Industrial and commercial power utilization energy efficiency evaluation method and system based on multi-modal data

Abstract

The invention discloses an industrial and commercial power utilization energy efficiency assessment method and system based on multi-modal data, and relates to the technical field of power utilization energy efficiency assessment, and the method comprises the steps: obtaining a local timestamp and a park reference timestamp of a park management server; calculating a time offset according to the local timestamp and the park reference timestamp; generating batch data packets according to the time offset; and the batch data packet is sent to the park management server, and the park management server is used for performing power utilization energy efficiency evaluation according to the batch data packet and generating a power utilization energy efficiency evaluation report. According to the method, the evaluation report can be generated in combination with the time offset, so that power utilization energy efficiency evaluation is realized, and the accuracy and the reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of electricity energy efficiency assessment technology, and in particular to a method and system for assessing industrial and commercial electricity energy efficiency based on multimodal data. Background Technology

[0002] In industrial parks, existing systems integrate data from multiple sources, including smart meters, various sensors, and enterprise resource planning (ERP) platforms, to accurately assess electricity efficiency. However, when data acquisition terminals within the park, especially newly deployed equipment, are configured to synchronize using different network time services (e.g., an internal time server within the park versus a time server on the public internet), this configuration discrepancy leads to small but persistent time skews between different data streams. These skews prevent energy consumption data from accurately matching specific production activities, affecting the accuracy of energy attribution, carbon emission calculations, and hindering cross-enterprise collaborative energy efficiency optimization. Existing systems struggle to identify and correct these multiple time-sourced electricity consumption data, impacting assessment accuracy and reliability.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The main objective of this invention is to propose a method and system for evaluating the energy efficiency of industrial and commercial electricity consumption based on multimodal data. This method can generate an evaluation report by combining time offsets to achieve energy efficiency evaluation, thereby improving accuracy and reliability.

[0005] On one hand, embodiments of the present invention provide a method for evaluating the energy efficiency of industrial and commercial electricity consumption based on multimodal data, including the following steps:

[0006] Obtain the local timestamp and the park reference timestamp from the park management server;

[0007] Calculate the time offset based on the local timestamp and the park reference timestamp;

[0008] Generate batch data packets based on the time offset;

[0009] The batch data packets are sent to the park management server, which then performs an energy efficiency assessment based on the batch data packets and generates an energy efficiency assessment report.

[0010] On the other hand, embodiments of the present invention provide an industrial and commercial power efficiency assessment system based on multimodal data, comprising:

[0011] The data acquisition module is used to acquire the local timestamp and the park reference timestamp from the park management server;

[0012] The offset calculation module is used to calculate the time offset based on the local timestamp and the park reference timestamp;

[0013] A data packet generation module is used to generate batch data packets based on the time offset.

[0014] The energy efficiency assessment module is used to send the batch data packets to the park management server, and the park management server is used to perform an energy efficiency assessment based on the batch data packets and generate an energy efficiency assessment report.

[0015] The embodiments of this application include at least the following beneficial effects: The embodiments of this application first obtain the local timestamp and the park reference timestamp of the park management server, then calculate the time offset based on the local timestamp and the park reference timestamp, then generate batch data packets based on the time offset, and finally send the batch data packets to the park management server. The park management server performs power consumption energy efficiency assessment and generates a power consumption energy efficiency assessment report. Thus, the assessment report can be generated by combining the time offset to realize power consumption energy efficiency assessment, thereby improving accuracy and reliability.

[0016] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description and the drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0018] Figure 1 This is a flowchart illustrating a method for evaluating the energy efficiency of industrial and commercial electricity consumption based on multimodal data, as described in an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of the structure of an industrial and commercial power energy efficiency assessment system based on multimodal data, according to an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments.

[0021] In related technologies, within industrial parks, existing systems integrate data from multiple sources, including smart meters, various sensors, and enterprise resource planning platforms, to accurately assess electricity efficiency. However, when data acquisition terminals within the park, especially newly deployed equipment, are configured to synchronize using different network time services (e.g., an internal time server within the park versus a time server on the public internet), this configuration difference leads to small but persistent time discrepancies between different data streams. These discrepancies prevent energy consumption data from accurately matching specific production activities, affecting the accuracy of energy consumption attribution, carbon emission calculations, and hindering cross-enterprise collaborative energy efficiency optimization. Existing systems struggle to identify and correct these concurrent electricity consumption data from multiple time sources, impacting assessment accuracy and reliability.

[0022] For example, in a typical industrial park, an advanced industrial and commercial electricity efficiency assessment system has been deployed and put into operation. The core function of this system is to integrate data from multiple sources, including real-time metering data from smart meters, operational status and environmental data from various sensors on the production line, and production orders and material information from the Enterprise Resource Planning (ERP) platform. To ensure accurate correlation of this heterogeneous data during energy efficiency assessment, the system deploys a primary Network Time Server (NTP) within the park as a unified time reference, assuming that all data collection points—whether smart meters, production sensors, or the ERP system—precisely align their clocks to this internal reference. This unified time reference is considered a fundamental prerequisite for achieving refined energy efficiency analysis. For instance, only when instantaneous power data precisely matches the timestamp of a specific production event can the system accurately attribute energy consumption, thereby identifying high-energy-consuming processes or potential equipment malfunctions.

[0023] However, as the industrial park's production scale gradually expands, several new production areas have been built within the park to meet the growing production demands, and a large number of new high-energy-consuming equipment have been introduced. The deployment and commissioning of these new devices are often handled by different project teams or external contractors, and are usually accompanied by strict deadlines. Against this backdrop, data acquisition terminals in these new areas, such as newly added smart meters, process control sensors, and environmental monitoring equipment, may be configured to use an NTP server on the public internet for time synchronization, rather than strictly connecting to the main network time server deployed within the park, due to reasons such as rapid deployment, ease of operation, or a lack of in-depth understanding of the park's overall IT architecture. This practice is not uncommon in large, decentralized industrial environments. For example, equipment vendors may default to connecting their products to public NTP, or the internal NTP access policy may not have been fully established during the initial construction of the new area's network infrastructure. This seemingly minor configuration difference inadvertently introduces a second independent time base into the park's data acquisition ecosystem, breaking the original system design's assumption of "a single time base shared throughout the entire park."

[0024] While both public internet NTP servers and internal campus NTP servers strive to provide accurate time services, a small but persistent systematic time discrepancy typically exists between them due to factors such as their respective network environments, server hierarchies, and network transmission latency. For example, the time of a public NTP server may consistently be tens of milliseconds ahead of the internal NTP server. This means that the timestamps of data recorded by smart meters and sensors in newly built areas are offset from the data in the original core areas from the source. This offset is often very subtle, insufficient to trigger regular system alarms, and difficult to detect during routine data checks. However, when the system attempts to merge data from different areas, this time offset causes the data from different production areas to be unable to achieve precise alignment on the timeline.

[0025] In industrial parks, multi-source data acquisition terminals use different network time bases, resulting in systematic time offsets. It is necessary to identify, quantify, and correct these time offsets to ensure accurate time consistency of various types of electricity consumption data during fusion. Based on this, reliable sharing and collaborative optimization of energy consumption and carbon emission data across enterprises can be achieved.

[0026] The embodiments of this application will be explained in detail below with reference to the accompanying drawings:

[0027] Figure 1 This is an optional flowchart of a method for evaluating the energy efficiency of industrial and commercial electricity consumption based on multimodal data, provided in an embodiment of this application. Figure 1The method may include, but is not limited to, steps S101 to S104.

[0028] Step S101: Obtain the local timestamp and the park reference timestamp from the park management server;

[0029] Step S102: Calculate the time offset based on the local timestamp and the park reference timestamp;

[0030] Step S103: Generate batch data packets based on the time offset;

[0031] Step S104: Send the batch data packet to the park management server. The park management server is used to perform an energy efficiency assessment based on the batch data packet and generate an energy efficiency assessment report.

[0032] Steps S101 to S104 shown in the embodiments of this application can generate an evaluation report by combining time offset, so as to realize the evaluation of power energy efficiency and improve accuracy and reliability.

[0033] In some embodiments, steps S101-S104 may involve first acquiring the local timestamp and the park reference timestamp from the park management server. The local timestamp can be automatically acquired through the clock module inside the data acquisition terminal; for example, a smart meter records the current local time while recording electricity consumption data. The park reference timestamp can be provided by the park management server via the Network Time Protocol (NTP) service, and the data acquisition terminal can periodically request it from the park management server. It is understood that the local timestamp refers to the current time information recorded by the data acquisition terminal or local device, reflecting the local time state at the time of data acquisition. The park reference timestamp refers to the standard time information provided by the park management server, serving as a unified time reference for the entire park.

[0034] Then, the time offset is calculated based on the local timestamp and the park reference timestamp. The initial time offset can be obtained by subtracting the park reference timestamp from the local timestamp. To improve the accuracy of the calculation, multiple measurements can be taken and the average value calculated. It can be understood that the time offset refers to the difference between the local timestamp and the park reference timestamp; it quantifies the deviation between the local device time and the park's standard time.

[0035] Then, based on the time offset, batch data packets are generated. Electricity consumption data, local timestamps, and calculated time offsets can be integrated to generate batch data packets. For example, electricity consumption data collected over a period of time can be packaged with the corresponding local timestamps and time offsets into a single data structure, forming a batch of data. In essence, a batch data packet is a collection that encapsulates information such as electricity consumption data, timestamps, and time offsets, aiming to achieve batch transmission and unified processing of data.

[0036] Finally, the batch data packets are sent to the park management server. The park management server uses these packets to perform an energy efficiency assessment and generate an energy efficiency assessment report. Upon receiving the batch data packets, the park management server calibrates its local timestamp based on the time offset contained within, thus obtaining a unified timestamp. Based on the energy consumption data under this unified timestamp, the park management server can perform energy efficiency assessments and generate reports. For example, the park management server can analyze the energy load curves of different production units over a specific time period, combine this with production mode information, calculate energy efficiency indicators and energy consumption attributions, and ultimately generate a detailed assessment report.

[0037] This embodiment effectively solves the time deviation problem caused by multiple time sources in traditional industrial and commercial park electricity efficiency assessments by introducing time offset calculation and batch data packet generation at the data acquisition end. First, by acquiring the local timestamp and the park's reference timestamp, the difference between the local equipment and the park's standard time is quantified. Then, based on the calculated time offset, the electricity consumption data, local timestamp, and time offset are encapsulated into batch data packets, ensuring the integrity of data transmission and the synchronization of time information. Finally, after receiving these batch data packets, the park management server can use the time offset to accurately calibrate the local timestamp, thereby obtaining electricity consumption data under a unified time stamp. Based on this calibrated data, the park management server can perform accurate electricity efficiency assessments and generate reliable assessment reports. The entire process forms a closed loop, solving the time synchronization problem from the data acquisition source, ensuring the accuracy and effectiveness of subsequent energy efficiency assessments, and thus providing a solid data foundation for park energy efficiency management.

[0038] By introducing local timestamps, park reference timestamps, and time offset calculations, this embodiment addresses the problem at its source. By precisely calculating the time offset and integrating it into batch data packets, it ensures that the data contains crucial time synchronization information before being transmitted to the park management server. The park management server can then use this information to perform precise time calibration, thereby obtaining electricity consumption data under a unified time stamp. This mechanism enables precise matching of energy consumption data with specific production activities, significantly improving the accuracy of energy consumption attribution and carbon emission calculations, and providing a reliable data foundation for cross-enterprise energy efficiency collaboration and optimization. Therefore, this embodiment represents a breakthrough in solving the challenge of integrating multi-time-source data, bringing significant technological advancements to energy efficiency management in industrial and commercial parks.

[0039] In some embodiments, step S102, calculating the time offset based on the local timestamp and the park reference timestamp, may include, but is not limited to, the following steps:

[0040] Send network probe signals to the park management server to obtain network round-trip time information;

[0041] Calculate the degree of network communication fluctuation based on network round-trip time information;

[0042] Calculate the initial offset based on the local timestamp and the park reference timestamp;

[0043] If the fluctuation of network communication is less than the preset fluctuation threshold, the time offset is obtained by weighted averaging of multiple initial offsets.

[0044] If the fluctuation in network communication exceeds a preset fluctuation threshold, median filtering is applied to multiple initial offsets to obtain the time offset.

[0045] In some embodiments, due to potential fluctuations in the network communication environment, the calculated time offset may be inaccurate if the real-time network communication status is not considered, thus affecting the reliability of subsequent energy efficiency assessments. To address this, a network probe signal can be sent to the park management server to obtain network round-trip time information. Data packets can be sent to the park management server via a network protocol (e.g., Internet Control Message Protocol, ICMP) and their responses can be received to measure the round-trip time of the data packets transmitted over the network. This allows the acquisition of network round-trip time information, reflecting the network latency between the terminal device and the park management server. The network round-trip time information may include a series of continuously measured round-trip time values.

[0046] Then, based on the network round-trip time information, the degree of network communication fluctuation is calculated. This can be done through statistical analysis methods, such as calculating the standard deviation, variance, or mean absolute deviation of the round-trip time information, to quantify the stability of network latency. The higher the degree of network communication fluctuation, the more unstable the network, and vice versa. The preset fluctuation threshold is a configurable parameter used to distinguish between stable and unstable states of network communication.

[0047] Then, calculate the initial offset based on the local timestamp and the park's reference timestamp. This initial offset can be obtained by calculating the difference between the time recorded by the local device and the time provided by the park management server, without considering network fluctuations. To improve accuracy, multiple measurements can be performed and the average value can be used as the initial offset.

[0048] If the network communication fluctuation is less than a preset fluctuation threshold, it indicates that the network environment is relatively stable. In this case, a weighted average of multiple initial offsets can be performed to obtain the time offset. This effectively reduces the impact of random errors and makes the calculated time offset more accurate. Different weights can be assigned to different measurement results based on factors such as the confidence level or measurement time of each measurement, and then a weighted average can be performed. For example, the average fluctuation can be calculated as the preset fluctuation threshold by statistically analyzing historical network environment data.

[0049] If network communication fluctuations exceed a preset fluctuation threshold, it indicates significant fluctuations or anomalies in the network environment. In this case, median filtering can be applied to multiple initial offsets to obtain the time offset. This effectively suppresses the impact of outliers (i.e., extreme offsets caused by network jitter or packet loss) on the final result, improving the robustness of time offset calculation. Median filtering, by selecting the median in the data sequence as a representative value, effectively resists outlier interference.

[0050] This embodiment monitors network communication status in real time and adaptively selects different time offset calculation strategies based on the degree of network communication fluctuation. Specifically, when the network is stable, weighted averaging can fully utilize the accuracy advantage of multiple measurement data to obtain a more accurate offset; while when the network is unstable, median filtering can effectively avoid measurement errors caused by network anomalies, ensuring the robustness of offset calculation. Thus, regardless of changes in the network environment, high-precision and high-reliability time offsets can be obtained, providing a solid time synchronization foundation for subsequent power efficiency assessments.

[0051] To illustrate this technical solution more clearly, a specific example is used below. Assume the terminal device needs to synchronize its time with the park management server. The terminal device first sends a series of network probe signals to the park management server, such as sending 10 ICMP (Internet Control Message Protocol) packets and recording the round-trip time (RTT) of each packet. Based on these 10 RTT values, their standard deviation is calculated as the degree of network communication fluctuation. Simultaneously, the terminal device obtains its local timestamp and the park reference timestamp from the park management server, calculating 10 initial offsets. If the calculated RTT standard deviation is less than a preset fluctuation threshold (e.g., 5 milliseconds), the network is considered stable. In this case, the 10 initial offsets are weighted and averaged to obtain the final time offset. For example, the weights can be calculated based on the RTT value corresponding to each initial offset, with smaller RTT values ​​having greater weights. If the RTT standard deviation is greater than the preset fluctuation threshold, the network is considered unstable. In this case, the 10 initial offsets are filtered by median, i.e., the median value after sorting the 10 initial offsets is taken as the final time offset to eliminate extreme errors caused by network anomalies. In this way, regardless of network conditions, a more accurate and optimized time offset can be obtained.

[0052] Through the above technical solution, this embodiment can effectively solve the problems of insufficient accuracy and robustness in complex network environments. By introducing network detection and fluctuation assessment, and dynamically adjusting the calculation strategy based on the assessment results, this embodiment significantly improves the accuracy and stability of time offset calculation, thereby ensuring the reliability of industrial and commercial power efficiency assessment based on multimodal data and avoiding assessment deviations caused by time synchronization errors.

[0053] In some embodiments, step S103, generating batch data packets based on the time offset, may include, but is not limited to, the following steps:

[0054] Obtain electricity consumption data;

[0055] Based on a preset time interval, electricity consumption data, local timestamps, and time offsets are integrated to obtain batch data;

[0056] Perform hash calculations on the batch data to obtain a batch data digest;

[0057] Digitally sign the batch data digest to generate a digital signature certificate;

[0058] The batch data and digital signature credentials are encapsulated to obtain a batch data packet.

[0059] In some embodiments, if the generated batch data packets are not effectively guaranteed for data integrity and authenticity, there is a risk that the data may be tampered with or forged during transmission, thereby affecting the accuracy and reliability of subsequent electricity efficiency assessments. Therefore, electricity consumption data can be obtained first. Electricity consumption data can be collected in real time or periodically from various electrical devices, smart meters, or sensors within industrial and commercial parks; this data forms the basis for energy efficiency assessments.

[0060] Then, based on a preset time interval, the electricity consumption data, local timestamps, and time offsets are integrated to obtain batch data. The preset time interval refers to a fixed time period set by the system according to actual needs or data processing capabilities, such as every 5 minutes, 15 minutes, or 1 hour. Within this preset time interval, the acquired electricity consumption data, corresponding local timestamps, and calculated time offsets are logically combined and categorized to form structured batch data. The purpose is to organize the scattered raw data into units that are easy to transmit and process.

[0061] Next, hash the batch data to obtain a batch data digest. A cryptographic hash function (such as the SHA-256 algorithm) can be used to process the integrated batch data, generating a fixed-length, unique hash value. This hash value is the batch data digest, acting like a "fingerprint" for the data. Any small modification to the batch data will cause a significant change in the hash value, allowing for quick detection of data tampering. Finally, digitally sign the batch data digest to generate a digital signature credential. The sender's private key can be used to encrypt the batch data digest, generating a digital signature. This digital signature credential verifies the data's origin, ensuring that the data was generated by a legitimate sender and has not been tampered with after signing.

[0062] Finally, the batch data and digital signature certificate are encapsulated to obtain a batch data packet. The original batch data and the generated digital signature certificate can be packaged together to form a complete batch data packet. This batch data packet is used for subsequent data transmission.

[0063] This embodiment effectively addresses the integrity and authenticity risks that data may face during transmission by introducing hash calculation and digital signature mechanisms during the batch data packet generation process. Specifically, the generation of batch data digests ensures the consistency of batch data content before and after transmission; any unauthorized modifications will be revealed by hash value mismatches. Simultaneously, asymmetric encryption technology is used to digitally sign the batch data digests, allowing the recipient to verify the signature's validity using the sender's public key, thereby confirming that the data originates from the claimed sender and has not been tampered with after signing. Thus, the data integrity and authenticity of the batch data packets are reliably guaranteed before they are sent to the park management server.

[0064] To illustrate this technical solution more clearly, a specific example is used below. Assume a smart meter in an industrial park collects electricity consumption data every minute. For energy efficiency assessment, the system needs to send a batch data packet to the park management server every 15 minutes. Specifically, at the end of each 15-minute cycle, the system acquires all electricity consumption data from those 15 minutes and combines this data with the local timestamp of each data point and a pre-calculated time offset to form a batch data. This batch data is then input into a hash algorithm, such as SHA-256, to generate a unique batch data digest. Next, the local data processing unit uses its private key to digitally sign the batch data digest, generating a digital signature credential. Finally, the original batch data and the generated digital signature credential are encapsulated together to form a complete batch data packet, which is then sent to the park management server. When the park management server receives this batch data packet, it uses the sender's public key to verify the digital signature credential and independently calculates the hash value of the received batch data. Only when the signature verification is successful and the calculated hash value matches the hash value in the credential will the batch of data packets be considered valid and tamper-proof, and then used for subsequent energy efficiency assessment.

[0065] Through the above technical solution, the batch data packets generated in this embodiment have higher data integrity and authenticity. This significantly reduces the risk of data being tampered with or forged during transmission, ensuring that the electricity consumption data received by the park management server is accurate and reliable. Therefore, electricity efficiency assessments based on this reliable data will be more accurate and convincing, providing a solid data foundation and decision-making basis for the refined energy efficiency management of industrial and commercial parks.

[0066] In some embodiments, step S104, based on the batch data package, performs an energy efficiency assessment and generates an energy efficiency assessment report, which may include, but is not limited to, the following steps:

[0067] Step S201: Verify the batch data packets and obtain the verification results;

[0068] Step S202: If the verification result is successful, then perform time calibration according to the batch data packet to obtain a unified time stamp;

[0069] Step S203: Based on the unified time stamp, conduct an energy efficiency assessment and generate an energy efficiency assessment report.

[0070] In some embodiments, the integrity and accuracy of batch data packets may be affected by factors such as network transmission, equipment failure, or malicious attacks, leading to reduced reliability of the assessment results. Furthermore, time synchronization issues between different terminal devices and the park management server can also introduce assessment errors, affecting the accuracy of energy efficiency analysis. If these problems are not addressed, the generated electricity energy efficiency assessment report may not accurately reflect the actual energy efficiency of industrial and commercial electricity consumption, thus impacting the effectiveness of decision-making.

[0071] Therefore, batch data packets can be verified first to obtain verification results. This aims to ensure the integrity and authenticity of received batch data packets and prevent data from being tampered with or damaged during transmission. The verification results can indicate whether the data packets are valid and whether they have passed integrity checks.

[0072] If the verification result is successful, time calibration is performed based on the batch data package to obtain a unified timestamp. This aims to eliminate potential discrepancies between the local timestamp and the park's reference timestamp, ensuring that all electricity consumption data is analyzed under a unified time benchmark. A unified timestamp refers to a timestamp that is consistent throughout the entire park management system after calibration.

[0073] Then, based on a unified time stamp, an energy efficiency assessment is conducted, generating an energy efficiency assessment report. This assessment process will be based on high-quality, time-synchronized energy consumption data, thereby generating a more accurate and reliable energy efficiency assessment report.

[0074] This embodiment introduces a verification mechanism for batch data packets, effectively identifying and excluding data packets that may have been tampered with or corrupted during transmission, thus ensuring the authenticity and integrity of the data used for energy efficiency assessment from the source. Specifically, the verification step ensures that only data packets that meet preset security and integrity standards are used for subsequent processing. Furthermore, by calibrating the time information in the batch data packets and obtaining a unified time stamp, the time inconsistency problem caused by clock drift or network latency between different terminal devices and the park management server is resolved, ensuring that all electricity consumption data is analyzed under a unified time reference. Therefore, the data used for electricity energy efficiency assessment is not only reliable but also highly consistent in the time dimension, greatly improving the accuracy and credibility of the assessment results.

[0075] To illustrate this technical solution more clearly, a specific example is used below. Suppose a smart meter terminal in an industrial park periodically sends electricity consumption data to the park management server. During data transmission, due to network instability or potential malicious attacks, some data in batch data packets may be lost or tampered with. Upon receiving the batch data packets, the park management server will first verify them. For example, by verifying digital signature credentials. If the verification result shows that the data packet has been tampered with or is incomplete, the data packet will be marked as invalid, thus preventing inaccurate data from entering the subsequent evaluation process. If the verification is successful, but there is a slight drift between the smart meter terminal's internal clock and the park management server's park reference timestamp, causing a discrepancy between the local timestamp and the actual time, this embodiment will perform time calibration based on the information in the batch data packets. For example, by calculating and compensating for the time offset, the local timestamp will be adjusted to a unified time stamp. Only after the data packets are verified as valid and the time is calibrated to a unified benchmark will this electricity consumption data be used for the final energy efficiency assessment, thus ensuring the accuracy and reliability of the assessment results.

[0076] Through the above technical solution, this embodiment can significantly improve the accuracy and reliability of electricity energy efficiency assessment. First, the data packet verification mechanism effectively avoids assessment deviations caused by data corruption or tampering, ensuring the authenticity of the assessment data. Second, the time calibration step eliminates time synchronization errors between different devices, enabling the analysis of electricity consumption data under a unified time benchmark, which is crucial for accurately identifying electricity consumption patterns and calculating energy efficiency indicators. Finally, the electricity energy efficiency assessment report generated based on verified and time-calibrated data can more realistically and accurately reflect the actual energy efficiency of industrial and commercial electricity consumption, providing users with more valuable decision-making basis, thereby improving the overall level of energy management.

[0077] In some embodiments, step S201, verifying the batch data packets to obtain verification results, may include, but is not limited to, the following steps:

[0078] Based on the batch data packets, obtain the sender's public key, credential verification failure count, and failure time interval;

[0079] If the number of failed credential verifications exceeds the preset consecutive failure threshold and the time interval between failures is less than the preset pause verification time interval, then the verification result is determined to be an invalid data packet.

[0080] If the number of failed credential verifications is less than the preset consecutive failure threshold and the time interval between failures is greater than the preset pause verification time interval, then the digital signature credential in the batch data packet is decrypted according to the public key to obtain the first data digest.

[0081] A hash calculation is performed on the batch data in the batch data packet to obtain a second data digest;

[0082] If the first data digest and the second data digest are the same, the verification result is determined to be successful; otherwise, the verification result is determined to be unsuccessful.

[0083] In some embodiments, simple verification may not be effective in dealing with malicious attacks or persistent data transmission anomalies, thus affecting the accuracy of energy efficiency assessments and system security. Furthermore, it could lead to system resources being consumed by invalid data packets, or even generate erroneous energy efficiency assessment reports due to maliciously tampered data packets. To address this, the sender's public key, credential verification failure count, and failure time interval can be obtained based on batch data packets. The sender's public key is crucial information for subsequent decryption of digital signature credentials and is typically configured during system initialization or registration. The credential verification failure count records the number of consecutive verification failures by a specific sender within a given period, and the failure time interval records the time difference between the previous failure time and the current time.

[0084] If the number of failed credential verifications exceeds a preset consecutive failure threshold, and the time interval between failures is less than a preset pause verification time interval, the verification result is determined to be an invalid data packet. The preset consecutive failure threshold is a configurable parameter, for example, it can be set to 3 or 5 times, used to define the degree of consecutive failures, and can be configured according to the network congestion level. The preset pause verification time interval is also a configurable parameter, for example, it can be set to 5 minutes or 10 minutes, used to determine whether failures occur within a short period, and can be configured according to the network congestion level. If both of the above conditions are met simultaneously, i.e., the sender fails verification multiple times consecutively within a short period, the system will directly determine that the batch of data packets is invalid and may take measures to suspend data reception for that sender to prevent malicious attacks or resource abuse.

[0085] If the credential verification failure count is less than the preset consecutive failure threshold, and the failure time interval is greater than the preset pause verification time interval, it indicates that the failures are not consecutive within a short period, and the system will enter the normal digital signature verification process. In this process, the digital signature credential in the batch data packet is decrypted using the public key to obtain the first data digest. The digital signature credential is generated by the sender encrypting the batch data digest using their private key; the first data digest obtained after decryption should be consistent with the original batch data digest. Simultaneously, a hash calculation is performed on the batch data in the batch data packet to obtain the second data digest. A hash calculation is a one-way function that can map data of arbitrary length to a fixed-length hash value, used to verify data integrity.

[0086] If the first and second data digests are the same, the verification result is considered successful. Otherwise, the verification result is considered unsuccessful. The decrypted first data digest can be compared with the second data digest obtained by hashing the batch data. If they are identical, it indicates that the batch data packets have not been tampered with during transmission and the sender's identity is legitimate; in this case, the verification result is considered successful. Conversely, if the first and second data digests are inconsistent, it indicates that the data packets may have been tampered with or the digital signature is invalid; in this case, the verification result is considered unsuccessful.

[0087] This embodiment effectively enhances the security and reliability of batch data packets by introducing a multi-layered verification mechanism. First, by monitoring the sender's credential verification failure count and failure time interval, and comparing them with preset continuous failure thresholds and preset pause verification time intervals, batch data packets from malicious sources or those with persistent transmission problems can be quickly identified and blocked. This predictive mechanism avoids complex subsequent decryption and hash calculations on obviously abnormal data packets, thus saving system resources and effectively resisting potential threats such as denial-of-service attacks. Second, for batch data packets that pass the initial screening, the public key is used to decrypt the digital signature credential, and the result is compared with the hash calculation result of the batch data itself. This ensures the integrity of the data during transmission and the authenticity of the sender's identity. The digital signature mechanism utilizes the characteristics of asymmetric encryption, ensuring that any tampering with the data packet will result in a hash value mismatch, which will be detected by the system. Therefore, this embodiment not only effectively prevents data tampering but also identifies and processes abnormal data streams at an early stage, ensuring the accuracy of subsequent energy efficiency assessments and the stable operation of the system.

[0088] To illustrate this technical solution more clearly, a specific example is used below. Suppose the park management server receives a batch data packet from a production unit. First, the park management server queries the production unit's historical verification records. In the first scenario, if the query shows that the production unit has failed verification three times consecutively within the past 5 minutes (i.e., the credential verification failure count is 3, and the failure interval is less than 5 minutes), while the system's preset consecutive failure threshold is 3, and the pause interval for verification is 5 minutes, then because the credential verification failure count equals the preset consecutive failure threshold, and the failure interval is less than the preset pause interval, the park management server will immediately determine that the batch data packet is invalid and may temporarily stop receiving data from that production unit to prevent potential malicious behavior or persistent failures. In the second scenario, if the query shows that the production unit's credential verification failure count is 1, and the most recent failure occurred 1 hour ago (i.e., the failure interval is greater than 5 minutes), then because the credential verification failure count is less than the preset consecutive failure threshold, and the failure interval is greater than the preset pause interval, the system will continue with digital signature verification.

[0089] Specifically, the park management server obtains the public key corresponding to the production unit. Then, it uses this public key to decrypt the digital signature credential in the batch data packet, obtaining a first data digest. Simultaneously, it performs a hash calculation on the batch data in the batch data packet to obtain a second data digest. If the first and second data digests are completely identical, the verification result is considered successful, and the batch data packet will be used for subsequent time calibration and energy efficiency assessment. If the first and second data digests are inconsistent, the verification result is considered a failure, the credential verification failure count for that production unit will increase, and the failure time interval will be updated.

[0090] Through the above technical solutions, this embodiment can significantly improve the security, efficiency, and robustness of batch data packet verification. Specifically, by introducing the judgment of credential verification failure count and failure time interval, the system can intelligently identify and filter out abnormal data packets that fail consecutively within a short period of time, effectively preventing malicious attacks and resource waste, and avoiding system performance degradation caused by repeated processing of invalid data packets. In addition, by combining the decryption of digital signature credentials and the hash calculation of batch data, the integrity of data during transmission and the authenticity of the sender's identity are ensured, thereby providing a highly reliable data source for subsequent electricity energy efficiency assessment. This multi-layered verification mechanism makes the generation of electricity energy efficiency assessment reports more accurate and reliable, further ensuring the overall security and operational efficiency of the industrial and commercial electricity energy efficiency assessment system.

[0091] In some embodiments, step S202, performing time calibration based on the batch data packets to obtain a unified time stamp, may include, but is not limited to, the following steps:

[0092] Monitor the internal clock operating parameters and environmental parameters of the monitoring terminal;

[0093] Based on environmental parameter information, calculate the instantaneous drift rate of the terminal's internal clock operating parameters;

[0094] Based on the instantaneous drift rate, time compensation is performed on the time offset in the batch data packets to obtain the compensated offset;

[0095] Based on the compensation offset, the local timestamps in the batch data packets are calibrated to obtain a unified time stamp.

[0096] In some embodiments, the internal clock of a terminal device may be affected by various environmental and operational parameters such as temperature, voltage fluctuations, and aging, resulting in dynamic clock drift. If calibration is performed solely based on the initially calculated time offset without real-time monitoring and compensation for this dynamic drift, the accuracy of time calibration may decrease, thereby affecting the accuracy of subsequent energy efficiency assessments.

[0097] To this end, we can first monitor the internal clock operating parameters and environmental parameters of the terminal. Sensors integrated into the terminal device can be used to acquire operating parameters related to clock accuracy in real time. Internal clock operating parameters may include crystal oscillator frequency, power supply voltage, and internal temperature, which directly affect the clock's stability and accuracy. Environmental parameters may include external ambient temperature, humidity, air pressure, and electromagnetic interference intensity, which may indirectly or directly affect the terminal clock's operating status. The purpose is to provide a comprehensive data foundation for subsequent calculations of clock drift.

[0098] Then, based on environmental parameter information, the instantaneous drift rate of the terminal's internal clock operating parameters is calculated. A pre-established mapping relationship can be used to analyze the relationship between the monitored environmental parameters and the terminal's internal clock operating parameters, thereby quantifying the clock's drift rate at the current moment. For example, a functional relationship can be established to map input parameters such as temperature and voltage to the clock's frequency deviation, thus calculating the instantaneous drift rate. The purpose is to dynamically capture the clock's drift trend.

[0099] Then, based on the instantaneous drift rate, time compensation is applied to the time offset in the batch of data packets to obtain the compensated offset. The calculated instantaneous drift rate can be applied to the existing time offset in the batch of data packets. For example, if the instantaneous drift rate indicates that the clock is accelerating, the time offset needs to be reduced accordingly; if the clock is decelerating, the compensation needs to be increased. This compensation is cumulative and can effectively offset any additional clock drift that may occur during data packet generation and transmission. Its purpose is to ensure that the time offset can more accurately reflect the true difference between the terminal's local time and the campus reference time.

[0100] Finally, based on the compensation offset, the local timestamps in the batch data packets are calibrated to obtain a unified time stamp. The time-compensated offset can then be applied to the local timestamps recorded in the batch data packets. For example, by adding or subtracting the compensation offset from the local timestamps, they can be converted into a unified time stamp consistent with the park's reference time. The purpose is to provide a high-precision, high-reliability time base, enabling all electricity consumption data to be accurately aligned and analyzed on the park management server.

[0101] This embodiment addresses potential accuracy issues by introducing real-time monitoring of the terminal's internal clock operating parameters and environmental parameters, and dynamically calculating the instantaneous drift rate of the clock based on this information. Specifically, when the terminal device encapsulates power consumption data along with a local timestamp and time offset into batch data packets and sends them to the park management server, this embodiment first uses the monitored environmental parameters on the terminal side or receiving side (depending on the implementation) to accurately calculate the instantaneous drift rate of the terminal's internal clock. Since clock drift is dynamic, calculating the instantaneous drift rate allows for more precise capture of clock deviations over short periods. Based on this, the instantaneous drift rate is used to compensate for the original time offset in the batch data packets, generating a more accurate compensated offset. This compensation mechanism effectively offsets clock drift caused by environmental changes or the device's own characteristics, enabling the time offset to more accurately reflect the true difference between local time and the park's reference time. Finally, by using this compensated offset to calibrate the local timestamp in the batch data packets, a highly accurate and uniform time stamp can be obtained, providing a reliable time reference for subsequent power efficiency assessments.

[0102] To illustrate this technical solution more clearly, a specific example is used below. Suppose a production workshop in an industrial park is equipped with a smart meter terminal. This terminal is responsible for collecting electricity consumption data and uploading it to the park's management server. During a certain period, the ambient temperature in the workshop fluctuates significantly, causing a slight change in the frequency of the crystal oscillator inside the smart meter terminal, which in turn causes its internal clock to drift. Specifically, the smart meter terminal continuously monitors its internal clock operating parameters (e.g., crystal oscillator frequency, internal temperature) and environmental parameters (e.g., external ambient temperature). When the external ambient temperature rises from 25°C to 35°C, the temperature sensor inside the terminal detects this change. Based on the pre-stored crystal oscillator temperature characteristic curve, the system can calculate the frequency deviation of the crystal oscillator at the current temperature, thus obtaining a positive instantaneous drift rate, for example, a drift of 0.1 seconds per hour.

[0103] At this point, if the local timestamp in the batch data packet is 10:00:00, and the initially calculated time offset is +5 seconds (meaning the local time is 5 seconds ahead of the park's reference time), the actual local time may have become 5.05 seconds ahead of the park's reference time due to drift caused by the aforementioned temperature changes during data packet generation and transmission. This embodiment compensates for the original time offset by +5 seconds based on the calculated instantaneous drift rate (0.1 seconds / hour). Assuming one hour has passed since the last calibration to the current data packet generation, the compensation offset will be adjusted to +5 seconds + 0.1 seconds = +5.1 seconds. Finally, after receiving the batch data packet, the park management server uses this compensated offset (+5.1 seconds) to calibrate the local timestamp 10:00:00 in the batch data packet, obtaining a unified time stamp of 09:59:54.9. Through this dynamic compensation mechanism, even if the terminal clock drifts, it ensures that the power consumption data is accurately aligned to the park's unified time reference, thereby guaranteeing the accuracy of subsequent energy efficiency assessments.

[0104] Through the above technical solution, this embodiment can significantly improve the accuracy and reliability of electricity consumption data time calibration. This embodiment effectively reduces the impact of dynamic clock drift on time calibration accuracy by real-time monitoring of the terminal's internal clock operating parameters and environmental parameters, and dynamically calculating the instantaneous drift rate for time compensation. Therefore, it ensures that the electricity consumption data received by the park management server has a highly consistent time stamp, avoiding energy efficiency assessment errors caused by time deviations. This results in more accurate and reliable electricity consumption energy efficiency assessment reports, providing a solid data foundation for refined energy management in industrial and commercial parks.

[0105] In some embodiments, step S203, which involves performing an energy efficiency assessment based on a unified time stamp and generating an energy efficiency assessment report, may include, but is not limited to, the following steps:

[0106] Step S301: Obtain production mode information of production units within the park. Production mode information includes start-up time, duration, and production load characteristics.

[0107] Step S302: Based on the production mode information, the electricity consumption data under the same time mark is segmented to obtain multiple data segments;

[0108] Step S303: Calculate the power load curve based on the data segment;

[0109] Step S304: Extract features from the electricity load curve to obtain curve features, including shape features, peak and valley values, and rate of change.

[0110] Step S305: Based on the curve characteristics, perform pattern recognition on the data segment to obtain the production mode type;

[0111] Step S306: Calculate energy efficiency indicators and energy consumption attribution based on production mode type and energy efficiency assessment rule set;

[0112] Step S307: Associate the energy efficiency indicators, energy consumption attribution, and production mode type to generate an electricity energy efficiency assessment report.

[0113] In some embodiments, if the impact of the actual production mode of production units in industrial and commercial parks on electricity consumption behavior is not fully considered, the resulting energy efficiency assessment report may lack specificity and make it difficult to accurately identify energy consumption problems in specific production processes. This limits the effectiveness of energy efficiency improvement measures and may result in the provided energy efficiency assessment report remaining only at the macro level, failing to provide enterprises with refined energy consumption management and optimization suggestions.

[0114] To this end, production mode information for production units within the park can be obtained first. This information includes start-up time, duration, and production load characteristics. Production mode information related to production activities can be collected; this information can come from the production management system, sensor data, or manual input. The purpose is to provide business context for subsequent electricity consumption data analysis. Simultaneously, based on the production mode information, electricity consumption data under a unified time marker is segmented into multiple data segments. Continuous electricity consumption data streams can be logically divided according to the start and end times or production stages of the production mode. For example, the start and end of a production shift, or the operating cycle of a large piece of equipment, can all serve as the basis for segmentation. The purpose is to link electricity consumption data with specific production activities, facilitating targeted analysis.

[0115] Then, based on the data segments, the electricity load curve is calculated. This allows for the depiction of the trend of electricity consumption over time within each data segment. This is typically accomplished by aggregating, smoothing, or interpolating the segmented electricity data to visually represent the electricity consumption characteristics under this production pattern. Feature extraction is then performed on the electricity load curve to obtain curve features, including shape characteristics, peak and valley values, and rate of change. Representative parameters can be identified from the electricity load curve using mathematical algorithms or statistical methods, such as the overall shape of the curve (e.g., stable or fluctuating), maximum value (peak), minimum value (valley), and the rate of increase or decrease in electricity consumption (rate of change). These features are key to quantifying electricity consumption behavior patterns.

[0116] Next, based on the curve characteristics, pattern recognition is performed on the data segment to obtain the production mode type. Machine learning algorithms or preset rules can be used to compare the extracted curve features with known production mode templates, thereby automatically identifying the specific production mode corresponding to the current data segment, such as "equipment operation mode," "idle standby mode," or "production line full-load operation mode." Based on the production mode type and the energy efficiency assessment rule set, energy efficiency indicators and energy consumption attribution are calculated. The energy efficiency assessment rule set is a predefined series of standards and formulas used to measure energy utilization efficiency under different production modes. Energy efficiency indicators can be unit product energy consumption, equipment operating efficiency, etc., while energy consumption attribution refers to the cause of specific energy consumption, such as whether it is due to equipment aging, improper operation, or the production process itself.

[0117] Finally, energy efficiency indicators, energy consumption attributions, and production mode types are correlated to generate an electricity energy efficiency assessment report. This means integrating the calculated energy efficiency data with their corresponding specific production mode types to form a structured and easy-to-understand report. This report not only demonstrates energy efficiency levels but also identifies energy consumption problems and their causes, and connects them to specific production activities, providing decision-makers with clear directions for improvement.

[0118] This embodiment addresses the lack of specificity in energy efficiency assessment reports by closely integrating electricity consumption data with production mode information of production units within the industrial park. Specifically, firstly, acquiring production mode information provides crucial business context for electricity consumption data analysis. Secondly, based on this production mode information, electricity consumption data under a unified time stamp is segmented, ensuring that the granularity of subsequent analysis matches actual production activities and avoiding the conflation of electricity consumption behavior at different production stages. Thus, calculating the electricity load curve for each data segment provides a clear view of the electricity consumption characteristics under a specific production mode. Furthermore, by extracting features such as shape, peak-valley values, and rate of change from the electricity load curve, a quantitative description of complex electricity consumption behavior is achieved. Based on these curve features, pattern recognition can automatically and accurately determine the specific production mode type corresponding to the current electricity consumption data, thereby transforming abstract electricity consumption data into business-meaningful production activity information. Finally, by combining the identified production mode types and the preset energy efficiency assessment rule set, refined energy efficiency indicators and energy consumption attributions are calculated and correlated with production mode types. This allows the generated electricity energy efficiency assessment report to not only reflect the overall energy efficiency level, but also to deeply analyze the energy consumption problems and their causes in specific production processes, providing enterprises with accurate energy consumption optimization basis.

[0119] To illustrate this technical solution more clearly, a specific example is used below. Assume a manufacturing plant in an industrial park whose production line typically operates in "full-load production mode" during the day (8:00-18:00), and in "equipment standby mode" at night (18:00-8:00 the next day), and in "equipment maintenance mode" every Tuesday afternoon (14:00-16:00). This embodiment first acquires information on these production modes, including the start time, duration, and expected production load characteristics of each mode. Upon receiving time-calibrated electricity consumption data, the system automatically segments the daily electricity consumption data based on this production mode information. For example, the data from 8:00-18:00 is divided into one segment, and the data from 18:00-8:00 the next day is divided into another segment, with the Tuesday afternoon maintenance data segment identified separately. For each data segment, its electricity load curve is calculated, and shape characteristics (such as curve stability and volatility), peak and valley values, and rate of change are extracted. For example, the curve for "full-load production mode" might show a high and stable load, while "equipment standby mode" might show a low and slightly fluctuating load. Through pattern recognition, the system can accurately determine the production mode type corresponding to each data segment. Subsequently, combined with a preset set of energy efficiency assessment rules (e.g., unit product energy consumption standards, equipment standby power consumption limits, etc.), the system calculates energy efficiency indicators (e.g., unit product power consumption) and energy consumption attributions (e.g., excessive standby power consumption is due to a device not being completely shut down) for each mode. Finally, the system associates these detailed energy efficiency indicators and energy consumption attributions with the specific production mode type to generate a report that clearly indicates that "in full-load production mode, unit product power consumption is slightly higher than the industry average, and it is recommended to check production process parameters; in equipment standby mode, the standby power consumption of a certain device is abnormally high, and it is recommended to carry out maintenance or optimize the standby strategy," thus providing the factory with precise directions for energy efficiency optimization.

[0120] Through the above technical solution, this embodiment overcomes the limitation of energy efficiency assessment reports lacking specificity. By deeply integrating electricity consumption data with production mode information, it achieves refined analysis and pattern recognition of electricity consumption behavior, making the energy efficiency assessment results more closely aligned with actual production and operation. Therefore, the generated electricity consumption energy efficiency assessment report not only provides macro-level energy efficiency indicators but also specifically identifies the production modes in which energy consumption problems exist and provides energy consumption attribution. This provides highly actionable energy efficiency improvement suggestions for industrial and commercial park managers and enterprise decision-makers, significantly improving the accuracy and effectiveness of energy efficiency management and contributing to deeper levels of energy conservation and consumption reduction.

[0121] In some embodiments, step S304 involves feature extraction of the electricity load curve to obtain curve features, which may include, but is not limited to, the following steps:

[0122] Step S401: Denoise the electricity consumption data under a unified time stamp;

[0123] Step S402: Reconstruct the time series of the denoised electricity consumption data and update the electricity load curve to improve the time resolution of the electricity load curve.

[0124] Step S403: Perform multi-dimensional transformation on the updated electricity load curve to generate multiple load curve views;

[0125] Step S404: Extract complementary feature sets from multiple load curve views;

[0126] Step S405: Based on the complementary feature set, extract features from the updated electricity load curve to obtain curve features.

[0127] In some embodiments, directly extracting features from the original electricity load curve may encounter problems such as noise interference in the data, insufficient data sampling resolution, and the inability of single-dimensional feature extraction to fully reflect the complex characteristics of the load curve. These problems may result in inaccurate or inefficient extracted curve features, thereby affecting the reliability of subsequent pattern recognition and energy efficiency assessment reports.

[0128] Therefore, noise reduction processing can be performed on the electricity consumption data under a unified time stamp. Signal processing techniques can be applied to eliminate random noise, instantaneous spikes, or abnormal fluctuations in the electricity consumption data. The aim is to improve the quality and purity of the electricity consumption data, ensure the accuracy of subsequent feature extraction, and avoid noise misleading the evaluation results.

[0129] Then, the denoised electricity consumption data is reconstructed using time series analysis to update the electricity load curve, thereby improving its temporal resolution. After denoising, interpolation, resampling, or other time series analysis methods can be used to increase the data point density of the electricity load curve, resulting in a more refined curve representation in the time dimension. The aim is to capture subtle changes and dynamic characteristics of electricity load on short time scales, providing a foundation for more accurate pattern recognition.

[0130] The updated electricity load curve is then subjected to multi-dimensional transformations to generate multiple load curve views. A single-dimensional electricity load curve can be mapped to different feature spaces through various mathematical transformations (such as Fourier transform, wavelet transform, and statistical feature extraction), allowing for observation and analysis of the load curve from multiple perspectives, including the frequency domain, time-frequency domain, and statistical characteristics. The aim is to comprehensively reveal the inherent structure and potential patterns of the electricity load curve, overcoming the limitations of a single view.

[0131] Extracting complementary feature sets from multiple load curve views. Features that complement each other and jointly describe key characteristics of the load curve can be selected and combined from different load curve views. For example, harmonic components can be extracted from the frequency view, and peak-valley values ​​and rates of change can be extracted from the time-domain view. The aim is to construct a comprehensive, non-redundant, and information-rich feature set to more accurately characterize the complex behavior of electrical loads.

[0132] Finally, based on the complementary feature set, feature extraction is performed on the updated electricity load curve to obtain curve features. These features can comprehensively reflect the shape characteristics, peak and valley values, rate of change, and other deep patterns of the electricity load curve, providing a solid data foundation for subsequent production pattern identification and energy efficiency assessment.

[0133] This embodiment effectively addresses issues such as noise interference, insufficient resolution, and incomplete features that may arise during feature extraction by introducing a series of refined processing steps. First, denoising the electricity consumption data under a unified time stamp effectively filters out random fluctuations and outliers, ensuring data purity for subsequent analysis. This improved data quality allows subsequent time series reconstruction to more accurately capture the true changes in electricity load. Second, time series reconstruction enhances the temporal resolution of the electricity load curve, revealing previously overlooked short-term electricity consumption behaviors and load fluctuations, thus enabling the identification of refined production patterns. Building on this, multi-dimensional transformations of the updated electricity load curve reveal its intrinsic characteristics from different mathematical and physical perspectives, such as frequency components, energy distribution, and statistical regularities, avoiding the limitations of single-dimensional analysis. Finally, complementary feature sets are extracted from these multi-dimensional views, ensuring that the obtained curve features are both comprehensive and discriminative, more accurately representing electricity consumption behavior under different production patterns, and providing more reliable input for subsequent energy efficiency assessment and energy consumption attribution.

[0134] To illustrate this technical solution more clearly, a specific example is provided below. When denoising electricity consumption data under a unified time stamp, methods such as wavelet denoising, moving average filtering, or Kalman filtering can be used to effectively suppress high-frequency noise and instantaneous spikes. For example, a threshold denoising method based on wavelet transform can be used, separating noise signals from the original electricity consumption data by selecting appropriate wavelet basis functions and threshold strategies. After denoising, when reconstructing the electricity consumption data for time series to improve time resolution, cubic spline interpolation, linear interpolation, or machine learning-based time series super-resolution techniques can be used. For example, if the original data sampling interval is 1 minute, it can be reconstructed into data with 10-second intervals through interpolation, thus depicting the dynamic changes of the load curve in greater detail. When performing multi-dimensional transformations on the updated electricity load curve to generate multiple load curve views, this may include, but is not limited to: performing a Fast Fourier Transform (FFT) on the curve to obtain a frequency domain view, analyzing its harmonic components and dominant frequency characteristics; performing wavelet packet decomposition to obtain a time-frequency domain view, analyzing the temporal distribution of different frequency components; calculating statistical features (such as mean, variance, skewness, kurtosis, maximum, and minimum values) to obtain a statistical view; or performing difference operations to obtain a rate of change view. When extracting complementary feature sets from these multiple load curve views, the amplitude and phase of the main harmonics and the proportion of dominant frequency energy can be extracted from the frequency domain view; the instantaneous changes in high-frequency energy within a specific time period can be extracted from the time-frequency domain view; the overall distribution characteristics of the load curve can be extracted from the statistical view; and load abrupt change points and trends can be extracted from the rate of change view. These features complement each other, together forming a comprehensive feature vector. Finally, based on the complementary feature set, feature extraction is performed on the updated electricity load curve to obtain curve features. For example, the extracted frequency features, time-frequency features, statistical features, and rate of change features are combined into a high-dimensional feature vector, which serves as the final representation of the electricity load curve and is used for subsequent production pattern recognition.

[0135] Through the above technical solutions, this embodiment can significantly improve the accuracy and robustness of electricity load curve feature extraction. Noise reduction effectively improves data quality and reduces evaluation errors; time series reconstruction allows for the full display of details in the electricity load curve, helping to identify more refined production patterns; multi-dimensional transformation and complementary feature set extraction ensure that the obtained curve features can comprehensively and deeply reflect the complex characteristics of electricity load. This embodiment can more effectively address the complexity and uncertainty present in actual industrial and commercial electricity consumption data, thereby providing more accurate and reliable curve features for electricity energy efficiency assessment, and ultimately improving the accuracy and guiding value of the entire energy efficiency assessment report.

[0136] In some embodiments, step S403 involves performing multi-dimensional transformation on the updated electricity load curve to generate multiple load curve views, which may include, but is not limited to, the following steps:

[0137] Spectral analysis was performed on the updated electricity load curve to identify instantaneous high-frequency noise components and harmonic characteristics;

[0138] Based on the instantaneous high-frequency noise components and harmonic characteristics, the updated power load curve is filtered.

[0139] The filtered electricity load curve is transformed in multiple dimensions to generate multiple load curve views.

[0140] In some embodiments, a spectrum analysis can be performed on the updated electricity load curve to identify instantaneous high-frequency noise components and harmonic characteristics. The aim is to reveal the hidden frequency components and periodic patterns within the curve. Spectrum analysis methods such as Fourier transform can be used to identify the instantaneous high-frequency noise components and harmonic characteristics present in the electricity load curve. The instantaneous high-frequency noise components may originate from measurement errors, instantaneous load fluctuations, or grid interference, while harmonic characteristics are typically related to the operation of nonlinear loads, such as current or voltage distortions generated by equipment like frequency converters and rectifiers.

[0141] Then, based on the instantaneous high-frequency noise components and harmonic characteristics, the updated load curve is filtered to effectively remove or suppress these unnecessary interference components, thereby obtaining a smoother load curve that better reflects the actual power consumption pattern. For example, a low-pass filter can be used to remove high-frequency noise, or a notch filter can be used to suppress specific harmonic frequencies.

[0142] The filtered electricity load curve is then subjected to multi-dimensional transformations to generate multiple load curve views. These multi-dimensional transformations can include, but are not limited to, wavelet transform, short-time Fourier transform, and empirical mode decomposition (EMD). Their purpose is to reveal the characteristics of the electricity load curve from different time scales, frequency ranges, or feature spaces. For example, wavelet transform can provide both time and frequency domain information, helping to capture the local transient characteristics of the curve; short-time Fourier transform can analyze the changes in the signal's spectrum over time; and empirical mode decomposition can decompose complex signals into a series of intrinsic mode functions, revealing fluctuation patterns at different scales. Generating multiple load curve views provides richer and more comprehensive information for subsequent feature extraction.

[0143] This embodiment improves the accuracy and robustness of subsequent feature extraction by performing spectral analysis and filtering on the updated electricity load curve before multi-dimensional transformation. Specifically, spectral analysis helps to accurately identify and quantify noise and harmonic interference in the curve. If these interferences are not processed, they may be amplified or introduce spurious features during multi-dimensional transformation, thus affecting the final energy efficiency assessment results. Targeted filtering removes these interference components, making the electricity load curve cleaner and more reflective of actual electricity consumption patterns. Based on this, multi-dimensional transformation generates multiple load curve views that are clearer and more effective, more accurately capturing key information such as the shape characteristics, peak and valley values, and rate of change of the electricity load curve, thus providing high-quality input data for subsequent pattern recognition and energy efficiency assessment.

[0144] Through the above technical solution, this embodiment introduces spectrum analysis and filtering steps, effectively avoiding interference from noise and harmonics on feature extraction. As a result, the generated load curve view has a higher signal-to-noise ratio and stronger representativeness, more accurately reflecting actual electricity consumption behavior and patterns. This not only improves the accuracy of load curve feature extraction but also provides a more reliable data foundation for subsequent production mode type identification, energy efficiency index calculation, and energy consumption attribution, thereby significantly improving the overall accuracy and reliability of industrial and commercial electricity energy efficiency assessment.

[0145] The beneficial effects of implementing the embodiments of the present invention include: the embodiments of this application first obtain the local timestamp and the park reference timestamp of the park management server, then calculate the time offset based on the local timestamp and the park reference timestamp, then generate batch data packets based on the time offset, and finally send the batch data packets to the park management server. The park management server performs power consumption energy efficiency assessment and generates a power consumption energy efficiency assessment report. Thus, the assessment report can be generated by combining the time offset to realize power consumption energy efficiency assessment, thereby improving accuracy and reliability.

[0146] like Figure 2 As shown, this embodiment of the invention also provides an industrial and commercial power efficiency assessment system based on multimodal data, comprising:

[0147] Data acquisition module 501 is used to acquire local timestamps and park reference timestamps from the park management server;

[0148] The offset calculation module 502 is used to calculate the time offset based on the local timestamp and the park reference timestamp;

[0149] The data packet generation module 503 is used to generate batch data packets based on the time offset.

[0150] The energy efficiency assessment module 504 is used to send batch data packets to the park management server. The park management server is used to conduct energy efficiency assessments based on the batch data packets and generate energy efficiency assessment reports.

[0151] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0152] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

Claims

1. A method for evaluating the energy efficiency of industrial and commercial electricity consumption based on multimodal data, characterized in that, Includes the following steps: Obtain the local timestamp and the park reference timestamp from the park management server; Calculate the time offset based on the local timestamp and the park reference timestamp; Generate batch data packets based on the time offset; The batch data packets are sent to the park management server, which then performs an energy efficiency assessment based on the batch data packets and generates an energy efficiency assessment report.

2. The method according to claim 1, characterized in that, The step of calculating the time offset based on the local timestamp and the park reference timestamp includes: Send a network probe signal to the park management server to obtain network round-trip time information; Calculate the degree of network communication fluctuation based on the network round-trip time information; Calculate the initial offset based on the local timestamp and the park reference timestamp; If the network communication fluctuation is less than a preset fluctuation threshold, then the time offset is obtained by weighted averaging of multiple initial offsets. If the network communication fluctuation is greater than a preset fluctuation threshold, then the median filter is applied to the multiple initial offsets to obtain the time offset.

3. The method according to claim 1, characterized in that, The step of generating batch data packets based on the time offset includes: Obtain electricity consumption data; According to a preset time interval, the electricity consumption data, the local timestamp, and the time offset are integrated to obtain batch data; Perform a hash calculation on the batch data to obtain a batch data digest; The batch data digest is digitally signed to generate a digital signature certificate; The batch data and the digital signature certificate are encapsulated to obtain the batch data packet.

4. The method according to claim 1, characterized in that, The step of performing an energy efficiency assessment based on the batch data packets and generating an energy efficiency assessment report includes: The batch of data packets is verified to obtain the verification results; If the verification result is successful, then time calibration is performed based on the batch data packets to obtain a unified time stamp; Based on the unified time stamp, an energy efficiency assessment is performed, and an energy efficiency assessment report is generated.

5. The method according to claim 4, characterized in that, The verification of the batch of data packets to obtain the verification result includes: Based on the batch data packets, obtain the sender's public key, credential verification failure count, and failure time interval; If the count of failed credential verification is greater than a preset consecutive failure threshold and the time interval between failures is less than a preset pause verification time interval, then the verification result is determined to be an invalid data packet. If the count of failed credential verification is less than a preset consecutive failure threshold and the time interval between failures is greater than a preset pause verification time interval, then the digital signature credential in the batch data packet is decrypted according to the public key to obtain a first data digest. A second data digest is obtained by hashing the batch data in the batch data packet. If the first data digest and the second data digest are the same, the verification result is determined to be successful; otherwise, the verification result is determined to be unsuccessful.

6. The method according to claim 4, characterized in that, The step of performing time calibration based on the batch data packets to obtain a unified time stamp includes: Monitor the internal clock operating parameters and environmental parameters of the monitoring terminal; Based on the environmental parameter information, calculate the instantaneous drift rate of the terminal's internal clock operating parameters; Based on the instantaneous drift rate, time compensation is performed on the time offset in the batch data packets to obtain the compensated offset; Based on the compensation offset, the local timestamps in the batch data packets are calibrated to obtain the unified time stamp.

7. The method according to claim 4, characterized in that, The step of conducting an energy efficiency assessment based on the unified time stamp and generating the energy efficiency assessment report includes: Obtain production mode information of production units within the park, including start-up time, duration, and production load characteristics; Based on the production mode information, the electricity consumption data under the unified time mark is segmented to obtain multiple data segments; Calculate the electricity load curve based on the data segment; Feature extraction is performed on the electricity load curve to obtain curve features, which include shape features, peak and valley values, and rate of change. Based on the curve characteristics, pattern recognition is performed on the data segment to obtain the production mode type; Based on the production mode type and energy efficiency assessment rule set, calculate energy efficiency indicators and energy consumption attribution; The energy efficiency indicators, energy consumption attribution, and production mode type are correlated to generate the electricity consumption energy efficiency assessment report.

8. The method according to claim 7, characterized in that, The step of extracting features from the electricity load curve to obtain curve features includes: The electricity consumption data under the unified time stamp is subjected to noise reduction processing; The denoised electricity consumption data is reconstructed into a time series, and the electricity load curve is updated to improve the time resolution of the electricity load curve. The updated electricity load curve is transformed in multiple dimensions to generate multiple load curve views; Extract complementary feature sets from the multiple load curve views; Based on the complementary feature set, feature extraction is performed on the updated electricity load curve to obtain the curve features.

9. The method according to claim 8, characterized in that, The process of performing multi-dimensional transformation on the updated electricity load curve to generate multiple load curve views includes: Spectral analysis was performed on the updated electricity load curve to identify instantaneous high-frequency noise components and harmonic characteristics; Based on the instantaneous high-frequency noise components and the harmonic characteristics, the updated power load curve is filtered. The filtered electricity load curve is transformed in multiple dimensions to generate the multiple load curve views.

10. A system for evaluating the energy efficiency of industrial and commercial electricity consumption based on multimodal data, characterized in that, include: The data acquisition module is used to acquire the local timestamp and the park reference timestamp from the park management server; The offset calculation module is used to calculate the time offset based on the local timestamp and the park reference timestamp; A data packet generation module is used to generate batch data packets based on the time offset. The energy efficiency assessment module is used to send the batch data packets to the park management server, and the park management server is used to perform an energy efficiency assessment based on the batch data packets and generate an energy efficiency assessment report.

Images