A low-carbon environment-friendly intelligent park management system

Abstract

This invention relates to the field of smart park management technology, specifically to a low-carbon and environmentally friendly smart park management system. It aims to solve the technical problem of false alarms and erroneous control caused by single-point sensor failures due to the lack of cross-system logical mechanism verification in existing smart park systems. The technical solution includes: acquiring monitoring data from multiple heterogeneous subsystems within the park through a multi-source data acquisition module; constructing and updating a data relationship model based on equipment physical connections, operating logic, and historical data using a data modeling module; and performing the following operations by a collaborative verification and intelligent decision-making module: when the main monitoring data is abnormal, retrieving associated auxiliary monitoring data according to the model and performing cross-validation, calculating a confidence score, and executing the corresponding decision-making strategy and control instructions based on the score interval. This invention has the beneficial effects of identifying single-point false alarms, reducing the frequency of erroneous actions, and improving the stability of cross-system linkage and the efficiency of low-carbon management in smart parks.

Description

Technical Field

[0001] This invention relates to the field of smart park management technology, and more specifically, to a smart park management system based on low-carbon and environmental protection. Background Technology

[0002] With the growing awareness of low-carbon and environmental protection, modern smart parks have integrated a large number of IoT sensing devices. Existing smart park management systems typically adopt a layered architecture, with each subsystem independently collecting data and displaying it on a central control platform.

[0003] Chinese Patent Announcement No. CN118072255A discloses a smart park multi-source data dynamic monitoring and real-time analysis system and method. Its technical solution uses edge computing to extract features from video, audio and sensor data, and uses the proximity of physical space to verify events. For example, when a temperature sensing device in a certain area alarms, the system automatically calls the camera or adjacent sensor in that area for auxiliary judgment.

[0004] The technical problem with this existing technology is that: First, the existing verification logic relies too heavily on the physical proximity of devices, ignoring the non-spatial coupling relationships based on operating mechanisms that are widespread in industrial parks. For example, a sudden increase in production line load can cause pressure fluctuations in remote water supply pumps, which makes the system unable to identify implicit causal risks across subsystems. Then, existing systems mostly use positive confirmation logic, that is, an alarm is triggered once the main sensor reaches the threshold and there is an environmental fluctuation. However, single-point false data generated by sensor aging, electromagnetic interference or communication jitter can easily induce erroneous cross-system linkage. For example, a false current surge caused by a current transformer failure may cause the system to misjudge as overload and cut off the entire production circuit, resulting in huge economic losses. Finally, the judgment logic of existing systems is mostly based on simple threshold triggers, lacking quantitative assessment of the authenticity of abnormal events, and lacking effective interception and self-healing mechanisms when issuing instructions, resulting in extremely low system fault tolerance.

[0005] In view of this, a smart park management system based on low carbon and environmental protection is proposed. Summary of the Invention

[0006] The purpose of this invention is to provide a smart park management system based on low carbon and environmental protection, so as to solve the technical problem in the existing smart park system that the lack of cross-system logic mechanism verification leads to false alarms and false linkage control caused by single-point sensor failure.

[0007] To solve the above-mentioned technical problems, the present invention provides a smart park management system based on low-carbon and environmental protection, comprising: The multi-source data acquisition module is used to acquire monitoring data from multiple heterogeneous subsystems within the park in real time. The monitoring data includes the operating status and energy consumption parameter data from at least two subsystems in energy management, environmental monitoring, and equipment operation and maintenance. The data modeling module constructs a data relationship model that reflects the causal or temporal correlation between different monitoring data based on the physical connection relationship of the equipment, the operation logic rules and historical monitoring data, and optimizes and updates the model based on new data and decision feedback. The collaborative verification and intelligent decision-making module is used to perform the following operations: S1: When any monitoring data defined as the main monitoring data becomes abnormal, at least one auxiliary monitoring data with a preset correlation relationship with it is retrieved according to the data relationship model. S2: Based on the abnormal characteristics of the main monitoring data and the real-time status of at least one auxiliary monitoring data, cross-validation is performed to calculate the event confidence score that characterizes the overall credibility of the abnormal event. S3: Based on the preset interval to which the event confidence score belongs, match the corresponding decision strategy, generate and execute control instructions associated with the decision strategy.

[0008] As a further improvement to this technical solution, the data modeling module is used to construct two types of association relationships, including: Based on the prior association between the physical connection of the equipment and the operation control logic; Based on the posterior statistical associations mined by time-series correlation analysis and causal discovery algorithms on historical monitoring data; In constructing the posterior statistical association, a Granger causality test is performed on historical monitoring data, and the test statistic is then used to determine the causality of the data. Data pairs whose values ​​are less than a preset significance level and whose calculated causal relationship coefficients exceed a preset strength threshold are identified as associated edges with a temporal influence relationship, and the causal relationship coefficients are quantified as the initial association weights of the associated edges.

[0009] As a further improvement to this technical solution, step S2 involves calculating the event confidence score. The method is as follows: ,in, The basic weighting coefficients for the main monitoring data; Based on the abnormal deviation of the main monitoring data The calculated basic anomaly contribution value, which is related to There is a positive correlation; For the first The association weight of each auxiliary monitoring data point in the data relationship model; This is a logical consistency function, whose input is the abnormal status of the main monitoring data. With the Current status of auxiliary monitoring data The output is a support score for anomalies in the main monitoring data; when When the direction of change is consistent with the direction expected based on the data relationship model, Output a positive value, otherwise output zero or a negative value; For the first The hardware health coefficient of each monitoring node is calculated in real time or updated periodically based on the node's historical failure rate, signal noise level, or communication stability. This represents the total number of auxiliary monitoring data used in this collaborative verification.

[0010] As a further improvement to this technical solution, the preset decision-making strategy in step S3 includes at least: High confidence execution strategy: When the event confidence score is higher than the first threshold, it is determined to be a high confidence real event. Then, based on the event type and combined with the data relationship model, the pre-set cross-system collaborative control strategy library is invoked, and an optimized control instruction set is sent to the relevant subsystems. Data false alarm interception strategy: When the event confidence score is lower than the second threshold and the original abnormal value of the main monitoring data exceeds the safety threshold, it is judged as a potential single point of failure of the data source, a maintenance check work order for the data source device is generated, and any linkage control instructions triggered by this abnormal event are prevented from being executed, and the existing control logic remains unchanged. Suspected event verification strategy: When the event confidence score is between the first threshold and the second threshold, the manual review process is triggered, the associated data group, confidence score and preliminary analysis conclusions are pushed to the management terminal, and the issuance of automatic control commands is suspended.

[0011] As a further improvement to this technical solution, the specific logic in the data false alarm interception strategy to prevent the execution of any linkage control commands triggered by this abnormal event includes: After the collaborative verification and intelligent decision-making module is started, the system creates and activates an instruction latch flag for each potentially affected actuator. When a false alarm is detected, the collaborative verification and intelligent decision-making module clears the instruction latch flag, so that the controllers of each actuator ignore all control instructions derived from this abnormal event and continue to execute the instructions output by their original control loops.

[0012] As a further improvement to this technical solution, the multi-source data acquisition module includes an edge gateway layer deployed on-site, and the edge gateway layer is configured as follows: Through built-in multiple protocol parsing engines, private protocol data packets from different vendor subsystems are parsed in real time and uniformly mapped into standardized information model objects. The information model objects include at least data values, physical units, collection timestamps, device spatial location codes, and the identifier of the subsystem to which they belong.

[0013] As a further improvement to this technical solution, the hardware health coefficient The real-time determination methods include: The system continuously monitors the data quality indicators of each monitoring node, including signal amplitude stability, noise spectrum characteristics, and the proportion of invalid data frames. When the data quality indicator of a node continuously deteriorates and falls below the health threshold, the system automatically performs at least one of the following operations: adjusts the association weight of the node. Set to zero, or set its logical consistency function. The output is forced to be zero.

[0014] As a further improvement to this technical solution, the system also includes an energy efficiency optimization control submodule, which is configured to execute the following energy-saving optimization process: The first category of key energy consumption data, which reflects the core production capacity activities of the park, is defined as the main monitoring data and monitored in real time. When the first type of key energy consumption data undergoes a preset change that conforms to the preset energy consumption optimization mode, the collaborative verification and intelligent decision-making module is activated. Based on the data relationship model, the second type of auxiliary energy consumption data with a correlation weight exceeding a preset threshold with the first type of key energy consumption data is selected as the core auxiliary monitoring data for collaborative verification. The second type of auxiliary energy consumption data corresponds to the energy consumption of the auxiliary system that supports the core production capacity activities. If the verification results indicate that the correlation between the first type and the second type of energy consumption data is a high-confidence real event, then based on the correlation confirmed by the collaborative verification, an optimization instruction to adjust the operating parameters of the auxiliary system is generated and issued. If the verification result is low confidence or a false alarm, then any direct shutdown or load reduction command to the auxiliary system triggered solely based on a single change in the first type of key energy consumption data shall be suppressed or rejected.

[0015] As a further improvement to this technical solution, the main monitoring data includes physical quantities such as the thermal state, force state, or substance concentration of the reaction equipment, and the auxiliary monitoring data includes physical or digital quantities reflecting electrical parameters, mechanical state, or logic switching quantities.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. In the intelligent park management system based on low - carbon environmental protection, by constructing a data association relationship model and introducing multiple associated auxiliary data for collaborative verification and cross - evidence collection when the master data is abnormal, it can effectively identify false signals caused by single - point sensor failures, noise interference, etc., avoiding the problem of rampant false alarms in the traditional single - threshold alarm mode, and making the triggering of alarms and subsequent control instructions based on a more comprehensive and credible evidence chain.

[0017] 2. In the intelligent park management system based on low - carbon environmental protection, the proposed confidence score mechanism and hierarchical decision - making strategy achieve a degree - based judgment of abnormal events, rather than a simple yes / no binary judgment. The system can take response measures with different intensities according to the confidence level, such as immediate execution, manual review, and interception of false alarms, so as to generate more refined and reasonable control instructions; combined with the association model, the system can insight into the dynamic relationship between production load and the energy consumption of auxiliary systems, and only implement energy - saving regulation when the associated evidence is sufficient, avoiding the comfort decline or production impact caused by one - size - fits - all control, and achieving precise carbon reduction on the premise of ensuring operation. Brief Description of the Drawings

[0018] Figure 1 It is the overall system block diagram of the present invention. Detailed Embodiments

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention. Embodiment

[0020] With the in - depth promotion of the concept of low - carbon environmental protection, various Internet of Things sensing devices are widely deployed in modern intelligent parks. Currently, common intelligent park management systems mostly adopt a hierarchical architecture, where each subsystem independently collects data and centrally displays it on the central platform; such systems usually use edge computing technology to perform local feature extraction on video, audio, and sensor data, and conduct preliminary event verification based on the proximity relationship of devices in the physical location; however, such systems still have obvious deficiencies: due to the lack of consistency verification of the logical relationships and operating mechanisms between different systems, the failure of a single sensor can easily trigger system false alarms and may lead to incorrect linkage control, affecting the safety and reliability of the normal operation of the park, and thus may not be able to achieve the concept of low - carbon environmental protection; In view of this, please refer to Figure 1 As shown, the purpose of the present invention is to provide an intelligent park management system based on low - carbon environmental protection, and this system includes: The multi-source data acquisition module is used to acquire monitoring data from multiple heterogeneous subsystems within the park in real time. The monitoring data includes the operating status and energy consumption parameter data from at least two subsystems in energy management, environmental monitoring, and equipment operation and maintenance. The data modeling module constructs a data relationship model that reflects the causal or temporal correlation between different monitoring data based on the physical connection relationship of the equipment, the operation logic rules and historical monitoring data, and optimizes and updates the model based on new data and decision feedback. The collaborative verification and intelligent decision-making module is used to perform the following operations: S1: When any monitoring data defined as the main monitoring data becomes abnormal, at least one auxiliary monitoring data with a preset correlation relationship with it is retrieved according to the data relationship model. S2: Based on the abnormal characteristics of the main monitoring data and the real-time status of at least one auxiliary monitoring data, cross-validation is performed to calculate the event confidence score that characterizes the overall credibility of the abnormal event. S3: Based on the preset interval to which the event confidence score belongs, match the corresponding decision strategy, generate and execute control instructions associated with the decision strategy; Further disclosure of the above system: Considering that the existing subsystems in the park adopt independent acquisition architectures, with inconsistent data protocols and formats and a lack of correlation mapping between physical quantities, the multi-source data acquisition module constructed in this system not only achieves standardized acquisition, but more importantly, it acquires heterogeneous data reflecting the coupling of the park's energy-environment-machine mechanism, such as motor current and pressure / concentration at the remote process end. This provides a data foundation for overcoming the limitations of physical proximity and conducting causal verification based on operational logic. Therefore, the multi-source data acquisition module is divided into a field sensing layer, an edge acquisition layer, and a data transmission layer: The field sensing layer deploys monitoring terminals adapted to different subsystems. For example, the energy management subsystem deploys smart meters, water meters, and gas meters to collect energy consumption parameters such as active power, cumulative water consumption, and gas flow; the environmental monitoring subsystem deploys temperature and humidity sensors, CO2 concentration sensors, and PM2.5 sensors. The system collects environmental parameters such as temperature, humidity, gas concentration, and dust content. The equipment operation and maintenance subsystem deploys vibration sensors, pressure transmitters, and current transformers to collect operating status parameters such as equipment vibration amplitude, pipeline pressure, and operating current. An industrial-grade edge gateway is deployed at the edge acquisition layer, with a built-in data buffer unit. The sampling frequency is set to a preset interval, for example, 500 milliseconds, to ensure real-time data acquisition. The data transmission layer uses a combination of wired Ethernet and wireless LoRa. Wired Ethernet is used for high-bandwidth data transmission from fixed devices, while wireless LoRa is used for low-power data transmission from distributed sensors. The transmission protocol uses the lightweight MQTT protocol, with data transmission latency controlled to within 1 second, ensuring data transmission stability and low latency. The acquisition accuracy of energy consumption parameters is set to 0.1-level industrial metrology standards.

[0021] Further explanation is needed: the 1-second transmission delay is based on the industrial-grade data transmission delay requirements in the "Technical Requirements for Data Transmission of Industrial Internet of Things Systems" (T / CIOT1001-2022). Combined with the park's wireless LoRa transmission distance (maximum 800 meters) and wired Ethernet networking structure, the transmission delay is optimized to ≤1 second by setting the MQTT protocol's QoS level (to QoS1), meeting the park's real-time management needs. The 0.1-level acquisition accuracy follows national / industry metrological standards such as "Electromechanical AC Energy Meter" (GB / T17215-2021) and "Smart Water Meter" (CJ / T188-2018). 0.1-level accuracy enables precise measurement of energy consumption parameters, providing a reliable data foundation for low-carbon energy efficiency analysis. The operating status data acquisition range needs to cover the entire rated operating range of the equipment, determined based on the product nameplate parameters of each piece of equipment in the park. For example, if the rated current of a 37kW water pump is 72A, the operating current acquisition range is set to 0-100A to ensure capture of the equipment's no-load, rated, and overload states. Through the aforementioned technical means, unified data collection from at least two heterogeneous subsystems in the park can be achieved, eliminating data silos and providing comprehensive and real-time basic data for subsequent data modeling and intelligent decision-making, thus supporting the overall low-carbon management and control of the park.

[0022] Considering that the existing system only realizes data aggregation and display without exploring the inherent causal and temporal relationships between data, and cannot support cross-system collaborative analysis, it is necessary to construct a quantitative correlation model based on device physical connections, operating logic, and historical data to provide the core logical basis for collaborative verification. The specific implementation is as follows: The data modeling module uses a directed weighted graph as its model carrier. Nodes represent monitoring data, edges represent the relationships between data, and edge weights represent the strength of these relationships. The modeling process consists of three stages. The first stage extracts the physical connection relationships of equipment, outlining the physical topology links of equipment within the park, such as the physical power supply and fluid connections between the central air conditioning unit and cooling water pumps and blowers, and the physical connections between the main equipment and auxiliary power supply equipment on the production line. The second stage outlines the operational logic rules, determining the timing control logic and condition triggering logic between equipment based on the equipment control manual and the park's operation and maintenance procedures, such as a 5-second delay before the blower starts after the air conditioning unit starts, and a 5-second delay before the main motor starts. The auxiliary pump operates synchronously; in the third stage, historical monitoring data from the past 12 months is imported, and an initial correlation framework is constructed by combining physical connections and operational logic. The correlation weights are then optimized through data fitting; the initial weights for strong logical correlations are 0.6 to 0.8, and the initial weights for weak logical correlations are 0.2 to 0.4; the model adopts an incremental update mechanism, importing new real-time data every 24 hours. At the same time, the weights and existence of the correlation edges are adjusted by combining manual review feedback and the execution effect of control commands. The iteration precision of the model optimization is set to 0.01 to balance the model optimization precision and iteration efficiency. When the weight adjustment range is less than 0.01, the model is considered to have converged. It should be further explained that the initial values ​​for strong / weak logical association weights are determined based on the park's 12-month equipment operation logs and maintenance records, statistically analyzing the synchronicity of data changes between devices with different physical / logical associations. For strong logical associations, such as the data synchronization change rate between air conditioning units and cooling water pumps ≥85%, the initial weight is set at 0.6-0.8; for weak logical associations, such as the data synchronization change rate between park lighting and ambient temperature 30%-50%, the initial weight is set at 0.2-0.4. The synchronization change rate is calculated as: number of data synchronization changes / total number of monitoring times × 100%. By employing the aforementioned technical means, a quantitative model reflecting the causal and temporal correlations between data is constructed, transforming discrete data into a correlated whole, enabling in-depth data fusion analysis, and providing core logical support for subsequent collaborative verification.

[0023] Considering that single-point monitoring data is susceptible to anomalies caused by sensor malfunctions, signal noise, and acquisition drift, relying solely on single-point data to determine anomalies is prone to false alarms. Therefore, it is necessary to introduce associated auxiliary monitoring data to improve the reliability of anomaly determination through multiple data corroborations. Thus, step S1 of the collaborative verification and intelligent decision-making module is implemented as follows: The system pre-defines the rules for dividing primary and auxiliary monitoring data. Primary monitoring data reflects the core energy consumption, core equipment status, and core environmental indicators of the park, such as the total energy consumption of central air conditioning, the main motor current of the production line, and the CO2 concentration in the core area of ​​the park. Auxiliary monitoring data is supporting data that has a physical or logical relationship with the primary monitoring data, such as cooling water pump energy consumption, fan operating status, and ambient temperature. The system monitors all primary monitoring data in real time. The normal range is determined by the historical 95th percentile, that is, the lower limit of the normal range is the 5th percentile of the historical data, and the upper limit is the 95th percentile of the historical data. When the real-time value exceeds the preset multiple of the upper and lower limits of the normal range, for example, 1.2 times, it is determined that the primary monitoring data is abnormal. When the primary monitoring data is triggered by an abnormality, the system traverses the data association model and retrieves auxiliary monitoring data that has an edge related to the primary monitoring data and whose association weight is not less than the preset weight. For example, when the preset weight is 0.3, it can not only eliminate a large number of weakly associated invalid data, but also retain more effective associated data, providing a reliable auxiliary data foundation for cross-validation. At least one auxiliary data is retrieved, and priority is given to retrieving the top 3 auxiliary data with the highest association weight. By following the steps above, we can accurately identify anomalies in the main monitoring data and select highly relevant auxiliary monitoring data to provide a data foundation for subsequent cross-validation and avoid introducing irrelevant data to interfere with the validation results. Considering that the existence of primary and secondary data alone cannot quantify the authenticity and credibility of anomalies, a multi-dimensional integrated scoring formula needs to be constructed. This formula should comprehensively consider the degree of anomaly in the primary data, the correlation strength of the secondary data, logical consistency, and hardware reliability to calculate a unified confidence score. Therefore, step S2 of the collaborative verification and intelligent decision-making module is implemented as follows: The event confidence score is calculated using a preset formula. The formula is ,in, The basic weighting coefficient for the primary monitoring data, set at 0.6, was determined through the following process: Using the Analytic Hierarchy Process (AHP), the correlation weights, logical consistency, and hardware health of the primary and auxiliary monitoring data were used as four evaluation indicators. The weight percentage of each indicator was determined through expert scoring and data verification of 50 typical anomaly events in the park. The primary monitoring data, as the core basis for anomaly triggering, accounted for 60% of the weighting, while the remaining indicators combined accounted for 40%. Therefore... The value is 0.6; Based on the abnormal deviation of the main monitoring data The calculated basic anomaly contribution value, which is related to There is a positive correlation. The abnormal deviation of the main monitoring data is calculated as follows: , The deviation is normalized to the range of 0 to 1. The normalization denominator of 3 is the deviation of the main monitoring data of all real abnormal events in the park within 12 months. 99% of the real abnormal events have a deviation of ≤3. Therefore, setting the denominator to 3 can normalize 99% of the real abnormal deviations to the range of 0-1. At the same time, extreme outliers, such as deviation >3, are uniformly set to 1 to avoid excessive interference of a single extreme value on the score. For the first The association weight of each auxiliary monitoring data in the association model is directly called by calling the weight value of the corresponding association edge in the model; This is a logical consistency function, whose input is the abnormal status of the main monitoring data. With the Current status of auxiliary monitoring data The output is a support score for anomalies in the main monitoring data. The output is 1 when the direction of change is consistent with the expected direction of the correlation model, 0.5 when the direction is consistent, -0.5 when the direction is opposite, and 0 when there is no correlation. The score of the logical consistency function is based on the logical correlation characteristics of the operation of the park equipment. The support of auxiliary data for the anomaly of the master data is divided into four levels. At the same time, it is evaluated by finite number of simulated abnormal scenarios. This score is used to quantify the support / opposition of auxiliary data. For the first In this embodiment, the hardware health coefficient of each monitoring node is set to 0 to 1, and is calculated in real time or updated periodically based on the node’s historical failure rate, signal noise level or communication stability. The total number of auxiliary monitoring data participating in this collaborative verification is 1 to 3 in this embodiment; the final calculated event confidence score is... The range of values ​​is limited to If multiple auxiliary data points positively support a calculated value exceeding 1, it is uniformly corrected to 1; if the calculated value is negative, it is uniformly corrected to 1. Through this normalization process, the scoring system is ensured to intuitively reflect the true probability distribution of abnormal events.

[0024] Through the above steps, multi-dimensional parameters are integrated and calculated to quantify the overall credibility of abnormal events, avoid scoring bias caused by a single factor, and provide an accurate and objective basis for subsequent classification decisions.

[0025] Considering that different confidence levels of abnormal events correspond to different processing requirements, high-confidence events require immediate execution of low-carbon optimization control, while low-confidence events require interception of false alarm commands. Therefore, it is necessary to set up hierarchical decision intervals and match differentiated decision strategies. Thus, step S3 of the collaborative verification and intelligent decision-making module is implemented as follows: The system presets three intervals for event confidence scoring. The first threshold is set to 0.8 in this embodiment, and the second threshold is set to 0.3. When the value is greater than 0.8, it is judged as a high-confidence real event, and a high-confidence execution strategy is matched; when When the value is less than 0.3, it is judged as a low-confidence potential false alarm, and the data false alarm interception strategy is applied; when When the value is between 0.3 and 0.8, it is judged as a suspected event and matched with the suspected event pending verification strategy. The system generates corresponding control instructions according to the matched strategy. The high confidence strategy generates cross-system collaborative optimization instructions, the false alarm interception strategy generates maintenance work orders and instruction blocking instructions, and the pending verification strategy generates manual review push instructions and executes the corresponding instructions. Through the above steps, a graded response to abnormal events can be achieved, and the handling strategies for different confidence scenarios can be accurately matched. This ensures timely handling of high-confidence anomalies while avoiding erroneous control caused by false alarms, thus balancing the efficiency of low-carbon management and system stability.

[0026] Considering that the physical connection and operational logic of equipment form the basis for data relationships, establishing such relationships prior to data mining ensures the rationality of the model's underlying logic and avoids logical biases caused by pure data mining. Therefore, the data modeling module includes prior relationships based on the physical connection and operational control logic of equipment, as specifically implemented below: The prior association relationships are divided into physical connection associations and operational logic associations. Physical connection associations are established by analyzing the physical topology of the park's equipment, such as the power supply link from the power distribution room to the power distribution cabinet to the electrical equipment, and the water supply link from the water pump to the pipeline to the water-using equipment. This establishes direct associations between upstream and downstream monitoring data, such as the output power of the power distribution cabinet and the energy consumption of the electrical equipment, and the pressure of the water pump and the flow rate of the water-using equipment. The association edges are bidirectional, with an initial weight of 0.7. In this embodiment, the initial weight is based on the energy / material transmission efficiency of the physical links of the park's equipment, such as an energy transmission efficiency of ≥90% for the power supply link and a fluid transmission efficiency of ≥85% for the water supply link. Data associations between physically connected devices are also established. The correlation is extremely strong, and the value is assigned based on the synchronous change rate (≥90%) of 12 months of historical data. The operation logic correlation is established based on the equipment control logic and operation and maintenance procedures, and the time sequence or conditional correlation is established. For example, after the air conditioning unit start command is issued, the fan operation status and air conditioning energy consumption are correlated in a time sequence. When the ambient temperature is higher than 28℃, the fresh air system air volume and the park CO2 concentration are correlated in a conditional way. The correlation edge is a one-way correlation, and the initial weight is assigned a value of 0.6. In this embodiment, the operation logic correlation is a manually set control relationship of the equipment. Affected by the control program and operation and maintenance, the data synchronous change rate (≥85%) is slightly lower than that of the physical connection correlation. Therefore, a value of 0.6 is assigned to conform to the correlation strength of actual operation. Through the above steps, relationships are constructed based on prior physical and logical knowledge, providing a basic logical framework for the model, ensuring the rationality of the relationships, and avoiding meaningless relationships generated by pure data mining.

[0027] Considering that the correlations between some data cannot be directly determined through prior knowledge, it is necessary to mine the true temporal and causal relationships through historical data to improve the comprehensiveness and accuracy of the model. Therefore, the data modeling module also includes posterior statistical correlations based on temporal correlation analysis and causal discovery algorithms for historical monitoring data. Specifically, when constructing the posterior statistical correlations, Granger causality tests are performed on the historical monitoring data, and the corresponding test statistics are... Data pairs whose values ​​are less than a preset significance level and whose calculated causal relationship coefficients exceed a preset strength threshold are identified as association edges with a time-series influence relationship. The causal relationship coefficients are then quantified as the initial association weights of these association edges. The specific implementation is as follows: The posterior statistical association was determined using the Granger causality test algorithm. Historical monitoring data with a time interval of 1 minute was selected, with a data length of no less than 3 months. The 1-minute time interval conforms to the temporal variation characteristics of the operating status of equipment in the park and can capture short-term temporal associations of the data. The data length of ≥3 months can cover different operating conditions in the park (weekdays / rest days, summer / winter), avoiding the one-sidedness of the data sample. The test results showed that the test results of the 3-month data sample were highly consistent with those of the 12-month data sample. First, the data is tested for stationarity using the ADF test. When the ADF test... When the value is <0.05, the data is considered stationary, following the general statistical standards for time series analysis. (ADF test) When the value is <0.05, the data can be determined to be a stationary series at a 95% confidence level, avoiding spurious regression in Granger causality tests caused by non-stationary data; After stabilization, Granger causality tests were conducted with a preset significance level of 0.05, which is a commonly used statistical significance level. This level can exclude spurious causal relationships caused by random factors at a 95% confidence level, ensuring that the discovered causal relationships are genuine associations between data. The threshold for causal relationship strength was 0.5. The causal relationship coefficient was calculated using the partial correlation coefficient, with a value ranging from 0 to 1. The larger the coefficient, the stronger the causal relationship. When the threshold for causal relationship strength is <0.5, the time-series influence is weak and has no practical control value. During the test, the Granger causality test statistic for each pair of data is calculated. Value, when When the value is less than 0.05, a Granger causal relationship is determined between the two sets of data. The causal coefficient is then calculated, reflecting the strength of the causal influence. When the causal coefficient is greater than 0.5, the data pair is identified as a valid association edge, and the causal coefficient is quantified as the initial association weight of that edge. The larger the coefficient, the higher the weight. For example, this can be used to examine historical data on the energy consumption of the central air conditioning system in a park and the ambient temperature. The value is 0.02, which is less than 0.05. The causal relationship coefficient is 0.72, which is greater than 0.5. A one-way association edge is established between ambient temperature and central air conditioning energy consumption, with an initial weight of 0.72. By employing the aforementioned techniques, Granger causality tests are used to uncover the true statistical associations between data, supplementing the deficiencies of prior associations, constructing a comprehensive and quantitative association model, and improving the accuracy of model association predictions.

[0028] Considering that high-confidence anomalies are genuine operational anomalies, cross-system collaborative control must be implemented immediately to achieve low-carbon optimization and fault handling, and to prevent the anomaly from escalating. Therefore, the high-confidence execution strategy is as follows: when the event confidence score is higher than the first threshold, it is determined to be a high-confidence genuine event. Then, based on the event type and the aforementioned data relationship model, a pre-set cross-system collaborative control strategy library is invoked, and an optimized control instruction set is sent to the relevant subsystems. The specific implementation is as follows: When the confidence score of an event exceeds the first threshold of 0.8, the system determines it as a high-confidence real event. Based on the type of abnormal event, such as abnormal energy consumption, equipment failure, or environmental exceedance, the system matches the corresponding strategy from the pre-set cross-system collaborative control strategy library in conjunction with the data association model. For example, for an abnormal increase in central air conditioning energy consumption, the system matches the air conditioning-water pump-fan collaborative load reduction strategy; for an abnormal exceedance of equipment current, the system matches the equipment shutdown-auxiliary system linkage shutdown strategy. The system sends an optimized control instruction set to the relevant subsystems. The instruction set contains the operating parameter adjustment values ​​and start / stop instructions of each subsystem, realizing cross-system collaborative control to reduce park energy consumption or handle equipment failures. Through the above strategies, precise cross-system collaborative control is implemented for high-confidence anomalies, enabling rapid low-carbon optimization and fault handling, and improving the efficiency of park management.

[0029] Since low-confidence anomalies are mostly single-point data source failures, it is necessary to intercept the linkage commands derived from false alarms to avoid energy waste and equipment risks caused by erroneous control. Therefore, the data false alarm interception strategy is as follows: when the event confidence score is lower than the second threshold and the original abnormal value of the main monitoring data exceeds the safety threshold, it is judged as a potential single-point data source failure, a maintenance check work order is generated for the data source device, and any linkage control commands triggered by this anomaly are prevented from being executed, maintaining the existing control logic unchanged. The specific implementation is as follows: When the event confidence score is below the second threshold of 0.3, and the original outlier value of the main monitoring data exceeds the safety threshold, the system determines it as a potential single-point data source failure. The system automatically generates a maintenance and verification work order for the data source device, which includes the device number, installation location, abnormal data type, and verification requirements, and pushes it to the operation and maintenance terminal. At the same time, the system prevents the execution of all linkage control commands derived from this abnormal event, maintaining the original control logic of each subsystem unchanged and avoiding erroneous control. It should be further noted that the safety threshold of the main monitoring data is determined based on the rated parameters in the equipment product manual and the "Code for Construction and Quality Acceptance of Industrial Automation Instrumentation Engineering" (GB50093-2013). For example, the safety threshold for motor operating current is 1.2 times the rated current, and the safety threshold for ambient CO2 concentration is 1500ppm, ensuring that a potential failure is only determined when the data exceeds the safe operating range of the equipment / environment, avoiding unnecessary work order generation under normal fluctuations. By employing the above strategies, false alarm commands can be effectively intercepted, ineffective linkage control can be avoided, and energy waste and equipment operation risks in the park can be reduced.

[0030] Since abnormal events with medium confidence levels cannot be accurately determined by algorithms, manual review is required to balance control efficiency and decision-making accuracy. Therefore, the pending verification strategy for suspected events is to trigger a manual review process when the event confidence score is between the first and second thresholds. This involves pushing the associated data group, confidence score, and preliminary analysis conclusions to the management terminal and temporarily suspending the issuance of automatic control commands. The specific implementation is as follows: When the event confidence score is between 0.3 and 0.8, the system triggers a manual review process. The system pushes the real-time data sets of the main monitoring data and associated auxiliary monitoring data, the event confidence score, and preliminary anomaly type analysis conclusions to the park management terminal. Simultaneously, the system suspends the issuance of all automatic control commands, awaiting the manual review results. After manual review confirms the anomaly is genuine, the high-confidence execution strategy is manually triggered; after confirming a false alarm, the data false alarm interception strategy is manually triggered. By employing the above strategies, suspected abnormal events are subject to manual review to avoid algorithmic misjudgments and balance control efficiency with decision-making accuracy.

[0031] Considering that simple instruction interception cannot completely block the transmission of false alarm instructions, an instruction latching mechanism is needed to shield false alarm instructions from the execution end and ensure the stability of the control logic. Therefore, the specific logic of the data false alarm interception strategy to prevent the execution of any linkage control instructions triggered by the current abnormal event includes: after the collaborative verification and intelligent decision-making module is started, the system creates and activates an instruction latch flag for each potentially affected actuator; when a false alarm is determined, the collaborative verification and intelligent decision-making module clears the instruction latch flag, so that the controllers of each actuator ignore all control instructions derived from the current abnormal event and continue to execute the instructions output by their original control loops. The specific implementation is as follows: After the collaborative verification and intelligent decision-making module is activated, the system determines the actuators that may be affected by the abnormal event, such as water pumps, fans, and air conditioning units, based on the data association model. The system creates an instruction latch flag for each affected actuator, initially in an active state, and stores the flag in the actuator's local controller. During collaborative verification, the system-generated linkage control commands carry the identifier of the abnormal event and are transmitted to the controllers of each actuator. When the system determines that a data false alarm has occurred, the collaborative verification and intelligent decision-making module sends an instruction latch clearing command to each actuator, deactivating the instruction latch flag. The active state of the instruction latch flag does not exceed a preset time. In this embodiment, the processing time for the collaborative verification and intelligent decision-making module is set to 30 seconds. If no clearing command is received within 30 seconds, the flag bit automatically expires to avoid permanent activation of the flag bit due to system communication failure, which would affect the normal control of the actuator. When the actuator controller receives a control command, it first checks the status of the command latch flag bit. If the flag bit is active, the command enters the latched pending state and is not executed temporarily. If the system determines it to be a high-confidence event and issues a confirmation execution command, the flag bit is deactivated and the command is released for execution. If the system determines it to be a false alarm or no confirmation signal is received within 30 seconds, the flag bit remains active or automatically expires, and the controller discards the pending command and continues to execute the original control loop command. By employing the aforementioned technical means, the instruction latching mechanism is used to shield false alarm instructions from the execution end, completely blocking invalid linkage control and ensuring the stable operation of the original control logic of each subsystem.

[0032] Considering that different vendors' subsystems use proprietary communication protocols and have inconsistent data packet formats, making direct fusion analysis impossible, an edge gateway is needed to implement protocol parsing and standardized data mapping. Therefore, the multi-source data acquisition module includes an edge gateway layer deployed on-site. This edge gateway layer is configured to: use built-in multi-protocol parsing engines to parse and uniformly map proprietary protocol data packets from different vendors' subsystems into standardized information model objects in real time. Each information model object includes at least the data value, physical unit, acquisition timestamp, device spatial location code, and subsystem identifier. The specific implementation is as follows: The edge gateway layer incorporates multiple protocol parsing engines, including Modbus RTU, Modbus TCP, BACnet, OPCUA, and proprietary serial port protocols, covering the communication protocols of mainstream vendor subsystems within the campus. After receiving proprietary protocol data packets from various subsystems, the edge gateway uses the corresponding parsing engine to parse the packets in real time, extracting information such as raw data values, physical quantity types, and device identifiers. Subsequently, the parsed data is uniformly mapped into standardized information model objects. These information model objects include data values, physical units, acquisition timestamps, device spatial location codes, and subsystem identifiers. The standardized data is then uploaded to the cloud data platform via the MQTT protocol. By employing the aforementioned technical means, protocol barriers between different vendor subsystems are eliminated, achieving standardization and unification of multi-source heterogeneous data, and providing a unified data carrier for subsequent data modeling and analysis.

[0033] Considering that hardware degradation at monitoring nodes can lead to decreased data quality and unreliable data can interfere with collaborative verification results, it is necessary to monitor data quality in real time and dynamically adjust node weights and logical consistency outputs. Therefore, the hardware health coefficient... The real-time determination methods include: the system continuously monitors the data quality indicators of each monitoring node, including signal amplitude stability, noise spectrum characteristics, and the proportion of invalid data frames; when the data quality indicator of a node is detected to be continuously deteriorating and falling below the health threshold, at least one of the following operations is automatically performed: the association weight of the node is adjusted. Set to zero, or set its logical consistency function. The output is forced to be zero, and the specific implementation is as follows: The system continuously collects raw data from each monitoring node, monitoring three types of data quality indicators: signal amplitude stability (calculating the coefficient of variation of signal amplitude per unit time); noise spectrum characteristics (extracting high-frequency noise components of the signal and calculating the noise power ratio); and invalid data frame ratio (statistically calculating the proportion of invalid data frames due to transmission failures or verification errors to the total number of data frames per unit time). The system presets data quality health thresholds: signal amplitude variation coefficient not greater than 0.1, noise power ratio not greater than 5%, and invalid data frame ratio not greater than 1%. The signal amplitude variation coefficient is the ratio of the signal standard deviation to the mean. This is achieved by statistically analyzing the normal operating conditions within the park. When the coefficient of variation of the monitoring node data is greater than 0.1, it indicates that the signal amplitude fluctuates drastically and the data reliability is significantly reduced. Using Fast Fourier Transform (FFT) to extract the noise spectrum of the signal, the noise power ratio of normal monitoring nodes is ≤5%. When the ratio is greater than 5%, it indicates that there is serious noise interference in the sensor or transmission link, and the true data value is masked. Following the "Industrial Communication Network Fieldbus Specification" (GB / T20501-2019), the proportion of invalid data frames in normal industrial communication is ≤1%. When the proportion is greater than 1%, it indicates a communication link failure or abnormal gateway acquisition, and data integrity cannot be guaranteed. If any metric of a node remains below the health threshold for three consecutive data collection periods, it is considered a data quality degradation; the system automatically performs an adjustment operation, changing the node's correlation weight. Set to zero, or force its logical consistency function. The output is set to zero to eliminate interference from the node's data on collaborative verification. The three-cycle sampling period is set because it matches the gateway's sampling frequency of 500 milliseconds; three cycles equate to 1.5 seconds. Tests show this cycle effectively filters out instantaneous signal fluctuations, such as single data anomalies caused by electromagnetic interference, preventing misjudgments of node health status and ensuring timely identification of hardware degradation. When the system detects that the node's data quality indicators have recovered to above the health threshold for 10 consecutive sampling cycles, the system automatically cancels the forced zeroing operation and gradually restores its weight in the correlation model, achieving data self-healing and logical reconstruction of the monitoring network. By employing the aforementioned technical means, the data quality of nodes can be monitored in real time, interference from unreliable data can be dynamically eliminated, the accuracy of event confidence calculation can be improved, and the stability of system decision-making can be ensured.

[0034] Considering that existing low-carbon energy-saving control relies solely on changes in core production capacity energy consumption data, which can easily trigger erroneous auxiliary system shutdown commands, it is necessary to combine auxiliary energy consumption data for collaborative verification to achieve precise energy-saving control. Therefore, the system also includes an energy efficiency optimization control submodule, which is configured to execute the following energy-saving optimization process: Define the first type of key energy consumption data reflecting the core production capacity activities of the park as the main monitoring data and monitor it in real time; when the first type of key energy consumption data undergoes a preset change that conforms to the preset energy consumption optimization mode, activate the collaborative verification and intelligent decision-making module, and select the correlation weight between the first type of key energy consumption data and the data relationship model. The second type of auxiliary energy consumption data exceeding a preset threshold is used as core auxiliary monitoring data for collaborative verification. This second type of auxiliary energy consumption data corresponds to the energy consumption of the auxiliary systems supporting the core production capacity activities. If the verification results indicate that the correlation between the first and second types of energy consumption data is a high-confidence real event, then based on the correlation confirmed by the collaborative verification, an optimization instruction to adjust the operating parameters of the auxiliary system is generated and issued. If the verification results are low-confidence or false alarms, then any direct shutdown or load reduction instruction for the auxiliary system triggered solely by a single change in the first type of key energy consumption data is suppressed or rejected. Specifically, this is implemented as follows: The energy efficiency optimization and control submodule defines the first type of key energy consumption data reflecting the core production capacity activities of the park as the main monitoring data, such as the total energy consumption of the main equipment of the production line and the air conditioning energy consumption of the core production area, and monitors them in real time. The preset energy consumption optimization mode is: the core production capacity energy consumption decreases by more than 10% for 5 minutes. When the first type of key energy consumption data changes, the collaborative verification and intelligent decision-making module is activated. In this embodiment, the core production capacity energy consumption decreases by more than 10% for 5 minutes is based on the operating characteristics of the core production capacity of the park. A limited number of energy-saving optimization tests with different decrease durations and decrease magnitudes are carried out. The results show that a 10% decrease in 5 minutes is the critical value for stable load reduction of the core production capacity. Under this change, the energy-saving optimization space of the auxiliary system is the largest, and it will not affect the normal operation of the core production capacity. If the decrease duration is less than 5 minutes or the magnitude is less than 10%, the core production capacity change is an instantaneous fluctuation and has no actual energy-saving optimization value. Based on the data association model, the system selects a second type of auxiliary energy consumption data whose association weight with the first type of key energy consumption data exceeds a preset weight as core auxiliary monitoring data. In this embodiment, the preset weight is 0.4, which is determined by testing the energy-saving optimization effect under three weight thresholds of 0.3, 0.4, and 0.5. The auxiliary energy consumption data selected under the 0.4 threshold has a strong correlation with the core production capacity energy consumption, and the energy-saving optimization instructions are highly effective, covering more auxiliary systems and improving energy efficiency. The second type of auxiliary energy consumption data corresponds to the energy consumption of auxiliary systems, such as cooling water pump energy consumption, fresh air system energy consumption, and lighting system energy consumption. The system performs collaborative verification of the primary and auxiliary energy consumption data and calculates the event confidence score. If the verification result is a high-confidence real event (score > 0.8), the system generates auxiliary system operating parameter adjustment instructions based on the confirmed association relationship, such as adjusting the water pump frequency, reducing the fresh air volume, and dimming the lighting brightness, and sends them to the corresponding subsystems. If the verification result is low confidence or false alarm (score < 0.3), the system suppresses or rejects the auxiliary system's direct shutdown or load reduction command triggered solely by a single change in the first type of key energy consumption data, in order to avoid erroneous control. Through the above steps, precise low-carbon and energy-saving control is achieved, with optimization instructions issued only when there are high-confidence correlation changes, suppressing erroneous control triggered by single data anomalies, and reducing the energy consumption of auxiliary systems while ensuring the operation of the core production capacity of the park.

[0035] Considering that the primary monitoring data needs to reflect the core operating status, and the auxiliary monitoring data needs to provide corroboration for the core status, it is necessary to distinguish between different types of primary and auxiliary monitoring data to adapt to the monitoring needs of multiple scenarios. Therefore, the primary monitoring data includes physical quantities that reflect the thermal state, force state, or substance concentration of the equipment, while the auxiliary monitoring data includes physical or digital quantities that reflect electrical parameters, mechanical state, or logic switching quantities. The specific implementation is as follows: The primary monitoring data selects physical quantities reflecting the core physical state of the equipment. Equipment thermal status includes shell temperature, medium temperature, and heat dissipation temperature; force status includes pipeline pressure, equipment vibration amplitude, and mechanical stress; and substance concentration includes CO2 concentration, dust concentration, and harmful gas concentration in the industrial park. Auxiliary monitoring data selects physical or digital quantities reflecting the supporting operating status. Electrical parameters include equipment operating current, voltage, power, and power factor; mechanical status includes bearing speed, valve opening, and mechanical displacement; and logic switching quantities include equipment start / stop status, valve on / off status, and circuit on / off status. For example, if the equipment thermal status is abnormal, such as a shell temperature of 85℃ as the primary monitoring data, the equipment operating current and start / stop switching quantities could be used as auxiliary monitoring data, such as a current of 75A and an on / off status of "running." Similarly, if the substance concentration is abnormal, such as a CO2 concentration of 1600ppm as the primary monitoring data, the fresh air system airflow and fan start / stop status could be used as auxiliary monitoring data, such as an airflow of 2000m³ / h and a fan that is off. Through the above technical means, the classification of primary and secondary monitoring data covers multi-dimensional operating parameters, adapts to the monitoring needs of different types of equipment and scenarios in the park, and improves the system's versatility and scenario adaptability.

[0036] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A smart park management system based on low-carbon and environmental protection, characterized in that, include: The multi-source data acquisition module is used to acquire monitoring data from multiple heterogeneous subsystems within the park in real time. The monitoring data includes the operating status and energy consumption parameter data from at least two subsystems in energy management, environmental monitoring, and equipment operation and maintenance. The data modeling module constructs a data relationship model that reflects the causal or temporal correlation between different monitoring data based on the physical connection relationship of the equipment, the operation logic rules and historical monitoring data, and optimizes and updates the model based on new data and decision feedback. The collaborative verification and intelligent decision-making module is used to perform the following operations: S1: When any monitoring data defined as the main monitoring data becomes abnormal, at least one auxiliary monitoring data with a preset correlation relationship with it is retrieved according to the data relationship model. S2: Based on the abnormal characteristics of the main monitoring data and the real-time status of at least one auxiliary monitoring data, cross-validation is performed to calculate the event confidence score that characterizes the overall credibility of the abnormal event. S3: Based on the preset interval to which the event confidence score belongs, match the corresponding decision strategy, generate and execute control instructions associated with the decision strategy.

2. The smart park management system based on low-carbon and environmental protection according to claim 1, characterized in that, The data modeling module is used to construct two types of relationships, including: Based on the prior association between the physical connection of the equipment and the operation control logic; Based on the posterior statistical associations mined by time-series correlation analysis and causal discovery algorithms on historical monitoring data; In constructing the posterior statistical association, a Granger causality test is performed on historical monitoring data, and the test statistic is then used to determine the causality of the data. Data pairs whose values ​​are less than a preset significance level and whose calculated causal relationship coefficients exceed a preset strength threshold are identified as associated edges with a temporal influence relationship, and the causal relationship coefficients are quantified as the initial association weights of the associated edges.

3. The smart park management system based on low-carbon and environmental protection according to claim 1, characterized in that, In step S2, the event confidence score is calculated. The method is as follows: ,in, The basic weighting coefficients for the main monitoring data; Based on the abnormal deviation of the main monitoring data The calculated basic anomaly contribution value, which is related to There is a positive correlation; For the first The association weight of each auxiliary monitoring data point in the data relationship model; This is a logical consistency function, whose input is the abnormal status of the main monitoring data. With the Current status of auxiliary monitoring data The output is a support score for anomalies in the main monitoring data; when When the direction of change is consistent with the direction expected based on the data relationship model, Output a positive value, otherwise output zero or a negative value; For the first The hardware health coefficient of each monitoring node is calculated in real time or updated periodically based on the node's historical failure rate, signal noise level, or communication stability. This represents the total number of auxiliary monitoring data used in this collaborative verification.

4. The smart park management system based on low-carbon and environmental protection according to claim 1, characterized in that: The preset decision-making strategies in step S3 include at least the following: High confidence execution strategy: When the event confidence score is higher than the first threshold, it is determined to be a high confidence real event. Then, based on the event type and combined with the data relationship model, the pre-set cross-system collaborative control strategy library is invoked, and an optimized control instruction set is sent to the relevant subsystems. Data false alarm interception strategy: When the event confidence score is lower than the second threshold and the original abnormal value of the main monitoring data exceeds the safety threshold, it is judged as a potential single point of failure of the data source, a maintenance check work order for the data source device is generated, and any linkage control instructions triggered by this abnormal event are prevented from being executed, and the existing control logic remains unchanged. Suspected event verification strategy: When the event confidence score is between the first threshold and the second threshold, the manual review process is triggered, the associated data group, confidence score and preliminary analysis conclusions are pushed to the management terminal, and the issuance of automatic control commands is suspended.

5. The smart park management system based on low-carbon and environmental protection according to claim 4, characterized in that, The specific logic in the data false alarm interception strategy to prevent the execution of any linkage control commands triggered by this abnormal event includes: After the collaborative verification and intelligent decision-making module is started, the system creates and activates an instruction latch flag for each potentially affected actuator. When a false alarm is detected, the collaborative verification and intelligent decision-making module clears the instruction latch flag, so that the controllers of each actuator ignore all control instructions derived from this abnormal event and continue to execute the instructions output by their original control loops.

6. The smart park management system based on low-carbon and environmental protection according to claim 1, characterized in that: The multi-source data acquisition module includes an edge gateway layer deployed on-site, and the edge gateway layer is configured as follows: Through built-in multiple protocol parsing engines, private protocol data packets from different vendor subsystems are parsed in real time and uniformly mapped into standardized information model objects. The information model objects include at least data values, physical units, collection timestamps, device spatial location codes, and the identifier of the subsystem to which they belong.

7. The smart park management system based on low-carbon and environmental protection according to claim 3, characterized in that: The hardware health coefficient The real-time determination methods include: The system continuously monitors the data quality indicators of each monitoring node, including signal amplitude stability, noise spectrum characteristics, and the proportion of invalid data frames. When the data quality indicator of a node continuously deteriorates and falls below the health threshold, the system automatically performs at least one of the following operations: adjusts the association weight of the node. Set to zero, or set its logical consistency function. The output is forced to be zero.

8. The smart park management system based on low-carbon and environmental protection according to claim 1, characterized in that, The system also includes an energy efficiency optimization control submodule, which is configured to perform the following energy-saving optimization process: The first category of key energy consumption data, which reflects the core production capacity activities of the park, is defined as the main monitoring data and monitored in real time. When the first type of key energy consumption data undergoes a preset change that conforms to the preset energy consumption optimization mode, the collaborative verification and intelligent decision-making module is activated. Based on the data relationship model, the second type of auxiliary energy consumption data with a correlation weight exceeding a preset threshold with the first type of key energy consumption data is selected as the core auxiliary monitoring data for collaborative verification. The second type of auxiliary energy consumption data corresponds to the energy consumption of auxiliary systems that support the core production capacity activities; If the verification results indicate that the correlation between the first type and the second type of energy consumption data is a high-confidence real event, then based on the correlation confirmed by the collaborative verification, an optimization instruction to adjust the operating parameters of the auxiliary system is generated and issued. If the verification result is low confidence or a false alarm, then any direct shutdown or load reduction command to the auxiliary system triggered solely based on a single change in the first type of key energy consumption data shall be suppressed or rejected.

9. The smart park management system based on low-carbon and environmental protection according to claim 1, characterized in that, The main monitoring data includes physical quantities such as the thermal state, force state, or substance concentration of the reaction equipment, while the auxiliary monitoring data includes physical or digital quantities reflecting electrical parameters, mechanical state, or logic switching quantities.

Images