cross-domain energy scheduling collaboration log are as follows: The hierarchical power distribution results are encapsulated into structured parameters, and an energy scheduling instruction set is constructed through the instruction parameter compilation engine to optimize the scheduling of lighting energy loads. Real-time monitoring of actual power transmission and solar energy load optimization scheduling delay and deviation. If the deviation is greater than the delay limit, the energy scheduling instruction set is triggered to retry and record an alarm, and a cross-domain energy scheduling collaboration log is generated simultaneously.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for optimizing and scheduling the comprehensive lighting energy load of a zero-carbon energy storage building according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for optimizing and scheduling the comprehensive lighting energy load of a zero-carbon energy storage building according to any one of claims 1 to 8 are implemented.

Images