Energy efficiency prediction analysis platform based on artificial intelligence

Abstract

The invention discloses an energy efficiency prediction analysis platform based on artificial intelligence. The platform comprises an energy consumption acquisition module, an energy efficiency prediction module, an energy consumption analysis module, an energy saving optimization suggestion module and a comprehensive management module. The energy consumption acquisition module is used for acquiring energy consumption related information; the energy prediction module is used for processing the energy consumption related information to obtain energy prediction information; the energy consumption analysis module is used for performing energy consumption analysis to obtain energy consumption analysis information; the energy-saving optimization suggestion module is used for performing energy-saving optimization analysis and generating an energy-saving optimization suggestion; and the comprehensive management module is used for carrying out overall comprehensive management. According to the invention, more comprehensive and accurate energy efficiency prediction can be realized.

Description

Technical Field

[0001] The present invention relates to the field of predictive analysis, and particularly to an energy efficiency prediction and analysis platform based on artificial intelligence. Background Art

[0002] With the acceleration of the global industrialization and urbanization processes, energy consumption has increased sharply, and the energy issue has increasingly become a key factor restricting sustainable economic development and environmental protection. At the enterprise level, whether it is a manufacturing enterprise or a commercial office complex, the proportion of energy costs in the total operating costs has been continuously rising, and the need for refined energy management is becoming more and more urgent.

[0003] On the one hand, traditional energy management methods mostly rely on manual meter reading and empirical judgment, suffering from problems such as untimely and inaccurate data collection, insufficient analysis depth, and difficulty in coping with complex and changing energy consumption scenarios. For example, in the face of the energy consumption monitoring of numerous production equipment in a large factory, the manual method cannot capture energy consumption fluctuations in real time, making it difficult to detect abnormal energy consumption conditions of equipment in a timely manner.

[0004] On the other hand, in recent years, artificial intelligence technology has developed vigorously, and deep learning algorithms have demonstrated excellent performance in fields such as time series prediction and data analysis and mining. Big data technology provides support for the storage and processing of massive energy data, and the progress of sensor technology makes it possible to collect high-precision and diverse energy data. Against this background, an energy efficiency prediction and analysis platform based on artificial intelligence has emerged. Summary of the Invention

[0005] In order to solve the technical problems existing in the existing solutions, an energy efficiency prediction and analysis platform based on artificial intelligence is proposed.

[0006] The present invention solves the above technical problems through the following technical solutions. The present invention includes an energy consumption collection module, an energy efficiency prediction module, an energy consumption analysis module, an energy-saving optimization suggestion module, and a comprehensive management module;

[0007] The energy consumption collection module is used to collect information related to energy consumption;

[0008] The energy prediction module is used to process information related to energy consumption to obtain energy prediction information;

[0009] The energy consumption analysis module is used to perform energy consumption analysis to obtain energy consumption analysis information;

[0010] The energy-saving optimization suggestion module is used to perform energy-saving optimization analysis and generate energy-saving optimization suggestions;

[0011] The comprehensive management module is used to perform overall comprehensive management.

[0012] Furthermore, the specific process of the energy consumption collection module collecting information related to energy consumption is as follows:

[0013] Connect data acquisition equipment and set up various types of sensor interfaces to access energy acquisition equipment, including power sensors, water meters and gas meter sensors;

[0014] Then, the data collection frequency is set. Different collection frequencies are set according to the stability of energy consumption and the analysis accuracy requirements, and data is collected according to the set collection frequencies.

[0015] Data preprocessing: After preliminary cleaning and verification of the collected data and removal of obviously erroneous data points, information related to energy consumption is obtained;

[0016] Furthermore, the energy prediction module processes the energy consumption related information, and the specific process of obtaining the energy prediction information is as follows;

[0017] First, data feature extraction is performed to extract multiple features from the collected energy consumption related information, including time features, equipment operation status features, and environmental features;

[0018] Use wavelet transform and Fourier transform to conduct time-frequency analysis and mine the hierarchical features of energy consumption data, including periodicity and trend;

[0019] After that, model selection and training are performed, using deep learning models, including long short-term memory networks (LSTM) or gated recurrent units (GRU);

[0020] After that, historical energy consumption data is collected and divided into training set, validation set and test set according to the preset ratio;

[0021] The model is trained using the training set, and the weight parameters of the model are continuously adjusted through the back-propagation algorithm. Hyperparameters are tuned on the validation set to ensure that the model has good generalization ability. Finally, the prediction accuracy of the model is evaluated on the test set, and the prediction error is required to be within an acceptable range.

[0022] Finally, real-time prediction and updating are performed. As new energy consumption-related information is collected, it is input into the trained prediction model in real time. The prediction model quickly outputs energy forecast information for a period of time in the future. At the same time, the model is incrementally trained regularly using newly accumulated data.

[0023] Furthermore, the energy consumption analysis module performs energy consumption analysis, and the specific process of obtaining energy consumption analysis information is as follows:

[0024] The energy consumption analysis process includes energy consumption composition analysis and energy consumption trend analysis;

[0025] The energy consumption composition analysis includes data classification and sorting, visualization table making, and correlation analysis implementation;

[0026] The energy consumption trend analysis includes time series processing, trend curve drawing, factor analysis, energy consumption baseline setting, and anomaly monitoring.

[0027] Furthermore, the specific process of the energy consumption composition analysis is as follows:

[0028] Data classification and sorting: Extract all the original energy consumption data within a preset time period (such as the past month) from the database, including the timestamp, energy type identifier, equipment number, and the corresponding energy consumption value;

[0029] Conduct the first round of classification according to the energy type, and classify the data into different energy data subsets of electricity, water, and gas respectively;

[0030] Within each energy subset, further subdivide according to the equipment category;

[0031] Finally, form a data structure according to the regional dimension;

[0032] Visualization table making: Visualize the energy consumption composition ratio in the form of a table;

[0033] Correlation analysis implementation: Select the associated equipment group in the preset area as the analysis object, and collect the operation status data of the associated equipment and the corresponding total energy consumption data within the preset time period;

[0034] Build a correlation analysis model, and use the Pearson correlation coefficient calculation method to measure the linear correlation between the start sequence of different equipment and the total energy consumption;

[0035] For each pair of equipment (equipment A and equipment B), synchronously traverse their operation status data in the time series, and observe the impact of the start status of equipment B on the total energy consumption within a certain time window after equipment A is started;

[0036] Specifically in the calculation, assume that the start time series of equipment A is these, and the start time series of equipment B is T B , and the total energy consumption series is E. For each time point t, when t ∈ T A and t + αt ∈ T B (Δt is the set time window), extract the corresponding energy consumption value E t , and calculate the Pearson correlation coefficient r between the start sequence of equipment A and B and these energy consumption values AB，E , and the formula is:

[0037]

[0038] Furthermore, the specific process of the energy consumption trend analysis is as follows:

[0039] First, perform time series processing: Retrieve historical data from the database storing energy consumption data, and select the corresponding time for summarization according to the required speed of energy consumption change.

[0040] After that, process the summarized data, remove short-term noise using moving average or exponential smoothing methods to highlight the long-term trend.

[0041] Perform trend curve plotting and factor analysis: Plot the trend of the processed data, with time (month or year) as the horizontal axis and the processed energy consumption value as the vertical axis, and use a plotting tool to draw a line chart.

[0042] Comprehensively consider the impact of external factors on energy consumption. On the one hand, collect production adjustment information within the enterprise, compare it with the energy consumption trend chart, and if the energy consumption curve shows a significant upward trend during the launch period of new products, record it to clarify the direct impact of production activities on energy consumption.

[0043] On the other hand, collect meteorological data and energy market price fluctuation information, and conduct integrated analysis with the energy consumption change situation.

[0044] Set the energy consumption baseline and monitor for anomalies: The energy consumption baseline is set based on the historical average energy consumption. Query data for the past year from the database and calculate the average energy consumption value as the baseline; if referring to the industry standard energy consumption value, obtain the standard energy consumption index matching the enterprise scale and production type from the industry database or specification documents as the baseline.

[0045] In the daily energy consumption monitoring process, for each newly obtained energy consumption data point, calculate its deviation ratio from the baseline according to the formula D = (E1 - E2) / E2 * 100, where E1 is the current actual energy consumption value and E2 is the baseline energy consumption value.

[0046] When D exceeds the preset range, it indicates an anomaly.

[0047] Furthermore, the specific process of generating the energy-saving optimization suggestions is as follows:

[0048] First, construct an energy-saving strategy library: Collect and organize energy-saving strategies, including equipment operation optimization strategies, energy substitution strategies, equipment upgrade strategies, and management measure optimization strategies.

[0049] Quantitatively evaluate each energy-saving strategy, establish the corresponding relationship between the strategy and the expected energy-saving effect, and store it in the energy-saving strategy library.

[0050] After that, perform intelligent recommendation:

[0051] Based on the results of the energy consumption analysis module, combined with the energy-saving strategy library, use intelligent algorithms including rule-based reasoning and collaborative filtering, etc., to generate corresponding energy-saving optimization suggestions for different scenarios.

[0052] Present the generated energy-saving optimization suggestions in an intuitive report form, including text descriptions, data comparisons, and implementation steps;

[0053] Furthermore, the process of comprehensive management by the comprehensive management module is as follows:

[0054] User permission management:

[0055] Establish a multi-role user system, including system administrators, energy managers, and ordinary users;

[0056] Data storage and management:

[0057] The Xi'an database management system, including a distributed relational database or a time series database, stores energy consumption data, prediction results, and analysis reports;

[0058] Classify and store the data, establish a reasonable data table structure, and improve the data query efficiency through technologies such as index optimization.

[0059] Regularly perform data backups in a combination of full backups and incremental backups, and back up the data to local storage devices and cloud storage;

[0060] At the same time, formulate a data archiving strategy and archive historical data according to preset rules.

[0061] System integration and expansion:

[0062] Provide open interfaces to integrate the platform with other existing enterprise management systems.

[0063] The present invention has the following advantages compared with the prior art: This energy efficiency prediction and analysis platform based on artificial intelligence can achieve accurate energy prediction. With the help of a deep learning model, it fully considers the time series characteristics of energy consumption data, accurately captures the dynamic change law of energy consumption, and has a high accuracy in predicting future energy demand. Through rigorous data feature extraction, covering multi-dimensional features such as time, equipment operation status, and environment, and using wavelet transform and Fourier transform to mine deep information, it provides a solid foundation for model training, enabling the prediction error to be controlled within an acceptable range;

[0064] At the same time, regular incremental training enables the model to adapt to changes in the energy consumption pattern and always fit the actual situation. Whether it is dealing with short-term production adjustments or long-term energy consumption trend changes, it can provide accurate forward-looking guidance on energy use for enterprises.

[0065] Classify and statistically analyze energy consumption data from multiple dimensions of energy type, equipment category, and region, draw visual charts to intuitively present the energy consumption ratio, accurately locate major energy consumers, and clarify the key directions for energy-saving transformation. Conduct correlation analysis to deeply explore the co-variation of energy consumption among equipment and regions;

[0066] The energy consumption trend analysis is scientific and effective. Use time series methods such as moving average and exponential smoothing to process historical energy consumption data, and draw trend curves to clearly show the increasing and decreasing trends of energy consumption. Combine internal and external factors within the enterprise for comprehensive analysis, deeply explain the reasons for energy consumption changes, set a reasonable energy consumption baseline, monitor deviations in real time, and promptly detect abnormal energy consumption periods or regions, providing a precise entry point for energy-saving optimization to ensure that the enterprise's energy consumption is always under control.

[0067] The energy-saving strategy library is rich and practical, covering energy-saving strategies in multiple fields such as equipment operation, energy substitution, equipment upgrading, and management measures, and quantitatively evaluating the expected effects. Enterprises can quickly match applicable strategies according to their own energy consumption situations to achieve personalized customization of energy-saving solutions.

[0068] The intelligent recommendation system is efficient and convenient. Based on the energy consumption analysis results and the strategy library, use intelligent algorithms to generate targeted energy-saving optimization suggestions for different scenarios, presented in an intuitive report, which is convenient for management personnel to understand and execute, greatly shortening the cycle from the formulation to the implementation of the energy-saving plan, and accelerating the process of the enterprise's energy-saving and efficiency-increasing. Brief Description of the Drawings

[0069] Figure 1 is the structural block diagram of the present invention. Detailed Embodiment

[0070] The following will make a detailed description of the embodiments of the present invention. These embodiments are implemented on the premise of the technical solution of the present invention, and detailed implementation manners and specific operation processes are given. However, the protection scope of the present invention is not limited to the following embodiments.

[0071] As Figure 1 shown, this embodiment provides a technical solution: an energy efficiency prediction and analysis platform based on artificial intelligence, including an energy consumption collection module, an energy efficiency prediction module, an energy consumption analysis module, an energy-saving optimization suggestion module, and a comprehensive management module;

[0072] The energy consumption collection module is used to collect information related to energy consumption;

[0073] The energy prediction module is used to process information related to energy consumption to obtain energy prediction information;

[0074] The energy consumption analysis module is used to conduct energy consumption analysis to obtain energy consumption analysis information;

[0075] The energy-saving optimization suggestion module is used to conduct energy-saving optimization analysis and generate energy-saving optimization suggestions;

[0076] The comprehensive management module is used to conduct overall comprehensive management.

[0077] The specific process of the energy consumption collection module collecting energy consumption-related information is as follows:

[0078] Connect the data collection device, set up various types of sensor interfaces to connect to energy collection devices, including power sensors, water meters, and gas meter sensors, to meet the collection requirements of different energy consumption data. These sensors can real-time monitor key parameters such as the flow, power, and pressure of energy use, and transmit the data to the collection module through wired (such as RS485, Ethernet) or wireless (such as Wi-Fi, Bluetooth, LoRa) transmission methods.

[0079] For high-power equipment in large industrial scenarios, use dedicated smart meters with high-precision measurement and data remote transmission functions to accurately collect real-time electricity consumption parameters of the equipment, such as three-phase voltage, current, active power, reactive power, etc.

[0080] Then set the data collection frequency. According to the stability of energy consumption and the requirements of analysis accuracy, set different collection frequencies, and collect data according to the set collection frequency; for relatively stable energy consumption sources, such as building lighting electricity, general office equipment electricity, etc., it can be set to collect data once every 15 minutes; for energy consumption with large fluctuations, such as large machinery and heat treatment equipment in industrial production, the collection frequency is increased to once a minute or even higher to ensure capturing the dynamic changes of energy consumption.

[0081] Data preprocessing: After initially cleaning and verifying the collected data and removing obvious error data points, obtain energy consumption-related information, such as abnormal high or low values caused by sensor failures. Through data verification algorithms, compare the data change ranges of adjacent collection cycles. If it exceeds a reasonable threshold, mark the data point as suspicious data, and correct or supplement it in combination with statistical features such as the average value and median of historical data.

[0082] Furthermore, the specific process of the energy prediction module processing energy consumption-related information to obtain energy prediction information is as follows;

[0083] First, extract data features, extract various features from the collected energy consumption-related information, including time features (such as hours, dates, seasons, etc.), equipment operation status features (equipment startup duration, shutdown duration, load rate, etc.), and environmental features (temperature, humidity, etc., because environmental factors may affect the efficiency of some energy consumption equipment, such as air conditioning systems);

[0084] Wavelet transform and Fourier transform are used to perform time-frequency analysis to explore the hierarchical characteristics of energy consumption data, including periodicity and trend, providing rich information for subsequent model training.

[0085] After that, model selection and training are carried out, using deep learning models, including long short-term memory networks (LSTM) or gated recurrent units (GRU). Considering the time series characteristics of energy consumption data, these models can effectively capture the dynamic changes of data and make accurate predictions on future energy consumption.

[0086] After that, historical energy consumption data is collected and divided into training set, validation set and test set according to the preset ratio;

[0087] The model is trained using the training set, and the weight parameters of the model are continuously adjusted through the back-propagation algorithm. Hyperparameters are tuned on the validation set to ensure that the model has good generalization capabilities. Finally, the prediction accuracy of the model is evaluated on the test set, requiring the prediction error to be within an acceptable range; for example, the mean absolute error (MAE) is controlled below a certain value.

[0088] Finally, real-time prediction and update are performed. As new energy consumption-related information is collected, it is input into the trained prediction model in real time. The prediction model quickly outputs energy forecast information for a period of time in the future (such as the next hour, the next day, etc., set according to actual needs). At the same time, the model is incrementally trained regularly (such as weekly or monthly) using newly accumulated data to adapt to potential changes in energy consumption patterns and continuously improve prediction accuracy.

[0089] The specific process of the energy consumption analysis module performing energy consumption analysis and obtaining energy consumption analysis information is as follows:

[0090] The energy consumption analysis process includes energy consumption composition analysis and energy consumption trend analysis;

[0091] Energy consumption composition analysis includes data classification and sorting, visualization table production and correlation analysis implementation;

[0092] Energy consumption trend analysis includes time series processing, trend curve drawing and factor analysis, energy consumption baseline setting and abnormality monitoring.

[0093] Furthermore, the specific process of the energy consumption composition analysis is as follows:

[0094] Data classification and sorting: extract all raw energy consumption data within a preset period of time (such as the past month) from the database, including timestamp, energy type identification, equipment number and corresponding energy consumption value;

[0095] For the first round of classification according to energy types, the data are classified into different energy data subsets of electricity, water, and gas respectively;

[0096] Within each energy subset, further subdivision is carried out according to equipment categories; taking the electricity subset as an example, the equipment management table is associated according to the equipment number, and the energy consumption data of different categories of equipment such as office equipment, production equipment, and lighting systems are identified and stored in different data tables or data frames respectively for subsequent analysis. The same method is applied to the classification of water and gas energy consumption data by equipment.

[0097] Finally, according to the regional dimension, a data structure is formed. For example, through the equipment installation location information or the regional identification field, the energy consumption data of various types of equipment are divided into different regional groups such as different floors, workshops, and departments, forming a multi-level nested data structure, such as taking the region as the top-level directory, with each subordinate energy type and then further subdivided into equipment categories.

[0098] Visualization table production: The production of the energy consumption composition ratio is visualized as a table; taking the example of drawing a pie chart to show the proportion of each energy type in the total energy consumption. First, calculate the total energy consumption of each energy type within the selected time period. Assume that the total energy consumption data is stored in a data table named "total_energy_consumption", the energy type field is "energy_type", and the energy consumption value field is "consumption_value". The SQL statement used is as follows;

[0099] "SELECT energy_type, SUM(consumption_value) AS total_consumption FROM total_energy_consumption GROUP BY energy_type"

[0100] The statistical results of the total energy consumption of each energy type are obtained.

[0101] Import the above statistical results into a plotting tool (such as the matplotlib library in Python or a professional business intelligence software), use the energy type as the sector label of the pie chart, and the corresponding proportion of the total energy consumption as the sector area ratio to draw a clear and intuitive pie chart, so that users can clearly see the weight distribution of electricity, water, gas, etc. in the overall energy consumption at a glance.

[0102] When drawing a bar chart to show the energy consumption comparison of different equipment categories or regions, similarly, first summarize the energy consumption data of each equipment category or region. For the bar chart of equipment categories, group and count the energy consumption by equipment category. The statement used is as follows:

[0103] The statement “SELECT equipment_category, SUM(consumption_value) AS category_consumption FROM equipment_energy_data GROUP BY equipment_category” is used, and then a bar chart is plotted with the equipment category on the horizontal axis and the energy consumption value on the vertical axis to highlight the differences in energy consumption among different equipment categories and identify the major energy consumers. For example, the bar charts corresponding to several high-energy-consuming machines among the production equipment will be significantly higher than those of other equipment.

[0104] Association analysis implementation: Select the associated equipment group in the preset area as the analysis object, and collect the operation status data of the associated equipment (including start time, stop time, operating power, etc.) and the corresponding total energy consumption data within the preset time period; the data storage format can be a time series data table, and each row record contains a timestamp, the status values of each equipment, and the total energy consumption value.

[0105] Build an association analysis model and use the Pearson correlation coefficient calculation method to measure the linear correlation between the start order of different equipment and the total energy consumption;

[0106] For each pair of equipment (equipment A and equipment B), traverse their operation status data synchronously in the time series, and observe the impact of the start status of equipment B on the total energy consumption within a certain time window (such as 10 minutes) after equipment A is started;

[0107] Specifically in the calculation, assume that the start time series of equipment A is these, and the start time series of equipment B is T B , and the total energy consumption series is E. For each time point t, when t ∈ T A and t + Δt ∈ T B (Δt is the set time window), extract the corresponding energy consumption value E t , and calculate the Pearson correlation coefficient r between the start order of equipment A and B and these energy consumption values AB，E , and the formula is:

[0108]

[0109] If equipment B is started within 10 minutes after equipment A is started, X AB，t is recorded as 1, and if it is not started, it is recorded as 0. is the average energy consumption, is the average value of the start order variable. After calculating the correlation coefficient, judge the strength of the correlation by looking at the absolute value. Close to 1 is a strong positive correlation, close to -1 is a strong negative correlation, and close to 0 is a weak correlation. Once it is detected that the correlation of a certain pair of equipment is abnormal and there is obvious energy consumption waste, mark it immediately for subsequent optimization.

[0110] The specific process of the energy consumption trend analysis is as follows:

[0111] First, perform time series processing: Retrieve historical data from the database storing energy consumption data. According to the required speed of energy consumption change, select the corresponding time for summarization. If you want to see short-term changes, summarize by day; if you want to consider long-term trends, summarize by month. Taking monthly summarization as an example, using the instruction "SELECT MONTH(date) AS month, SUM(consumption_value) AS monthly_consumption FROM energy_data GROUP BY MONTH(date)", the data can be grouped together by month.

[0112] After that, process the summarized data. Use the moving average or exponential smoothing method to remove short-term noise and highlight the long-term trend;

[0113] For example, for simple moving average, set a window size, such as 3 months. The moving average value MA_t at each time point t is calculated as follows: where E_i is the energy consumption value at time point i. Exponential smoothing is similar. Give a smoothing coefficient, such as 0.2, and smooth the data step by step according to the formula.

[0114] Conduct trend curve plotting and factor analysis: Plot the trend of the processed data. Use time (month or year) as the horizontal axis and the processed energy consumption value as the vertical axis. Use a plotting tool to draw a line chart, so that it can be clearly seen whether the energy consumption is rising, falling or remaining stable.

[0115] Comprehensively consider the impact of external factors on energy consumption. On the one hand, collect production adjustment information within the enterprise, such as records during periods of new product launch, production line capacity expansion, etc. Compare it with the energy consumption trend chart. If the energy consumption curve shows a significant upward trend during the new product launch period, record it to clarify the direct impact of production activities on energy consumption;

[0116] On the other hand, collect meteorological data (such as temperature and humidity in different seasons) and energy market price fluctuation information, and conduct integrated analysis with the energy consumption change situation. For example, the increase in summer temperature causes a significant increase in air conditioner power consumption, driving up electricity energy consumption, which is reflected in the trend chart; the increase in energy prices prompts enterprises to reduce the output of high-energy-consuming links, thereby reducing energy consumption. These causal relationships should be recorded in detail to explain the internal reasons for energy consumption changes.

[0117] Setting the energy consumption baseline and abnormal monitoring: The energy consumption baseline is set based on the historical energy consumption average value. Query data for the past year from the database and calculate the average energy consumption value as the baseline; if referring to the industry standard energy consumption value, obtain the standard energy consumption index matching the enterprise scale and production type from the industry database or specification documents as the baseline;

[0118] In the daily energy consumption monitoring process, for each newly obtained energy consumption data point, calculate its deviation ratio from the baseline according to the formula D = (E1 - E2) / E2 * 100, where E1 is the current actual energy consumption value and E2 is the baseline energy consumption value;

[0119] When D exceeds the preset range, it indicates an anomaly.

[0120] Set a reasonable deviation threshold, such as between -20% and 20%. Once the deviation exceeds the threshold, the system immediately triggers an alarm, marks the time period or area with abnormal energy consumption, and provides a basis for accurately formulating energy-saving optimization measures.

[0121] The specific process of generating the energy-saving optimization suggestions is as follows:

[0122] First, construct an energy-saving strategy library: collect and organize energy-saving strategies, including equipment operation optimization strategies (such as reasonably adjusting the start and stop times of equipment, optimizing the equipment load rate, and dynamically adjusting the cooling capacity of the air conditioning system according to the personnel distribution), energy substitution strategies (replacing some traditional energy sources with clean energy sources such as solar energy and wind energy when conditions permit), equipment upgrade strategies (recommending to replace with new equipment with higher energy efficiency), and management measure optimization strategies (such as implementing energy quota management and conducting energy-saving training to improve employees' energy-saving awareness);

[0123] Quantitatively evaluate each energy-saving strategy, establish the corresponding relationship between the strategy and the expected energy-saving effect (such as energy-saving percentage, cost savings amount, etc.), and store it in the energy-saving strategy library for quickly matching applicable strategies according to the specific energy consumption situation.

[0124] After that, perform intelligent recommendation:

[0125] Based on the results of the energy consumption analysis module, combined with the energy-saving strategy library, use intelligent algorithms including rule-based reasoning and collaborative filtering, etc., to generate corresponding energy-saving optimization suggestions for different scenarios. For example, for the situation where the energy consumption of a certain workshop is relatively high and the equipment idle time is relatively long, recommend the equipment start-stop optimization strategy, and elaborate on how to reasonably arrange the equipment startup according to the production schedule and the expected energy-saving effect.

[0126] Present the generated energy-saving optimization suggestions in an intuitive report form, including text descriptions, data comparisons, and implementation steps, etc., to facilitate the understanding and execution by management personnel.

[0127] The process of comprehensive management by the comprehensive management module is as follows:

[0128] User permission management:

[0129] Establish a multi-role user system, including system administrators, energy management personnel, and ordinary users;

[0130] Different permissions are assigned to each role. The system administrator has the highest permissions and can perform operations such as system configuration, user management, data backup and recovery, etc.; energy management personnel are responsible for the daily monitoring and analysis of energy data, as well as the formulation and implementation of energy-saving optimization plans; ordinary users can only view the energy consumption information related to themselves, such as the energy consumption data of their departments.

[0131] Ensure the legality of user identities through methods such as username and password verification and permission tokens, prevent illegal access and data tampering, and ensure the safe and stable operation of the system.

[0132] Data storage and management:

[0133] The Xi'an database management system, including a distributed relational database or a time series database, stores energy consumption data, prediction results, and analysis reports;

[0134] Classify and store the data, establish a reasonable data table structure, and improve the data query efficiency through technologies such as index optimization.

[0135] Regularly perform data backups in a combination of full backups and incremental backups, and back up the data to local storage devices and cloud storage to prevent data loss;

[0136] At the same time, formulate a data archiving strategy, archive historical data according to preset rules, release storage space, and ensure the efficient operation of the database.

[0137] System integration and expansion:

[0138] Provide open interfaces to integrate the platform with other existing enterprise management systems (such as enterprise resource planning ERP, manufacturing execution system MES, etc.), realize the interconnection and interoperability of energy data with production, financial, and other data, and provide more comprehensive data support for enterprise comprehensive decision-making.

[0139] Adopt a modular design concept to facilitate the expansion of subsequent functional modules. For example, according to the development of new energy technologies and the energy-saving needs of enterprises, add new energy power generation prediction modules, energy trading auxiliary decision-making modules, etc., to improve the adaptability and value of the platform.

[0140] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, such as two, three, etc., unless otherwise specifically and clearly defined.

[0141] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0142] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. An energy efficiency prediction and analysis platform based on artificial intelligence, characterized in that: It includes energy consumption collection module, energy efficiency prediction module, energy consumption analysis module, energy-saving optimization suggestion module and comprehensive management module; The energy consumption collection module is used to collect information related to energy consumption; The energy prediction module is used to process information related to energy consumption and obtain energy prediction information; The energy consumption analysis module is used to perform energy consumption analysis and obtain energy consumption analysis information; The energy-saving optimization suggestion module is used to perform energy-saving optimization analysis and generate energy-saving optimization suggestions; The comprehensive management module is used for overall integrated management.

2. The energy efficiency prediction and analysis platform based on artificial intelligence according to claim 1, characterized in that: The specific process of the energy consumption collection module collecting energy consumption related information is as follows: Connect data acquisition equipment and set up various types of sensor interfaces to access energy acquisition equipment, including power sensors, water meters and gas meter sensors; Then, the data collection frequency is set. Different collection frequencies are set according to the stability of energy consumption and the analysis accuracy requirements, and data is collected according to the set collection frequencies. Data preprocessing involves preliminary cleaning and verification of the collected data, removing obviously erroneous data points, and obtaining information related to energy consumption.

3. The energy efficiency prediction and analysis platform based on artificial intelligence according to claim 1, characterized in that: The energy prediction module processes the energy consumption related information, and the specific process of obtaining the energy prediction information is as follows; First, data feature extraction is performed to extract multiple features from the collected energy consumption related information, including time features, equipment operation status features, and environmental features; Use wavelet transform and Fourier transform to conduct time-frequency analysis and mine the hierarchical features of energy consumption data, including periodicity and trend; After that, model selection and training are performed, using deep learning models, including long short-term memory networks (LSTM) or gated recurrent units (GRU); After that, historical energy consumption data is collected and divided into training set, validation set and test set according to the preset ratio; The model is trained using the training set, and the weight parameters of the model are continuously adjusted through the back-propagation algorithm. Hyperparameters are tuned on the validation set to ensure that the model has good generalization ability. Finally, the prediction accuracy of the model is evaluated on the test set, and the prediction error is required to be within an acceptable range. Finally, real-time prediction and updating are performed. As new energy consumption-related information is collected, it is input into the trained prediction model in real time. The prediction model outputs energy prediction information for a preset period of time in the future. At the same time, the model is incrementally trained regularly using newly accumulated data.

4. The energy efficiency prediction and analysis platform based on artificial intelligence according to claim 1, characterized in that: The specific process of the energy consumption analysis module performing energy consumption analysis and obtaining energy consumption analysis information is as follows: The energy consumption analysis process includes energy consumption composition analysis and energy consumption trend analysis; Energy consumption composition analysis includes data classification and sorting, visualization table production and correlation analysis implementation; Energy consumption trend analysis includes time series processing, trend curve drawing and factor analysis, energy consumption baseline setting and abnormality monitoring.

5. The energy efficiency prediction and analysis platform based on artificial intelligence according to claim 3 is characterized by: The specific process of the energy consumption composition analysis is as follows: Data classification and sorting: extract all raw energy consumption data within a preset time period from the database, including timestamp, energy type identification, equipment number and corresponding energy consumption value; The first round of classification is carried out according to energy type, and the data are classified into different energy data subsets of electricity, water and gas; Within each energy subset, further subdivision is done based on equipment category; Finally, the data structure is formed according to the regional dimension; Visualization table production: Make a table to visualize the energy consumption composition ratio; Implementation of joint analysis: select the associated equipment group in the preset area as the analysis object, collect the operating status data of the associated equipment within the preset time and the corresponding total energy consumption data; A correlation analysis model was constructed, and the Pearson correlation coefficient calculation method was used to measure the linear correlation between the startup sequence of different devices and the total energy consumption; For each pair of devices (device A and device B), their operating status data is synchronously traversed in the time series. When device A is turned on, the impact of the on-state of device B on the total energy consumption is observed within a certain time window; In the specific calculation, it is assumed that the startup time sequence of backup device A is these, and the startup time sequence of device B is T B , the total energy consumption sequence is E, for each time point t, when t∈T A And t+Δt∈T B (Δt is the set time window), extract the corresponding energy consumption value E t , calculate the Pearson correlation coefficient r between the power-on sequence of devices A and B and these energy consumption values AB，E , the formula is:

6. The energy efficiency prediction and analysis platform based on artificial intelligence according to claim 3 is characterized by: The specific process of the energy consumption trend analysis is as follows: First, perform time series processing: extract historical data from the database that stores energy consumption data, and select the corresponding time for aggregation based on the required energy consumption change speed; Then process the aggregated data, use moving average or exponential smoothing method to remove short-term noise and highlight long-term trends; Draw trend curves and factor analysis: Draw a trend graph for the processed data, with time as the horizontal axis and the processed energy consumption value as the vertical axis, and use the drawing tool to draw a line graph; Comprehensively consider the impact of external factors on energy consumption. On the one hand, collect production adjustment information within the enterprise and compare it with the energy consumption trend chart. If the energy consumption curve rises significantly during the launch of new products, record it to clarify the direct impact of production activities on energy consumption; On the other hand, meteorological data and information on energy market price fluctuations are collected and integrated with energy consumption changes for analysis; Energy consumption baseline setting and abnormal monitoring: The energy consumption baseline is set based on the historical energy consumption average value. The data of the past year is queried from the database to calculate the average energy consumption value as the benchmark; if the industry standard energy consumption value is referenced, the standard energy consumption index that matches the enterprise scale and production type must be obtained from the industry database or specification documents as the benchmark; In the daily energy consumption monitoring process, each time a new energy consumption data point is obtained, the deviation ratio from the baseline is calculated according to the formula D = (E1-E2) / E2*100, where E1 is the current actual energy consumption value and E2 is the baseline energy consumption value; When D exceeds the preset range, it indicates that there is an abnormality.

7. The energy efficiency prediction and analysis platform based on artificial intelligence according to claim 1, characterized in that: The specific process of generating the energy-saving optimization suggestion is as follows: First, build an energy-saving strategy library: collect and organize energy-saving strategies, including equipment operation optimization strategies, energy substitution strategies, equipment upgrade strategies, and management measures optimization strategies; Conduct quantitative evaluation on each energy-saving strategy, establish the corresponding relationship between the strategy and the expected energy-saving effect, and store it in the energy-saving strategy library; Then make intelligent recommendations: Based on the results of the energy consumption analysis module and combined with the energy-saving strategy library, intelligent algorithms including rule-based reasoning and collaborative filtering are used to generate corresponding energy-saving optimization suggestions for different scenarios; The generated energy-saving optimization suggestions are presented in the form of an intuitive report, including text descriptions, data comparisons and implementation steps.

8. The energy efficiency prediction and analysis platform based on artificial intelligence according to claim 1, characterized in that: The process of comprehensive management by the comprehensive management module is as follows: User Rights Management: Establish a multi-role user system, including system administrators, energy managers, and ordinary users; Data storage and management: Xi'an database management system, including distributed relational database or time series database, stores energy consumption data, forecast results and analysis reports; Categorize and store data, establish a reasonable data table structure, and improve data query efficiency through index optimization and other technologies; Regularly perform data backup by combining full backup and incremental backup to back up data to local storage devices and cloud storage; At the same time, formulate data archiving strategies and archive historical data according to preset rules; System integration and expansion: Provide open interfaces to integrate the platform with other existing management systems of the enterprise.

Images