Road and tunnel lighting LED luminaire energy efficiency data analysis system and method

Abstract

This invention discloses a data analysis system and method for energy efficiency of LED luminaires used in road and tunnel lighting, relating to the field of photoelectric monitoring technology. The system includes acquiring energy efficiency label registration data for LED luminaires used in road and tunnel lighting; performing scenario-based standardization processing on the registration data according to road and tunnel lighting energy efficiency standards to obtain standardized data; performing anomaly marking and producer analysis through a two-dimensional anomaly detection mechanism to obtain two-dimensional anomaly marking results, producer classification markings, and anomaly feature profile datasets; and dynamically updating hierarchical time-series data and related derived data in the hidden layer through a time-drift-aware hierarchical self-calibration incremental iteration mechanism to output dynamic energy efficiency analysis results. This invention can meet the analysis needs of different scenarios, improve the relevance and practicality of energy efficiency data applications, and provide reliable data support for the formulation and revision of energy efficiency standards for LED luminaires used in road and tunnel lighting and for energy-saving policy research.

Description

Technical Field

[0001] This invention relates to the field of lighting energy efficiency data technology, and in particular to a system and method for analyzing energy efficiency data of LED lighting fixtures for road and tunnel lighting. Background Technology

[0002] With the rapid development of LED lighting technology, LED luminaires have been widely used in road and tunnel lighting due to their high luminous efficiency. Exploring their energy-saving potential is of great significance for promoting green lighting and implementing energy-saving policies. Energy efficiency standards, as the core basis for guiding enterprise technological progress and regulating market order, require a large amount of real and valid product energy efficiency data for their formulation and revision. The energy efficiency data for LED luminaires used in road and tunnel lighting mainly comes from the energy efficiency labeling registration system.

[0003] Currently, the registration system has accumulated data on nearly 80,000 product models, each containing over 50 parameters, resulting in a massive and complex dataset. Existing processing methods have significant limitations: firstly, reliance on manual data processing and statistics is not only costly in terms of manpower and time but also susceptible to human error, leading to insufficient data validity; secondly, the lack of scientific and standardized analytical methods makes it difficult to efficiently reveal the distribution characteristics and trends of energy efficiency indicators, as well as the correlation between energy efficiency indicators and performance indicators such as power and color temperature, hindering support for the formulation and revision of energy efficiency standards and research on related energy-saving policies. Therefore, a technical solution capable of automated data processing and systematic statistical analysis is urgently needed to address the aforementioned problems of low data processing efficiency and non-standardized analysis. Summary of the Invention

[0004] In view of the existing problems mentioned above, a data analysis system and method for energy efficiency of LED lighting fixtures for roads and tunnels is proposed.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0006] In a first aspect, the present invention includes the following steps:

[0007] Obtain energy efficiency label filing data for LED lighting products for road and tunnel lighting; perform scenario-based standardization processing on the filing data according to the energy efficiency standards for road and tunnel lighting to obtain standardized data; and construct a usable map based on the standardized data using a three-level hierarchical coding system.

[0008] Based on the standardized data, a statistical analysis task of time series parameter evolution is performed hierarchically to generate hierarchical time series statistical results, which are then stored in a hidden layer.

[0009] Based on the available graph and hierarchical time-series statistical results, parameter association datasets are obtained by performing parameter association mining on the filing data through a hierarchical attention network with lighting scene differentiation.

[0010] Based on the standardized data, hierarchical time-series statistical results and parameter association dataset, anomaly labeling and producer analysis are performed through a two-dimensional anomaly detection mechanism to obtain two-dimensional anomaly labeling results, producer hierarchical labeling and anomaly feature profile dataset.

[0011] Based on real-time energy efficiency label filing data entry, historical data adjustment and multi-dimensional analysis dataset, the hierarchical time-series data and related derived data in the hidden layer are dynamically updated through a time-series drift-aware layered self-calibration incremental iteration mechanism, and dynamic energy efficiency analysis results are output.

[0012] Furthermore, the method for constructing a usable map based on the standardized data using three-level hierarchical coding includes:

[0013] Based on the differences in energy efficiency standards for road and tunnel lighting, the product type, luminous efficacy, and power consumption parameters in the filing data are used to distinguish the scene categories of roads and tunnels. Standardized data is obtained by unifying parameter formats, correcting outliers, and labeling scene data.

[0014] Based on the standardized data, product, parameter, and producer information codes are extracted and normalized according to hierarchical standards. The weight calculation rules for each level's codes are dynamically adjusted based on the energy efficiency characteristics of these three levels. Anomalies within a level are identified based on the energy efficiency baseline, and the coding weight of anomaly information is reduced through a penalty factor. Groups are formed according to product type and power, parameter type and distribution, and producer scale and compliance level, and the relationships between enterprises, products, and parameters within each group are established. A hierarchical energy efficiency correlation map is constructed using the following formula:

[0015] ;

[0016] in, The comprehensive quantization value of the usable graph constructed for the three-level hierarchical coding, where m=1,2,3 represent the product level, parameter level, and producer level, respectively. Let m be the dynamic weight of the m-th level. This is the encoded value for the m-th level. This represents the maximum value of the m-th level of encoding. Let m be the power-law adjustment factor at the m-th level. Let m be the set of neighboring nodes within the m-th level. Let be the association strength between a node in the m-th level and its neighboring node k. Let be the mean of the association strength between neighboring nodes within the m-th level. It is a very small constant. This is the minimum value of the m-th level encoding. This is another power-law adjustment factor at the m-th level. This is the abnormal penalty coefficient. Let m be the number of outlier data in the m-th level. This represents the total number of data points at the m-th level.

[0017] Furthermore, a method for generating hierarchical time-series statistical results and storing them hierarchically in a hidden layer by performing a statistical analysis task of time-series parameter evolution hierarchically includes:

[0018] Based on the three-level hierarchy of product, parameter, and producer, time-series analysis objects are divided, and a hierarchical time-series analysis unit is constructed. Based on the scenario-adaptive time-series cycle, robust discrete quantification is used to measure the periodic fluctuations of parameters within the three levels. The quantitative trend slope and stability of energy efficiency indicators are fitted within the three levels and the time-series cycle. Considering the seasonal sensitivity of roads and the stable environment of tunnels, the modulation coefficients of the time-series cycle are dynamically configured to obtain hierarchical time-series statistical results, which are then stored hierarchically in a hidden layer. The formula is:

[0019] ;

[0020] in, For hierarchical time series statistics, This represents the net change in energy efficiency of level L within the time window t. Let L be the variance of the k-th parameter at time t. Let L be the energy efficiency correlation between level L and its neighboring level s at time t. Let L be the temporal distance between hierarchy L and its neighbor s. For dynamic weights.

[0021] Furthermore, the method for obtaining a parameter association dataset by performing parameter association mining on the registration data through a hierarchical attention network differentiated by lighting scenes includes:

[0022] Based on the standardized data, available graphs, and hierarchical time-series statistical results, the parameter correlation strength is obtained by distinguishing between road and tunnel scenarios and focusing on a three-tiered attention network and self-attention mechanism involving products, parameters, and producers. The formula is:

[0023] ;

[0024] in, For scene identification, For hierarchical identification, , For the two parameters to be associated, Available spectral quantization values ​​for three-level hierarchical coding. For hierarchical time series statistics, For dynamic weight parameters, The prior correlation strength of intrinsic parameters at the same level. The query vector for parameters. The key vector for parameters;

[0025] Based on the parameter association strength, parameter association mining is performed, traversing all dimensions of scene, level, time window, and parameter pair combination, and binding dimension identifiers with association strength values ​​to obtain parameter association dataset.

[0026] Furthermore, methods for obtaining two-dimensional anomaly labeling results, producer hierarchical labeling, and anomaly feature profiling datasets include:

[0027] Based on the standardized data, hierarchical time-series statistical results, and parameter association datasets, a two-dimensional anomaly score is obtained by comparing parameter and time fluctuation anomalies with parameter association mining results against a scenario-customized association benchmark to determine the inverse or insufficient association strength between parameters. The formula is:

[0028] ;

[0029] in, For producer identification, To standardize data, Let L be the time series mean of level L, parameter k, and time t, and let be the standard deviation of the fluctuation of level L, parameter k, and time t. For time fluctuation anomalies of power, For producer differentiation weights, For normal correlation baseline, To correlate with abnormal powers of exponentiation, This is the time-series trend penalty coefficient. The power of the difference in trends over time;

[0030] Based on the standardized data, anomaly-marked datasets are obtained by threshold determination; according to the comparison between the two-dimensional anomaly degree and the preset anomaly threshold, when the two-dimensional anomaly degree exceeds the preset threshold, the dataset is determined to be abnormal, otherwise it is determined to be normal.

[0031] Based on the aforementioned anomaly-labeled dataset, by accumulating the anomaly degree of the same producer across time windows, across products, parameters, and producer levels, and using the industry's maximum anomaly degree as a benchmark, the cross-level anomaly degree of each producer is normalized to a unified interval. Then, a piecewise mapping function is used to convert the normalized anomaly degree into a preset discrete level to obtain the producer classification labeling result. The formula is:

[0032] ;

[0033] in, The abnormality is accumulated to a power. For hierarchical weights, The maximum sum of the anomalies of all producers is calculated. The maximum value of the classification. (*) represents the mapping hierarchical function;

[0034] Based on standardized data, hierarchical time-series statistical results, and producer-level feature datasets, a single-dimensional anomaly feature vector is obtained by calculating the most abnormal parameter and considering the anomaly occurrence time, hierarchy, and producer-level features. The formula is:

[0035] ;

[0036] in, This represents the parameter k that contributes the most to the anomaly score. The producer is coded in a three-level hierarchy;

[0037] Based on the single-dimensional abnormal feature vector, it is associated and stored according to the dimensions of scene, producer, level, and time, and finally a multi-dimensional abnormal feature profile dataset containing multi-dimensional indexes and abnormal feature mapping relationships is formed.

[0038] Furthermore, a method for dynamically updating hierarchical time-series data and associated derived data in the hidden layer through a time-drift-aware hierarchical self-calibration incremental iteration mechanism to output dynamic energy efficiency analysis results includes:

[0039] The formula for outputting the dynamic energy efficiency optimization priority index is:

[0040] ;

[0041] in, The energy efficiency optimization priority index is given for scenario s, producer p, parameter k, and time t. To correct the power order in stages, Contribute an exponential power to the correlation deviation. For the historical optimization cost of producer p, parameter k, and time t, The benchmark deviation penalty coefficient, The power of the reference deviation. This represents the total number of historical time windows.

[0042] Based on the hierarchical time-series data and associated derived data in the hidden layer that are dynamically updated according to the hierarchical self-calibration incremental iteration mechanism of time-series drift perception, the dynamic energy efficiency optimization priority index is output through the two-dimensional anomaly marking results, producer classification and parameter correlation strength and benchmark value, time-series mean, industry energy efficiency benchmark and historical optimization cost, through the value amplification of anomalies and classifications and the contribution quantification of parameter correlation deviation.

[0043] Based on the dynamic energy efficiency optimization priority index, the dynamic energy efficiency analysis results are output through the available spectrum.

[0044] The associated derived data includes anomaly label distribution and producer hierarchical label; the multi-dimensional analysis dataset includes two-dimensional anomaly label results, producer hierarchical label, hierarchical time-series statistical results, and anomaly feature profile dataset.

[0045] On the other hand, an energy efficiency data analysis system for LED lighting fixtures for roads and tunnels includes:

[0046] Data standardization module: used to standardize the multi-source energy efficiency data of LED lighting fixtures for road and tunnel lighting, including noise reduction, normalization and missing value imputation, to generate standardized data; the multi-source energy efficiency data includes product performance data, operating parameter data and producer registration data;

[0047] Hierarchical Feature Construction Module: Used to extract hierarchical features from standardized data and construct a three-level hierarchical code: product level, parameter level, and producer level; the hierarchical features include product compliance, parameter integrity, producer scale, and compliance attributes;

[0048] Hierarchical Time Series Analysis Module: Used to perform time series statistics on standardized data at three levels, calculate hierarchical time series mean, variance of fluctuation, energy efficiency trend and time series change slope, and generate hierarchical time series statistical results;

[0049] Parameter association mining module: used to fuse hierarchical features and hierarchical time-series statistical results, quantify the correlation strength between parameters through a scenario-based self-attention mechanism, and generate a parameter association dataset; the scenario-based approach includes differentiated processing for road scenarios and tunnel scenarios;

[0050] The two-dimensional anomaly analysis module is used to construct a two-dimensional anomaly detection model based on hierarchical time-series statistical results and parameter correlation datasets, calculate the anomaly degree, and output a two-dimensional anomaly label, producer classification, and anomaly feature profile dataset.

[0051] Energy efficiency optimization decision module: It integrates anomaly degree, producer classification, parameter correlation deviation and industry energy efficiency benchmarks to calculate the energy efficiency optimization priority index and output optimization priority ranking and rectification suggestions.

[0052] The beneficial effects of this invention are:

[0053] This invention enables the rapid import and integration of multiple files and multi-year data, replacing the traditional manual processing mode, significantly reducing manpower and time investment, and significantly improving data processing efficiency. Through standardized processing steps such as product type conversion, data format unification, and abnormal data labeling, it can accurately identify and mark abnormal or erroneous data, avoiding errors caused by manual operation, ensuring that subsequent analysis is based on high-quality data, meeting the analysis needs of different scenarios, improving the relevance and practicality of energy efficiency data applications, and providing reliable data support for the formulation and revision of energy efficiency standards for road and tunnel lighting LED lamps and the research on energy-saving policies. Attached Figure Description

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 A flowchart for analyzing energy efficiency data of LED lighting fixtures for road and tunnel lighting. Detailed Implementation

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0057] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" as used 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 different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0059] Reference Figure 1As an embodiment of the present invention, 500 registered models of LED lighting fixtures for roads and tunnels from 100 domestic manufacturers in 2023-2024 were selected (300 for roads and 200 for tunnels), covering more than 50 parameters such as rated power (10-300W), correlated color temperature (2700K-6500K), luminous efficacy (60-150lm / W), and power factor (0.7-0.99).

[0060] This invention includes the following steps:

[0061] Obtain energy efficiency label filing data for LED lighting products for road and tunnel lighting; perform scenario-based standardization processing on the filing data according to the energy efficiency standards for road and tunnel lighting to obtain standardized data; and construct a usable map based on the standardized data using a three-level hierarchical coding system.

[0062] Based on the standardized data, a statistical analysis task of time series parameter evolution is performed hierarchically to generate hierarchical time series statistical results, which are then stored in a hidden layer.

[0063] Based on the available graph and hierarchical time-series statistical results, parameter association datasets are obtained by performing parameter association mining on the filing data through a hierarchical attention network with lighting scene differentiation.

[0064] Based on the standardized data, hierarchical time-series statistical results and parameter association dataset, anomaly labeling and producer analysis are performed through a two-dimensional anomaly detection mechanism to obtain two-dimensional anomaly labeling results, producer hierarchical labeling and anomaly feature profile dataset.

[0065] Based on real-time energy efficiency label filing data entry, historical data adjustment and multi-dimensional analysis dataset, the hierarchical time-series data and related derived data in the hidden layer are dynamically updated through a time-series drift-aware layered self-calibration incremental iteration mechanism, and dynamic energy efficiency analysis results are output.

[0066] In this embodiment, the method for constructing a usable graph based on the standardized data through three-level hierarchical coding includes:

[0067] Based on the differences in energy efficiency standards for road and tunnel lighting, the product type, luminous efficacy, and power consumption parameters in the filing data are used to distinguish the scene categories of roads and tunnels. Standardized data is obtained by unifying parameter formats, correcting outliers, and labeling scene data.

[0068] Based on the standardized data, product, parameter, and producer information codes are extracted and normalized according to hierarchical standards. The weight calculation rules for each level's codes are dynamically adjusted based on the energy efficiency characteristics of these three levels. Anomalies within a level are identified based on the energy efficiency baseline, and the coding weight of anomaly information is reduced through a penalty factor. Groups are formed according to product type and power, parameter type and distribution, and producer scale and compliance level, and the relationships between enterprises, products, and parameters within each group are established. A hierarchical energy efficiency correlation map is constructed using the following formula:

[0069] ;

[0070] in, The comprehensive quantization value of the usable graph constructed for the three-level hierarchical coding, where m=1,2,3 represent the product level, parameter level, and producer level, respectively. Let m be the dynamic weight of the m-th level. This is the encoded value for the m-th level. This represents the maximum value of the m-th level of encoding. Let m be the power-law adjustment factor at the m-th level. Let m be the set of neighboring nodes within the m-th level. Let be the association strength between a node in the m-th level and its neighboring node k. Let be the mean of the association strength between neighboring nodes within the m-th level. It is a very small constant. This is the minimum value of the m-th level encoding. This is another power-law adjustment factor at the m-th level. This is the abnormal penalty coefficient. Let m be the number of outlier data in the m-th level. This represents the total number of data points at the m-th level.

[0071] The raw data was standardized for specific scenarios (distinguishing between road and tunnel scenarios), and a usable map was constructed using a three-level hierarchical coding system. Some results are shown below:

[0072] Product Type Producer ID Rated power (W) Luminous efficacy (lm / W) The three-level hierarchical coding comprehensive value h (the higher the value, the better the availability) Street lighting fixtures P01 150 120 0.89 Street lighting fixtures P02 80 95 0.76 Tunnel lighting fixtures P03 200 110 0.92 Tunnel lighting fixtures P04 60 70

[0073] The results show that tunnel lighting fixture P03 has the highest h value because its parameters are complete, there are no abnormalities, and the manufacturer is highly compliant; P04 has a low h value because there are 3 missing parameters and 2 abnormal values ​​(luminous efficacy is 30% lower than the industry average).

[0074] In this implementation example, the method for generating hierarchical time-series statistical results and storing them hierarchically in a hidden layer by performing a statistical analysis task of time-series parameter evolution hierarchically includes:

[0075] Based on the three-level hierarchy of product, parameter, and producer, time-series analysis objects are divided, and a hierarchical time-series analysis unit is constructed. Based on the scenario-adaptive time-series cycle, robust discrete quantification is used to measure the periodic fluctuations of parameters within the three levels. The quantitative trend slope and stability of energy efficiency indicators are fitted within the three levels and the time-series cycle. Considering the seasonal sensitivity of roads and the stable environment of tunnels, the modulation coefficients of the time-series cycle are dynamically configured to obtain hierarchical time-series statistical results, which are then stored hierarchically in a hidden layer. The formula is:

[0076] ;

[0077] in, For hierarchical time series statistics, This represents the net change in energy efficiency of level L within the time window t. Let L be the variance of the k-th parameter at time t. Let L be the energy efficiency correlation between level L and its neighboring level s at time t. Let L be the temporal distance between hierarchy L and its neighbor s. For dynamic weights.

[0078] The evolution of time-series parameters from Q1 2023 to Q2 2024 is statistically analyzed quarterly. Taking "luminous efficacy" as an example, the results of some hierarchical time-series statistics (T value, the higher the value, the better the time-series stability) show that the luminous efficacy of road lighting fixtures steadily increases with each quarter, and the fluctuation decreases (T value increases from 0.82 to 0.91); due to the stable environment, the time-series stability of tunnel lighting fixtures is always higher than that of road lighting fixtures (T value > 0.9).

[0079] In this embodiment, the method for obtaining a parameter association dataset by performing parameter association mining on the registration data through a hierarchical attention network differentiated by lighting scenes includes:

[0080] Based on the standardized data, available graphs, and hierarchical time-series statistical results, the parameter correlation strength is obtained by distinguishing between road and tunnel scenarios and focusing on a three-tiered attention network and self-attention mechanism involving products, parameters, and producers. The formula is:

[0081] ;

[0082] in, For scene identification, For hierarchical identification, , For the two parameters to be associated, Available spectral quantization values ​​for three-level hierarchical coding. For hierarchical time series statistics, For dynamic weight parameters, The prior correlation strength of intrinsic parameters at the same level. The query vector for parameters. The key vector for parameters;

[0083] Based on the parameter association strength, parameter association mining is performed, traversing all dimensions of scene, level, time window, and parameter pair combination, and binding dimension identifiers with association strength values ​​to obtain parameter association dataset.

[0084] In this implementation example, the method for obtaining the two-dimensional anomaly labeling results, producer hierarchical labeling, and anomaly feature profile dataset includes:

[0085] Based on the standardized data, hierarchical time-series statistical results, and parameter association datasets, a two-dimensional anomaly score is obtained by comparing parameter and time fluctuation anomalies with parameter association mining results against a scenario-customized association benchmark to determine the inverse or insufficient association strength between parameters. The formula is:

[0086] ;

[0087] in, For producer identification, To standardize data, Let L be the time series mean of level L, parameter k, and time t, and let be the standard deviation of the fluctuation of level L, parameter k, and time t. For time fluctuation anomalies of power, For producer differentiation weights, For normal correlation baseline, To correlate with abnormal powers of exponentiation, This is the time-series trend penalty coefficient. The power of the difference in trends over time;

[0088] Based on the standardized data, anomaly-marked datasets are obtained by threshold determination; according to the comparison between the two-dimensional anomaly degree and the preset anomaly threshold, when the two-dimensional anomaly degree exceeds the preset threshold, the dataset is determined to be abnormal, otherwise it is determined to be normal.

[0089] Based on the aforementioned anomaly-labeled dataset, by accumulating the anomaly degree of the same producer across time windows, across products, parameters, and producer levels, and using the industry's maximum anomaly degree as a benchmark, the cross-level anomaly degree of each producer is normalized to a unified interval. Then, a piecewise mapping function is used to convert the normalized anomaly degree into a preset discrete level to obtain the producer classification labeling result. The formula is:

[0090] ;

[0091] in, The abnormality is accumulated to a power. For hierarchical weights, The maximum sum of the anomalies of all producers is calculated. The maximum value of the classification. (*) represents the mapping hierarchical function;

[0092] Based on standardized data, hierarchical time-series statistical results, and producer-level feature datasets, a single-dimensional anomaly feature vector is obtained by calculating the most abnormal parameter and considering the anomaly occurrence time, hierarchy, and producer-level features. The formula is:

[0093] ;

[0094] in, This represents the parameter k that contributes the most to the anomaly score. The producer is coded in a three-level hierarchy;

[0095] Based on the single-dimensional abnormal feature vector, it is associated and stored according to the dimensions of scene, producer, level, and time, and finally a multi-dimensional abnormal feature profile dataset containing multi-dimensional indexes and abnormal feature mapping relationships is formed.

[0096] In this embodiment, a method for dynamically updating hierarchical time-series data and associated derived data in a hidden layer through a time-drift-aware hierarchical self-calibration incremental iteration mechanism, and outputting dynamic energy efficiency analysis results, includes:

[0097] The formula for outputting the dynamic energy efficiency optimization priority index is:

[0098] ;

[0099] in, The energy efficiency optimization priority index is given for scenario s, producer p, parameter k, and time t. To correct the power order in stages, Contribute an exponential power to the correlation deviation. For the historical optimization cost of producer p, parameter k, and time t, The benchmark deviation penalty coefficient, The power of the reference deviation. This represents the total number of historical time windows.

[0100] Based on the hierarchical time-series data and associated derived data in the hidden layer that are dynamically updated according to the hierarchical self-calibration incremental iteration mechanism of time-series drift perception, the dynamic energy efficiency optimization priority index is output through the two-dimensional anomaly marking results, producer classification and parameter correlation strength and benchmark value, time-series mean, industry energy efficiency benchmark and historical optimization cost, through the value amplification of anomalies and classifications and the contribution quantification of parameter correlation deviation.

[0101] Based on the dynamic energy efficiency optimization priority index, the dynamic energy efficiency analysis results are output through the available spectrum.

[0102] The associated derived data includes anomaly label distribution and producer hierarchical label; the multi-dimensional analysis dataset includes two-dimensional anomaly label results, producer hierarchical label, hierarchical time-series statistical results, and anomaly feature profile dataset.

[0103] On the other hand, an energy efficiency data analysis system for LED lighting fixtures for roads and tunnels includes:

[0104] Data standardization module: used to standardize the multi-source energy efficiency data of LED lighting fixtures for road and tunnel lighting, including noise reduction, normalization and missing value imputation, to generate standardized data; the multi-source energy efficiency data includes product performance data, operating parameter data and producer registration data;

[0105] Hierarchical Feature Construction Module: Used to extract hierarchical features from standardized data and construct a three-level hierarchical code: product level, parameter level, and producer level; the hierarchical features include product compliance, parameter integrity, producer scale, and compliance attributes;

[0106] Hierarchical Time Series Analysis Module: Used to perform time series statistics on standardized data at three levels, calculate hierarchical time series mean, variance of fluctuation, energy efficiency trend and time series change slope, and generate hierarchical time series statistical results;

[0107] Parameter association mining module: used to fuse hierarchical features and hierarchical time-series statistical results, quantify the correlation strength between parameters through a scenario-based self-attention mechanism, and generate a parameter association dataset; the scenario-based approach includes differentiated processing for road scenarios and tunnel scenarios;

[0108] The two-dimensional anomaly analysis module is used to construct a two-dimensional anomaly detection model based on hierarchical time-series statistical results and parameter correlation datasets, calculate the anomaly degree, and output a two-dimensional anomaly label, producer classification, and anomaly feature profile dataset.

[0109] Energy efficiency optimization decision module: It integrates anomaly degree, producer classification, parameter correlation deviation and industry energy efficiency benchmarks to calculate the energy efficiency optimization priority index and output optimization priority ranking and rectification suggestions.

[0110] This embodiment also provides a computer device applicable to the method for analyzing energy efficiency data of LED lighting fixtures for road and tunnel lighting, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the method for analyzing energy efficiency data of LED lighting fixtures for road and tunnel lighting as proposed in the above embodiment.

[0111] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices 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 the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0112] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the method for analyzing energy efficiency data of LED lighting fixtures for road and tunnel lighting as proposed in the above embodiments. 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 Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0113] 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 it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for analyzing energy efficiency data of road and tunnel lighting LED luminaires, characterized in that, Includes the following steps: Obtain energy efficiency label filing data for LED lighting products for road and tunnel lighting; perform scenario-based standardization processing on the filing data according to the energy efficiency standards for road and tunnel lighting to obtain standardized data; and construct a usable map based on the standardized data using a three-level hierarchical coding system. Based on the standardized data, a statistical analysis task of time series parameter evolution is performed hierarchically to generate hierarchical time series statistical results, which are then stored in a hidden layer. Based on the available graph and hierarchical time-series statistical results, parameter association datasets are obtained by performing parameter association mining on the filing data through a hierarchical attention network with lighting scene differentiation. Based on the standardized data, hierarchical time-series statistical results and parameter association dataset, anomaly labeling and producer analysis are performed through a two-dimensional anomaly detection mechanism to obtain two-dimensional anomaly labeling results, producer hierarchical labeling and anomaly feature profile dataset. Based on real-time energy efficiency label filing data entry, historical data adjustment, and multi-dimensional analysis datasets, the hierarchical time-series data and related derived data in the hidden layer are dynamically updated through a time-series drift-aware, layered self-calibration incremental iteration mechanism, and dynamic energy efficiency analysis results are output.

2. The method of claim 1, wherein the method is a method of analyzing energy efficiency data of a road and tunnel lighting LED luminaire, characterized by, A method for constructing a usable graph based on the standardized data using three-level hierarchical coding includes: Based on the differences in energy efficiency standards for road and tunnel lighting, the product type, luminous efficacy, and power consumption parameters in the filing data are used to distinguish the scene categories of roads and tunnels. Standardized data is obtained by unifying parameter formats, correcting outliers, and labeling scene data. Based on the standardized data, product, parameter, and producer information codes are extracted and normalized according to hierarchical standards. The weight calculation rules for each level's codes are dynamically adjusted based on the energy efficiency characteristics of these three levels. Anomalies within a level are identified based on the energy efficiency baseline, and the coding weight of anomaly information is reduced through a penalty factor. Groups are formed according to product type and power, parameter type and distribution, and producer scale and compliance level, and the relationships between enterprises, products, and parameters within each group are established. A hierarchical energy efficiency correlation map is constructed using the following formula: ； in, The comprehensive quantization value of the usable graph constructed for the three-level hierarchical coding, where m=1,2,3 represent the product level, parameter level, and producer level, respectively. Let m be the dynamic weight of the m-th level. This is the encoded value for the m-th level. This represents the maximum value of the m-th level of encoding. Let m be the power-law adjustment factor at the m-th level. Let m be the set of neighboring nodes within the m-th level. Let be the association strength between a node in the m-th level and its neighboring node k. Let be the mean of the association strength between neighboring nodes within the m-th level. It is a very small constant. This is the minimum value of the m-th level encoding. This is another power-law adjustment factor at the m-th level. This is the abnormal penalty coefficient. Let m be the number of outlier data in the m-th level. This represents the total number of data points at the m-th level.

3. The method of claim 2, wherein the method further comprises: Methods for generating hierarchical time-series statistical results and storing them hierarchically in hidden layers by performing statistical analysis tasks on the evolution of time-series parameters according to hierarchy include: Based on the three-level hierarchy of product, parameter, and producer, time-series analysis objects are divided, and a hierarchical time-series analysis unit is constructed. Based on the scenario-adaptive time-series cycle, robust discrete quantification is used to measure the periodic fluctuations of parameters within the three levels. The quantitative trend slope and stability of energy efficiency indicators are fitted within the three levels and the time-series cycle. Considering the seasonal sensitivity of roads and the stable environment of tunnels, the modulation coefficients of the time-series cycle are dynamically configured to obtain hierarchical time-series statistical results, which are then stored hierarchically in a hidden layer. The formula is: ； in, For hierarchical time series statistics, This represents the net change in energy efficiency of level L within the time window t. Let L be the variance of the k-th parameter at time t. Let L be the energy efficiency correlation between level L and its neighboring level s at time t. Let L be the temporal distance between hierarchy L and its neighbor s. For dynamic weights.

4. The method for analyzing energy efficiency data of a road and tunnel lighting LED luminaire according to claim 3, wherein, A method for obtaining a parameter association dataset by performing parameter association mining on the aforementioned filing data using a hierarchical attention network differentiated by lighting scenes includes: Based on the standardized data, available graphs, and hierarchical time-series statistical results, the parameter correlation strength is obtained by distinguishing between road and tunnel scenarios and focusing on a three-level attention network and self-attention mechanism involving products, parameters, and producers. The formula is: ； in, For scene identification, For hierarchical identification, , For the two parameters to be associated, Available spectral quantization values ​​for three-level hierarchical coding. For hierarchical time series statistics, For dynamic weight parameters, The prior correlation strength of intrinsic parameters at the same level. The query vector for parameters. The key vector for parameters; Based on the parameter association strength, parameter association mining is performed, traversing all dimensions of scene, level, time window, and parameter pair combination, and binding dimension identifiers with association strength values ​​to obtain parameter association dataset.

5. The method for analyzing energy efficiency data of a road and tunnel lighting LED luminaire according to claim 4, wherein, Methods for obtaining two-dimensional anomaly labeling results, producer hierarchical labeling, and anomaly feature profiling datasets include: Based on the standardized data, hierarchical time-series statistical results, and parameter association datasets, a two-dimensional anomaly score is obtained by comparing parameter and time fluctuation anomalies with parameter association mining results against a scenario-customized association benchmark to determine the inverse or insufficient association strength between parameters. The formula is: ； in, For producer identification, To standardize data, Let L be the time series mean of level L, parameter k, and time t, and let be the standard deviation of the fluctuation of level L, parameter k, and time t. For time fluctuation anomalies of power, For producer differentiation weights, For normal correlation baseline, To correlate with abnormal powers of exponentiation, This is the time-series trend penalty coefficient. The power of the difference in trends over time; Based on the standardized data, anomaly-marked datasets are obtained by threshold determination; according to the comparison between the two-dimensional anomaly degree and the preset anomaly threshold, when the two-dimensional anomaly degree exceeds the preset threshold, the dataset is determined to be abnormal, otherwise it is determined to be normal. Based on the aforementioned anomaly-labeled dataset, by accumulating the anomaly degree of the same producer across time windows, across products, parameters, and producer levels, and using the industry's maximum anomaly degree as a benchmark, the cross-level anomaly degree of each producer is normalized to a unified interval. Then, a piecewise mapping function is used to convert the normalized anomaly degree into a preset discrete level to obtain the producer classification labeling result. The formula is: ； in, The abnormality is accumulated to a power. For hierarchical weights, The maximum sum of the anomalies of all producers is calculated. The maximum value of the classification. (*) represents the mapping hierarchical function; According to the standardized data, the hierarchical time sequence statistical result and the producer hierarchical characteristic data set, a single dimension abnormal characteristic vector is obtained through positioning operation of the most abnormal parameter and the abnormal occurrence time, level and producer three level characteristics , the formula is: ； wherein, denotes the parameter k that contributes most to the abnormality degree, is a producer three-level hierarchy encoding; Based on the single-dimensional abnormal feature vector, it is associated and stored according to the dimensions of scene, producer, level, and time, and finally a multi-dimensional abnormal feature profile dataset containing multi-dimensional indexes and abnormal feature mapping relationships is formed.

6. The method for analyzing energy efficiency data of a road and tunnel lighting LED luminaire according to claim 5, wherein, Methods for dynamically updating hierarchical time-series data and associated derived data in hidden layers through a time-drift-aware, layered self-calibration incremental iterative mechanism to output dynamic energy efficiency analysis results include: The formula for outputting the dynamic energy efficiency optimization priority index is: ； in, The energy efficiency optimization priority index is given for scenario s, producer p, parameter k, and time t. To correct the power order in stages, Contribute an exponential power to the correlation deviation. For the historical optimization cost of producer p, parameter k, and time t, The benchmark deviation penalty coefficient, The power of the reference deviation. This represents the total number of historical time windows. Based on the hierarchical time-series data and associated derived data in the hidden layer that are dynamically updated according to the hierarchical self-calibration incremental iteration mechanism of time-series drift perception, the dynamic energy efficiency optimization priority index is output through the two-dimensional anomaly marking results, producer classification and parameter correlation strength and benchmark value, time-series mean, industry energy efficiency benchmark and historical optimization cost, through the value amplification of anomalies and classifications and the contribution quantification of parameter correlation deviation. Based on the dynamic energy efficiency optimization priority index, the dynamic energy efficiency analysis results are output through the available spectrum. The associated derived data includes anomaly label distribution and producer hierarchical label; the multi-dimensional analysis dataset includes two-dimensional anomaly label results, producer hierarchical label, hierarchical time-series statistical results, and anomaly feature profile dataset.

7. A data analysis system for energy efficiency of LED luminaires used for road and tunnel lighting, for performing the method described in any one of claims 1-6, characterized in that, include: Data standardization module: used to standardize the multi-source energy efficiency data of LED lighting fixtures for road and tunnel lighting, including noise reduction, normalization and missing value imputation, to generate standardized data; the multi-source energy efficiency data includes product performance data, operating parameter data and producer registration data; Hierarchical Feature Construction Module: Used to extract hierarchical features from standardized data and construct a three-level hierarchical code: product level, parameter level, and producer level; the hierarchical features include product compliance, parameter integrity, producer scale, and compliance attributes; Hierarchical Time Series Analysis Module: Used to perform time series statistics on standardized data at three levels, calculate hierarchical time series mean, variance of fluctuation, energy efficiency trend and time series change slope, and generate hierarchical time series statistical results; Parameter association mining module: used to fuse hierarchical features and hierarchical time-series statistical results, quantify the correlation strength between parameters through a scenario-based self-attention mechanism, and generate a parameter association dataset; the scenario-based approach includes differentiated processing for road scenarios and tunnel scenarios; The two-dimensional anomaly analysis module is used to construct a two-dimensional anomaly detection model based on hierarchical time-series statistical results and parameter correlation datasets, calculate the anomaly degree, and output a two-dimensional anomaly label, producer classification, and anomaly feature profile dataset. Energy efficiency optimization decision module: It integrates anomaly degree, producer classification, parameter correlation deviation and industry energy efficiency benchmarks to calculate the energy efficiency optimization priority index and output optimization priority ranking and rectification suggestions.

Images