Port air quality monitoring method and system based on multi-source data analysis

Abstract

The application provides a port air quality monitoring method and system based on multi-source data analysis, obtains a port multi-source monitoring data set containing port gas, meteorological and operation activity monitoring data; then performs cross-type data correlation mapping processing on the port multi-source monitoring data set, obtains a multi-source data correlation mapping table, mines abnormal correlation patterns, and identifies correlation combinations exceeding a preset threshold; in combination with historical pollution correlation records, trace analysis is performed on the abnormal correlation pattern set to generate pollution trace clues; then, pollution diffusion trend description is derived, a monitoring response instruction is generated and sent to a terminal in combination with the current monitoring state. The application comprehensively integrates multi-source data, deeply mines correlation patterns, accurately traces pollution sources, early warns pollution events, and effectively improves the accuracy and timeliness of port air quality monitoring.

Description

Technical Field

[0001] This invention relates to the field of air quality monitoring, and more specifically, to a port air quality monitoring method and system based on multi-source data analysis. Background Technology

[0002] During port operations, air quality is influenced by a complex array of factors. Traditional port air quality monitoring methods are often limited to a single data source, such as relying solely on port gas monitoring equipment to obtain gas composition data, or focusing only on the general impact of meteorological data on air quality. However, the port environment is a highly dynamic and interconnected system; changes in gas composition are not only related to meteorological conditions but also closely linked to port operations.

[0003] Existing monitoring methods are insufficient to comprehensively consider the complex relationships between these multi-source factors. For example, relying solely on gas monitoring data makes it difficult to determine whether abnormal gas concentrations stem from changes in meteorological diffusion conditions or from malfunctions in port equipment. Furthermore, the lack of effective source tracing analysis methods in the face of abnormal air quality makes it impossible to quickly and accurately identify pollution sources and diffusion trends, hindering the development of timely and effective countermeasures and consequently impacting normal port operations and the surrounding environmental quality. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a port air quality monitoring method based on multi-source data analysis, the method comprising: A multi-source monitoring data set for the port is obtained, which includes port gas monitoring data, port meteorological monitoring data, and port operation activity monitoring data. The port gas monitoring data records gas composition monitoring information at different monitoring points in the port, the port meteorological monitoring data records meteorological status monitoring information for the port area, and the port operation activity monitoring data records operational status monitoring information for port operation equipment. Cross-type data association mapping processing is performed on the port multi-source monitoring data set to establish the association relationship between the port gas monitoring data and the port meteorological monitoring data, and at the same time, the association relationship between the port gas monitoring data and the port operation activity monitoring data is established to obtain a multi-source data association mapping table. Based on the multi-source data association mapping table, abnormal association patterns of port air quality are mined, and association combinations that exceed the preset association threshold range are identified to obtain an abnormal association pattern set. By combining historical pollution correlation records of the port, the abnormal correlation pattern set is analyzed for source tracing, and the feature information associated with historical pollution events in the abnormal correlation pattern set is extracted to generate source tracing clues for port air quality pollution. Based on the source tracing clues of port air quality pollution, a description of pollution diffusion trend is derived, the description of pollution diffusion trend is integrated with the current port air quality monitoring status, a port air quality monitoring response command is generated, and the port air quality monitoring response command is sent to the port monitoring and control terminal. In another aspect, embodiments of the present invention also provide a port air quality monitoring system based on multi-source data analysis, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0005] Based on the above, this invention acquires a multi-source monitoring data set of a port, covering port gas, meteorological, and operational activity monitoring data. It then performs cross-type correlation mapping processing on this multi-source monitoring data set, establishing correlations between gas and meteorological data, and between gas and operational activities, forming a multi-source data correlation mapping table. This allows for in-depth analysis of the intrinsic connections between various factors. Based on this multi-source data correlation mapping table, abnormal correlation patterns are mined, and correlation combinations exceeding preset thresholds are identified, enabling timely detection of potential air quality anomalies. By combining historical pollution correlation records for source tracing analysis, feature information associated with historical events is extracted, generating pollution source tracing clues, enabling rapid and accurate location of pollution sources. Finally, a pollution diffusion trend description is derived and integrated with the current monitoring status to generate response commands, which are sent to the port monitoring and control terminal, effectively improving the accuracy, timeliness, and comprehensiveness of port air quality monitoring. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the execution flow of the port air quality monitoring method based on multi-source data analysis provided in the embodiments of the present invention.

[0007] Figure 2 This is a schematic diagram of exemplary hardware and software components of a port air quality monitoring system based on multi-source data analysis provided in an embodiment of the present invention. Detailed Implementation

[0008] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a port air quality monitoring method based on multi-source data analysis, provided in one embodiment of the present invention. The following is a detailed description of this port air quality monitoring method based on multi-source data analysis.

[0009] Step S110: Obtain a multi-source monitoring data set for the port. The multi-source monitoring data set for the port includes port gas monitoring data, port meteorological monitoring data, and port operation activity monitoring data. The port gas monitoring data records gas composition monitoring information at different monitoring points in the port. The port meteorological monitoring data records meteorological status monitoring information for the port area. The port operation activity monitoring data records operational status monitoring information for port operation equipment.

[0010] In this embodiment, port gas monitoring data is collected through a network of gas sensors deployed in various key areas of the port. These sensors are distributed across different functional areas such as the wharf front, storage yards, transport channels, and storage areas. Each monitoring point is equipped with a multi-parameter gas sensor, capable of simultaneously monitoring the concentration of multiple gaseous components, including sulfur dioxide, nitrogen oxides, particulate matter, and volatile organic compounds. The sampling frequency of the sensors is set according to monitoring requirements. The monitoring data is transmitted to the data center in real time via a wireless transmission module. During data transmission, an encrypted transmission protocol is used to encrypt the data to prevent it from being stolen or tampered with. Simultaneously, the sensors are calibrated regularly to ensure the accuracy of the monitoring data.

[0011] Port meteorological monitoring data is provided by multiple automatic weather stations located within the port area. These stations are deployed according to meteorological observation standards and can monitor meteorological elements such as wind direction, wind speed, temperature, humidity, air pressure, precipitation, and visibility in real time. The collection and transmission of meteorological data follow relevant meteorological data standards, and the data undergoes quality control before storage to remove outliers and invalid data.

[0012] Port operation monitoring data is acquired through integration with the port's production management system and equipment monitoring system, covering the operational status information of various operating equipment such as container cranes, gantry cranes, loading and unloading bridges, forklifts, tugboats, and transport trucks. Specifically, this includes data such as real-time equipment location, operating speed, workload, energy consumption, start-up and shutdown times, and fault alarms. During data collection, only parameters related to equipment operating status are acquired, avoiding the acquisition of operators' personal identification information or other privacy-related data. Data that may involve privacy is anonymized to ensure that data collection complies with relevant laws and regulations on data security and privacy protection.

[0013] Step S120: Perform cross-type data association mapping processing on the port multi-source monitoring data set, establish the association relationship between the port gas monitoring data and the port meteorological monitoring data, and at the same time establish the association relationship between the port gas monitoring data and the port operation activity monitoring data, to obtain a multi-source data association mapping table.

[0014] Step S121: Time synchronization and coordination are performed on the port gas monitoring data, port meteorological monitoring data and port operation activity monitoring data in the port multi-source monitoring data set so that different types of monitoring data are matched at the same time node.

[0015] Because port gas monitoring data, meteorological monitoring data, and operational activity monitoring data come from different acquisition systems, their timestamp generation mechanisms may differ, leading to inconsistencies in the time granularity and start time of the data. To achieve effective correlation across data types, time synchronization and coordination are first required. Using the timestamp of the port gas monitoring data as a benchmark, the timestamps of the meteorological and operational activity monitoring data are uniformly converted to the same time format and precision as the gas monitoring data. For data with incomplete timestamp matching, linear interpolation is used to fill in the data. For example, if the time interval of the gas monitoring data is one minute, while the time interval of the meteorological monitoring data is five minutes, the five minutes of meteorological data are evenly distributed across five one-minute time nodes, ensuring that there are corresponding meteorological and operational activity data at each one-minute time node. For discrete event records in the operational activity monitoring data, such as equipment start-up and shutdown operations, they are mapped to specific time nodes, and the operating status of the equipment is marked in the time period before and after that time node.

[0016] Step S122: Determine the core correlation items for cross-type data association. The core correlation items include the correlation items between the concentration of a specific gas component in the port gas monitoring data and the temperature and humidity parameters in the port meteorological monitoring data, the correlation items between the concentration of a specific gas component in the port gas monitoring data and the operating time of the operating equipment in the port operation activity monitoring data, and the correlation items between the wind force parameters in the port meteorological monitoring data and the operating power of the operating equipment in the port operation activity monitoring data.

[0017] After time synchronization was completed, core correlation terms for cross-type data were determined based on the objectives of port air quality monitoring and the influencing factors of pollutant generation and diffusion. The concentration of specific gas components in port gas monitoring data is a direct indicator of air quality, while temperature and humidity are important meteorological factors affecting gas diffusion and chemical reactions; therefore, the concentration of specific gas components and temperature / humidity parameters were identified as one core correlation term. Port operations are one of the main sources of pollutants, and the operating time of operating equipment directly affects pollutant emissions; therefore, the concentration of specific gas components and the operating time of operating equipment were identified as another core correlation term. Wind parameters (wind direction and speed) affect the direction and speed of pollutant diffusion, and the operating power of operating equipment is closely related to pollutant emissions; different power levels result in different emission characteristics. Therefore, wind parameters and the operating power of operating equipment were identified as a third core correlation term. The determination of these core correlation terms is based on an in-depth analysis of the factors influencing port air quality and can comprehensively reflect the inherent connections between different types of data.

[0018] Step S123: Set a correlation representation method for each core correlation item, and use numerical mapping to represent the correlation degree between different types of monitoring data. The correlation degree is derived by the consistency of the change trend of different types of monitoring data at the same time node.

[0019] For each core correlation item, a specific correlation representation method needs to be established to quantify the degree of correlation between different types of monitoring data. A numerical mapping method is used to represent the degree of correlation with a single value, the magnitude of which reflects the strength of the correlation. The derivation of the correlation degree is based on the consistency of the changing trends of different types of monitoring data at the same time point. For example, for the correlation item between a specific gas component concentration and the operating time of the equipment, the changing trends of the gas component concentration and the operating time of the equipment are observed within the same time series. If the gas component concentration increases as the operating time of the equipment increases, it indicates a strong positive correlation; if the concentration decreases as the operating time increases, there may be a negative correlation or other interfering factors. The numerical value of the correlation degree is derived through the analysis of the consistency of the above changing trends.

[0020] Step S124: Based on the set association representation method, calculate the association degree value of each core association item and record the association degree value corresponding to each core association item.

[0021] Step S1241: For each core correlation item, extract the numerical sequence of the two types of monitoring data corresponding to the same time node, wherein the numerical sequence contains the monitoring data values ​​of multiple consecutive time nodes.

[0022] For each core correlation item, values ​​of the two types of monitoring data at the same time point are extracted from the multi-source monitoring data set after time synchronization processing, forming two numerical sequences. For example, for the correlation item between the concentration of a specific gas component and the operating time of the equipment, the concentration values ​​of the specific gas component at multiple consecutive time points are extracted to form a concentration numerical sequence; at the same time point, the operating time values ​​of the equipment are extracted to form an operating time numerical sequence. The length of the above numerical sequences is determined according to the time span of the monitoring data and the analysis requirements to ensure that a sufficient number of sample points are included to reflect the changing trend of the data.

[0023] Step S1242: Analyze the changing trends of the two types of monitoring data numerical sequences, determine whether the increasing and decreasing directions of the two types of numerical sequences are consistent, calculate the matching situation of the increasing and decreasing magnitudes, and derive the consistency coefficient of the changing trends based on the consistency of the increasing and decreasing directions and the matching situation of the increasing and decreasing magnitudes.

[0024] A trend analysis was performed on the two types of extracted numerical sequences. First, the direction of increase or decrease in data at each time point was determined, i.e., whether the value at the next time point increased, decreased, or remained essentially unchanged compared to the previous time point. Then, the proportion of time points in both types of numerical sequences where the direction of increase or decrease was consistent was calculated; a higher proportion indicates better consistency. Next, the magnitude of increase or decrease in both types of numerical sequences at time points where the direction of increase or decrease was consistent was calculated, i.e., (value of the next time point - value of the previous time point) / value of the previous time point. The matching of magnitudes was then analyzed, for example, by calculating the absolute or relative difference in magnitudes; a smaller difference indicates better matching. Combining the consistency ratio of the direction of increase or decrease and the matching of magnitudes, a trend consistency coefficient was derived. This trend consistency coefficient comprehensively reflects the degree of consistency in the trends of the two types of numerical sequences.

[0025] Step S1243: Convert the trend consistency coefficient into a correlation degree value. The representation range of the correlation degree value corresponds to the derivation range of the trend consistency coefficient. Record the correlation degree value corresponding to each core correlation item. At the same time, record the time node range and numerical sequence length on which the correlation degree value is calculated.

[0026] The derived consistency coefficient of the trend is converted into a correlation degree value. The conversion method is determined based on the value range of the consistency coefficient, ensuring that the representation range of the correlation degree value corresponds to the derived range of the consistency coefficient. For example, the value range of the consistency coefficient is linearly mapped to 0-1, where 0 represents no correlation and 1 represents complete correlation. When recording the correlation degree value corresponding to each core correlation item, the time node range and the length of the numerical sequence on which the value was calculated are also recorded to facilitate subsequent evaluation and verification of the reliability of the correlation degree value. For example, if the numerical sequence length of a certain correlation item is too short, it may reduce the credibility of the correlation degree value, which needs to be considered in subsequent analysis.

[0027] Step S125: Determine the row and column identifiers of the multi-source data association mapping table. The row identifiers correspond to the type classification of port multi-source monitoring data, and the column identifiers correspond to the classification of core association items.

[0028] The row identifiers of the multi-source data association mapping table are determined according to the types of port multi-source monitoring data, divided into rows for port gas monitoring data, port meteorological monitoring data, and port operation activity monitoring data, with each row representing one type of monitoring data. The column identifiers correspond to the categories of core association items: columns for association items between specific gas component concentration and temperature and humidity parameters, columns for association items between specific gas component concentration and operating equipment runtime, and columns for association items between wind force parameters and operating equipment power; each column represents a core association item. Through clear row and column identifiers, the structure of the multi-source data association mapping table is clear, facilitating data retrieval and analysis.

[0029] Step S126: Fill the calculated correlation degree values ​​of each core correlation item into the corresponding positions of the multi-source data correlation mapping table according to the correspondence between row identifier and column identifier to form the initial multi-source data correlation mapping table.

[0030] Based on the correspondence between row and column identifiers, the correlation values ​​of each core correlation item calculated in step S124 are filled into the corresponding cells of the multi-source data correlation mapping table. For example, the concentration of a specific gas component belongs to port gas monitoring data, and the temperature and humidity parameters belong to port meteorological monitoring data. Their correlation values ​​are filled into the intersection cell of the port gas monitoring data row and the column for the correlation of specific gas component concentration and temperature and humidity parameters. This process is repeated until all the correlation values ​​of the core correlation items are filled in, forming the initial multi-source data correlation mapping table. The initial mapping table visually displays the degree of correlation established between different types of monitoring data through the core correlation items.

[0031] Step S127: Supplement the indirect relationships between each core related item, calculate the association transmission coefficient between different core related items, and integrate the association transmission coefficient into the initial multi-source data association mapping table to form a complete multi-source data association mapping table.

[0032] Besides direct core correlations, indirect correlations may also exist between different core correlations. For example, the correlation between the concentration of a specific gas component and the operating time of the equipment, and the correlation between the concentration of a specific gas component and temperature and humidity parameters, both involve the concentration of the specific gas component, thus indicating an indirect correlation. These indirect correlations are characterized by calculating the correlation transmission coefficient between different core correlations. The calculation of the correlation transmission coefficient is based on the numerical value of the correlation degree of each core correlation and the inherent relationship between the data types they involve. For example, the correlation transmission coefficient is derived by analyzing the role and influence of the commonly involved data types in different correlations. The calculated correlation transmission coefficients are then incorporated into the initial multi-source data correlation mapping table as additional columns or rows, or by using composite values ​​in the cells of the mapping table to represent the direct correlation degree and the indirect correlation transmission coefficient, thus forming a complete multi-source data correlation mapping table that can more comprehensively reflect the complex correlations between multi-source monitoring data.

[0033] Step S130: Based on the multi-source data association mapping table, mine abnormal association patterns of port air quality, identify association combinations that exceed the preset association threshold range, and obtain an abnormal association pattern set.

[0034] Step S131: Determine the association dimensions in the multi-source data association mapping table. The association dimensions include the association dimension between the gas composition in the port gas monitoring data and the meteorological state in the port meteorological monitoring data, the association dimension between the gas composition in the port gas monitoring data and the operating status of the operating equipment in the port operation activity monitoring data, and the association dimension between the meteorological state in the port meteorological monitoring data and the operating status of the operating equipment in the port operation activity monitoring data.

[0035] The correlation dimensions in the multi-source data correlation mapping table are determined based on the interaction relationships between data types. Gas composition in port gas monitoring data directly reflects air quality; meteorological conditions affect the diffusion and transformation of gas components, thus establishing a correlation dimension between gas composition and meteorological conditions. The operating status of operational equipment directly determines pollutant emissions, thereby affecting gas component concentrations, thus establishing a correlation dimension between gas composition and operational equipment operating status. Meteorological conditions also influence operational activities; for example, severe weather may alter the operating power or methods of operational equipment, and the operation of operational equipment may also have a minor impact on local meteorological conditions, thus establishing a correlation dimension between meteorological conditions and operational equipment operating status. These three correlation dimensions comprehensively cover the pairwise correlations between gas, meteorological, and operational activity data, forming the foundation for uncovering abnormal correlation patterns.

[0036] Step S132: Assess the correlation strength of each correlation dimension, adjust the assessment priority of each correlation dimension according to the port air quality monitoring requirements, and generate correlation dimension priority allocation results.

[0037] Each correlation dimension is assessed for correlation strength, based on factors such as its explanatory power for air quality anomalies in historical data, data stability and reliability, and its importance in the current monitoring scenario. For example, during peak port operation periods, the operational status of equipment may have a more significant impact on air quality; therefore, the correlation strength between gas composition and equipment operational status is relatively high. The assessment priority of each correlation dimension is adjusted according to the specific needs of port air quality monitoring, such as focusing on the source of a particular pollutant or the impact of meteorological conditions on pollution diffusion. If the current monitoring need is to identify the main emission sources of pollutants, the correlation dimension between gas composition and equipment operational status has the highest priority; if the need is to predict pollution diffusion trends, the correlation dimension between gas composition and meteorological conditions has the highest priority. Through comprehensive assessment and adjustment, a correlation dimension priority allocation result is generated, clarifying the importance of each correlation dimension in the process of anomaly correlation pattern mining.

[0038] Step S133: Perform pattern matching between the multi-source data association mapping table and the preset port normal air quality association pattern library, calculate the matching degree between each association combination in the multi-source data association mapping table and the corresponding association pattern in the port normal air quality association pattern library, and obtain the association pattern matching degree sequence.

[0039] The pre-defined port normal air quality correlation model library is constructed based on long-term, historical, multi-source monitoring data under normal operating conditions. It includes normal correlation models for various correlation dimensions under different seasons, operational intensities, and meteorological conditions. The current multi-source data correlation mapping table is matched with the normal correlation model library. For each correlation combination—that is, a specific correlation item under each correlation dimension, such as the correlation between sulfur dioxide concentration in the gas composition and temperature in the meteorological state, or the correlation with the operating time of the equipment—the matching degree between it and the corresponding correlation model in the normal correlation model library is calculated. The matching degree is calculated by comparing the differences between the numerical value and trend of the current correlation combination and the normal correlation model; the smaller the difference, the higher the matching degree. The matching degrees of all correlation combinations are arranged in order to obtain the correlation model matching degree sequence.

[0040] Step S134: Filter the association combinations in the association pattern matching degree sequence whose matching degree is lower than the preset matching degree threshold, and mark the association combinations whose matching degree is lower than the preset matching degree threshold as candidate abnormal association combinations.

[0041] Based on port air quality management requirements and statistical analysis of historical anomalies, a preset matching degree threshold is set. This threshold is used to determine whether correlation combinations deviate from normal patterns. The correlation pattern matching degree sequence is traversed, and correlation combinations with matching degrees below the preset matching degree threshold are filtered out. These correlation combinations show significant differences from the correlation patterns under normal conditions, potentially indicating air quality anomalies; therefore, they are marked as candidate anomaly correlation combinations. The setting of the preset matching degree threshold needs to comprehensively consider the false positive rate and the false negative rate, ensuring that it can effectively identify potential anomaly correlation combinations without excessively misclassifying normal fluctuations as anomalies.

[0042] Step S135: Analyze the frequency and duration of the candidate anomaly association combinations during the continuous monitoring period, and retain candidate anomaly association combinations whose frequency of occurrence exceeds a preset frequency threshold and whose duration exceeds a preset duration threshold.

[0043] Step S1351: Set the time range for the continuous monitoring period, which is determined based on the regular monitoring cycle of port air quality and the possible duration of pollution events.

[0044] The duration of continuous monitoring is determined based on the port's routine air quality monitoring cycle, such as one day or one week, and the typical duration of historical pollution events. For example, if the routine monitoring cycle at the port is one day, while the duration of a typical pollution event may range from several hours to several days, then a three-day continuous monitoring period would be set to ensure that potential persistent anomalies are captured. The time range should not be too short, lest long-lasting anomalies be missed; nor should it be too long, lest it increase the amount of data processing and the possibility of misjudgment.

[0045] Step S1352: Within the set time range, count the number of times each candidate anomaly association combination appears at different time nodes to obtain the frequency of occurrence of each candidate anomaly association combination.

[0046] Within the set continuous monitoring period, for each candidate anomaly association combination, it is checked whether it is marked as a candidate anomaly at each time point, i.e., whether the matching degree is lower than a preset threshold. The total number of times each candidate anomaly association combination appears throughout the entire time range is counted to obtain the occurrence frequency. The occurrence frequency reflects the frequency of occurrence of the candidate anomaly association combination; the higher the frequency, the more frequently it deviates from the normal pattern, and the more attention it may warrant.

[0047] Step S1353: For each candidate anomaly association combination, identify the time interval in which it occurs continuously within a set time range, where each continuous time interval contains multiple consecutive time nodes.

[0048] For example, step S13531: Obtain all occurrence time nodes of each candidate abnormal association combination within a set time range, sort all occurrence time nodes in chronological order, and form an ordered time node list.

[0049] For each candidate anomaly association combination, all time nodes marked as candidate anomalies within a set time range are extracted from the monitoring data. Then, these time nodes are sorted in chronological order to form an ordered list of time nodes. For example, if a candidate anomaly association combination is marked as anomaly at 8:00, 9:00 on the first day, and 10:00, 11:00, and 12:00 on the second day, the ordered list of time nodes would be [8:00 on the first day, 9:00 on the first day, 10:00 on the second day, 11:00 on the second day, 12:00 on the second day].

[0050] Step S13532: Set a continuous determination interval, which is determined based on the collection interval of port air quality monitoring data.

[0051] The continuous judgment interval is set based on the data collection interval of the port air quality monitoring data. For example, if the data collection interval is 1 hour, the continuous judgment interval can be set to 1 hour. That is, if the interval between two adjacent occurrence time points does not exceed 1 hour, they are considered to have occurred consecutively. If the collection interval is shorter, such as 10 minutes, the continuous judgment interval is also shortened to 10 minutes accordingly to accurately reflect the continuity of the data.

[0052] Step S13533: Check the time interval between two adjacent time nodes in the ordered time node list in turn, and determine whether the time interval is less than or equal to the set continuous judgment interval.

[0053] Following the order of the ordered time node list, the time difference between any two adjacent time nodes is calculated sequentially, i.e., the difference between the previous and subsequent time nodes. This time interval is then compared with the set continuous decision interval to determine whether it is less than or equal to the continuous decision interval.

[0054] Step S13534: If the time interval between two adjacent time nodes is less than or equal to the continuous judgment interval, then these two time nodes are classified into the same continuous time interval; if the time interval between two adjacent time nodes is greater than the continuous judgment interval, then the latter time node is taken as the starting node of the new continuous time interval, and so on, the entire ordered time node list is traversed to divide all the appearing time nodes into different continuous time intervals; each continuous time interval contains at least two consecutively appearing time nodes, and the number of time nodes contained in each continuous time interval and the specific time information of each time node are recorded.

[0055] Based on the comparison between the time interval of adjacent time nodes and the continuous decision interval, the time nodes are divided into continuous time intervals. If the adjacent time interval is less than or equal to the continuous decision interval, the two time nodes are grouped into the same continuous time interval, and the interval between the next time node and the last time node of the current interval is checked. If the adjacent time interval is greater than the continuous decision interval, the current continuous time interval ends, the next time node is taken as the starting node of the new continuous time interval, and the check of subsequent nodes continues. After traversing all time nodes, multiple continuous time intervals are obtained, each interval containing at least two consecutively occurring time nodes. At the same time, the number of time nodes contained in each continuous time interval and the specific time information of each time node, such as the start time and end time, are recorded.

[0056] Step S13535: Organize the divided continuous time intervals, remove continuous time intervals containing fewer than the preset number of time nodes, and retain continuous time intervals that meet the preset requirements.

[0057] To eliminate transient anomalies caused by random factors, a preset threshold is set. This threshold is determined based on the data collection interval and the typical duration characteristics of the anomaly. For example, if the collection interval is 1 hour and the anomaly typically lasts at least 3 hours, then the preset threshold is 3 time points. The divided continuous time intervals are then processed, and each interval is checked to see if the number of time points it contains reaches the preset threshold. If it is less than the threshold, the continuous time interval is considered to be likely caused by random fluctuations and is removed; continuous time intervals containing the number of time points that reaches or exceeds the preset threshold are retained.

[0058] Step S1354: Calculate the time difference between the start and end time nodes of each continuous time interval. This time difference is the duration of the corresponding continuous time interval.

[0059] For each retained continuous time interval, the time difference between its recorded start and end times is calculated, and this time difference is the duration of the continuous time interval. For example, if a continuous time interval starts at 8:00 AM on the first day and ends at 11:00 AM on the first day, with a collection interval of 1 hour and includes 4 time points (8:00, 9:00, 10:00, and 11:00), then the duration of this interval is 3 hours.

[0060] Step S1355: Compare the duration of all continuous time intervals for each candidate anomaly association combination, select the continuous time interval with the longest duration, and determine the duration of the candidate anomaly association combination as the duration of the longest duration.

[0061] Each candidate anomaly association combination may correspond to multiple preserved continuous time intervals. By comparing the durations of these intervals, the longest continuous time interval is identified, and this longest duration is determined as the duration of the candidate anomaly association combination. The duration reflects the longest period during which the anomaly association combination persists and is an important indicator for judging the severity of the anomaly.

[0062] Step S1356: Compare the occurrence frequency of each candidate abnormal association combination with a preset frequency threshold, and compare its duration with a preset duration threshold.

[0063] The preset frequency threshold and preset duration threshold are set based on port air quality safety standards and the characteristics of historical abnormal events. The frequency threshold represents the minimum number of times a candidate anomaly association combination occurs within a set continuous monitoring period; the duration threshold represents the shortest duration for which a candidate anomaly association combination persists. The frequency of occurrence of each candidate anomaly association combination is compared with the preset frequency threshold to determine whether it reaches or exceeds the threshold; simultaneously, its duration is compared with the preset duration threshold to determine whether it reaches or exceeds the threshold.

[0064] Step S1357: Select candidate abnormal association combinations whose occurrence frequency exceeds a preset frequency threshold and whose duration exceeds a preset duration threshold, and use these candidate abnormal association combinations as abnormal association combinations for the association dimension to be supplemented.

[0065] Only when the frequency of a candidate anomaly association combination exceeds a preset frequency threshold, and its duration also exceeds a preset duration threshold, is the anomaly of that association combination considered statistically significant and having a real impact, potentially indicating actual air quality anomalies. Candidate anomaly association combinations that simultaneously meet both conditions are selected as anomaly association combinations for supplementing the association dimension.

[0066] Step S1358: Record the occurrence time of abnormal association combinations of the dimensions to be supplemented, as well as the start and end time of the longest continuous time interval.

[0067] For the abnormal association combinations of the selected dimensions to be supplemented, record in detail all occurrence time nodes within the set time range, as well as the start and end time nodes of the longest continuous time interval. The above information has important reference value for subsequent anomaly analysis and source tracing, and can help determine the specific time period and duration of the anomaly.

[0068] Step S136: Supplement the missing association dimension information in the retained candidate anomaly association combinations to generate a complete anomaly association pattern.

[0069] Since the previously selected candidate anomaly association combinations may only involve partial association dimensions, such as only including anomalies related to gas composition and the operating status of work equipment, lacking information on the association with meteorological conditions, it is necessary to supplement the completeness of the association dimensions in order to comprehensively analyze the causes of the anomalies. By querying multi-source monitoring datasets, monitoring data for other association dimensions of the candidate anomaly association combination at the corresponding time point can be obtained, such as meteorological condition data. The relationship between these data and the current anomaly association combination can be analyzed to complete the missing association dimension information. For example, if a candidate anomaly association combination is an anomaly related to sulfur dioxide concentration and the operating time of a certain work equipment, meteorological data such as wind direction and wind speed for that time period can be supplemented to analyze whether meteorological factors have an impact on the anomaly association, thereby forming a complete anomaly association pattern that includes three association dimensions: gas, operation, and meteorology.

[0070] Step S137: Classify and integrate all complete abnormal association patterns according to the association dimension to obtain an abnormal association pattern set.

[0071] After completing the classification, all abnormal correlation patterns are categorized and integrated according to the correlation dimensions involved. For example, patterns mainly involving abnormalities in the correlation dimension between gas composition and meteorological conditions are grouped into one category, patterns mainly involving abnormalities in the correlation dimension between gas composition and operating equipment status are grouped into another category, and patterns involving abnormalities in all three correlation dimensions are grouped into a third category, and so on. Through classification and integration, a clearly structured and hierarchical set of abnormal correlation patterns is obtained, which facilitates subsequent source tracing analysis and processing.

[0072] Step S140: Combine the port's historical pollution association records to perform source tracing analysis on the set of abnormal association patterns, extract the feature information associated with historical pollution events in the set of abnormal association patterns, and generate source tracing clues for port air quality pollution.

[0073] Step S141: Obtain a set of historical pollution association records for the port. The set of historical pollution association records for the port includes multi-source data association patterns when historical pollution events occurred, pollution source types corresponding to historical pollution events, and diffusion path information of historical pollution events.

[0074] The Port Historical Pollution Association Record Collection is a database formed through detailed investigations and records of past port air quality pollution events. Each historical pollution association record contains multi-source data association patterns during the period in which the pollution event occurred, namely the correlation between data such as gas composition concentration, meteorological conditions, and operational activities at that time; it also records the finally determined pollution source type of the pollution event, such as ship emissions, truck transportation emissions, loading and unloading operations emissions, and storage yard dust; as well as information on the diffusion path of the pollution event, including the order in which pollutants spread from the source to various areas and the scope of impact.

[0075] Step S142: Compare the features of each abnormal association pattern in the abnormal association pattern set with the historical multi-source data association patterns in the port historical pollution association record set, and calculate the feature similarity between each abnormal association pattern and each historical multi-source data association pattern.

[0076] Step S1421: Extract the feature dimensions of each abnormal correlation pattern. The feature dimensions include the gas composition type involved in the abnormal correlation pattern, the meteorological state parameters involved in the abnormal correlation pattern, the operating parameters of the operating equipment involved in the abnormal correlation pattern, and the correlation strength characteristics between the parameters.

[0077] For each anomalous correlation pattern in the set of anomalous correlation patterns, specific feature dimensions are extracted from its three correlation dimensions: gas, meteorology, and operation. Gas composition types include specific pollutant types such as sulfur dioxide, nitrogen oxides, and particulate matter; meteorological state parameters include wind direction, wind speed, temperature, humidity, and air pressure; operational equipment parameters include equipment type, operating time, operating power, and workload; and the correlation strength characteristics between each parameter, i.e., the previously calculated correlation degree values ​​and the consistency of change trends, etc. These feature dimensions comprehensively describe the essential characteristics of the anomalous correlation pattern.

[0078] Step S1422: Extract the corresponding feature dimension of each historical multi-source data association pattern in the port historical pollution association record set, perform characterization processing on the parameters under each feature dimension, convert different types of parameters into a unified characterization form, calculate the degree of parameter difference between the abnormal association pattern and the historical multi-source data association pattern under each feature dimension, and assign corresponding dimension weights to each feature dimension according to the degree of influence of each feature dimension on the port air quality pollution source tracing.

[0079] From the historical pollution association records of the port, the corresponding feature dimensions of each historical multi-source data association pattern are extracted, namely, the gas composition type, meteorological state parameters, operating equipment parameters, and association strength features that are the same as the current abnormal association pattern. Since different types of parameters have different dimensions and representation forms (e.g., gas composition concentration is measured in milligrams per cubic meter, meteorological parameters in degrees Celsius or meters per second, and operating parameters in hours or kilowatts), these parameters need to be processed to convert them into a unified dimensionless representation form, for example, by normalizing the parameter values ​​to the range of 0-1. Then, for each feature dimension, the degree of difference between the parameter values ​​of the current abnormal association pattern and the parameter values ​​of the historical multi-source data association patterns is calculated, such as absolute difference or relative difference. Based on the importance of each feature dimension in pollution source tracing, corresponding dimension weights are assigned to each feature dimension. For example, gas composition type is directly related to the pollution source and has a higher weight; wind direction and wind speed in meteorological state parameters have a significant impact on the diffusion path and also have a higher weight; equipment type in operating equipment parameters has a higher weight, etc.

[0080] Step S1423: Based on the degree of parameter difference and corresponding dimension weight of each feature dimension, calculate the comprehensive feature similarity between the abnormal association pattern and the historical multi-source data association pattern. The comprehensive feature similarity is used to reflect the overall matching situation between the abnormal association pattern and the historical multi-source data association pattern.

[0081] The weighted difference value is obtained by multiplying the parameter difference of each feature dimension by its weight. Then, the weighted difference values ​​of all feature dimensions are combined, for example, by summing or averaging, to obtain a comprehensive difference index. The comprehensive feature similarity is inversely proportional to the comprehensive difference index; the smaller the comprehensive difference index, the higher the comprehensive feature similarity. The comprehensive feature similarity comprehensively reflects the overall matching between the current abnormal association pattern and the multi-source data association pattern of historical pollution events.

[0082] Step S1424: Record the comprehensive feature similarity between each abnormal association pattern and each historical multi-source data association pattern to form a feature similarity comparison table.

[0083] The comprehensive feature similarity between each calculated abnormal association pattern and each historical multi-source data association pattern is recorded according to the numbers of the abnormal association pattern and the historical association pattern, forming a feature similarity comparison table. This feature similarity comparison table shows the degree of similarity between the current abnormal pattern and the historical pattern.

[0084] Step S143: Filter historical multi-source data association patterns whose feature similarity exceeds the preset similarity threshold, and extract the pollution source type and diffusion path information of the historical pollution events corresponding to the historical multi-source data association patterns whose feature similarity exceeds the preset similarity threshold, as reference source tracing information.

[0085] A preset similarity threshold is set, determined based on the accuracy and recall requirements for identifying historical pollution events. From a feature similarity lookup table, historical multi-source data association patterns whose comprehensive feature similarity to the current anomaly association pattern exceeds the preset similarity threshold are selected. These historical patterns are highly similar to the current anomaly pattern, indicating that the current anomaly may be caused by reasons similar to historical events. The pollution source type and diffusion path information of the historical pollution events corresponding to these historical multi-source data association patterns are extracted and used as reference source tracing information.

[0086] Step S144: Analyze the spatiotemporal characteristics of the current port multi-source monitoring data set corresponding to the abnormal correlation pattern. The spatiotemporal characteristics include the distribution of monitoring points where the abnormal correlation pattern occurs and the time series characteristics of the abnormal correlation pattern.

[0087] Analyze the spatial and temporal characteristics of the current anomalous correlation patterns. Spatially, examine the distribution of monitoring points indicating abnormal gas concentrations within the port area. Are they concentrated in a specific area, such as the pier front, a storage yard, or a transport corridor, or scattered across multiple areas? What is the relative position of these monitoring points to port equipment, storage facilities, etc.? Temporally, analyze the onset time, peak time, and current trends of the anomalous correlation patterns, and whether they coincide with peak port operations, ship arrival and departure times, and the occurrence of specific meteorological conditions. Through spatiotemporal analysis, the scope of potential pollution sources can be further narrowed down, and key time points of anomaly occurrence can be identified.

[0088] Step S145: The reference source information is correlated and fused with the spatiotemporal features to determine the location range of the potential pollution source corresponding to the current abnormal correlation pattern and the possible pollution emission mode of the potential pollution source.

[0089] The system integrates the spatiotemporal characteristics of historical pollution source types and diffusion paths from the source tracing information with those of current anomaly correlation patterns. For example, if the source tracing information shows that a historical pollution event was caused by a certain type of loading and unloading equipment, and its diffusion path was mainly influenced by the wind direction at the time, and the current anomaly correlation pattern shows a monitoring point distribution similar to that historical diffusion path, and the anomaly occurred at the same time as the operation of that type of loading and unloading equipment, then it can be preliminarily determined that the current potential pollution source is likely to be that type of loading and unloading equipment, and its location range may be in an area similar to that of the historical pollution source. Simultaneously, based on the operating parameters of the current operating equipment and historical emission methods, the possible pollution emission methods of the potential pollution source can be inferred, such as direct emissions of exhaust gas and material dust emissions.

[0090] Step S146: Based on the location range of potential pollution sources and possible pollution emission methods, analyze the key factors affecting pollution generation. The key factors include the operating parameters of the equipment related to potential pollution sources and the meteorological state parameters related to pollution diffusion.

[0091] After determining the location and possible emission methods of potential pollution sources, further analysis is needed to identify the key factors influencing pollution generation. Operating parameters of equipment related to potential pollution sources, such as operating power, workload, fuel consumption rate, and maintenance status, directly affect the amount and concentration of pollutants generated. Meteorological parameters related to pollution diffusion, such as wind direction, wind speed, atmospheric stability, and temperature stratification, determine the diffusion speed, direction, and range of pollutants in the atmosphere. Analyzing these key factors allows for a deeper understanding of the mechanisms of pollution generation and diffusion.

[0092] Step S147: Integrate the location range of potential pollution sources, possible pollution emission methods and key factors to generate port air quality pollution source tracing clues that include pollution source location clues, pollution emission method clues and key influencing factors.

[0093] The analysis will integrate and summarize the location range of potential pollution sources, such as specific berths, storage yard numbers, and equipment numbers; possible pollution emission methods, such as combustion exhaust emissions, dust from material handling, and volatilization from leaks; and key factors affecting pollution generation and diffusion, such as specific operating parameters of equipment and meteorological parameters. This information will be presented in a clear and unambiguous manner to form a source tracing framework for port air quality pollution.

[0094] Step S150: Based on the source tracing clues of port air quality pollution, deduce the pollution diffusion trend description, integrate the pollution diffusion trend description with the current port air quality monitoring status, generate a port air quality monitoring response command, and send the port air quality monitoring response command to the port monitoring and control terminal.

[0095] Step S151: Extract the potential pollution source location range and key factors affecting pollution generation from the port air quality pollution source tracing clues, and determine the initial condition parameters required to deduce the pollution diffusion trend.

[0096] From the source tracing clues of air quality pollution in ports, the location range information of potential pollution sources is extracted and converted into specific geographic coordinates or relative location descriptions as the starting point of pollution diffusion. Simultaneously, key factors influencing pollution generation are extracted, such as the type of pollution source, the possible emission intensity (estimated based on operating parameters of equipment), the type of pollutant, and the initial concentration. This information collectively constitutes the initial condition parameters required to deduce the pollution diffusion trend. The accuracy of the initial condition parameters directly affects the reliability of the pollution diffusion trend deduction; therefore, they need to be as detailed and precise as possible.

[0097] Step S152: Obtain wind direction, wind speed, and atmospheric stability information from the current port meteorological monitoring data, and use the wind direction, wind speed, and atmospheric stability information as environmental impact parameters affecting pollution diffusion.

[0098] Current meteorological conditions are the main environmental factors influencing pollutant dispersion. Wind direction determines the direction of pollutant dispersion, wind speed affects the dispersion speed and transport distance, and atmospheric stability reflects the vertical mixing capacity of the atmosphere. When atmospheric stability is low (unstable), pollutants tend to disperse upwards and dilute quickly; when atmospheric stability is high (stable), pollutants are less likely to disperse and tend to accumulate near the ground. Obtaining these three key pieces of information from current port meteorological monitoring data will serve as environmental impact parameters affecting pollution dispersion, used for subsequent pollution dispersion trend derivation.

[0099] Step S153: Based on the initial condition parameters and the environmental impact parameters, build a pollution diffusion trend derivation logic to simulate the port area range that pollution may cover in different time periods.

[0100] Step S1531: Convert the potential pollution source location range in the initial condition parameters into a spatial coordinate representation form to determine the set of starting coordinate points for pollution emissions.

[0101] The geographical coordinates or relative location descriptions of potential pollution sources are converted into spatial coordinate representations based on port digital maps, such as coordinate points in a Cartesian coordinate system. If the pollution source is a region, such as a storage yard, the boundary coordinate points of that region are determined; if the pollution source is a specific piece of equipment, the center point of the equipment or the location of the emission outlet is used as the coordinate point. These coordinate points are combined to form the set of starting coordinate points for pollution emissions.

[0102] Step S1532: Based on the key factors in the initial condition parameters, derive the initial intensity characterization information of pollution emissions, which is derived based on the operating parameters of the equipment corresponding to the potential pollution source and the possible emission duration.

[0103] Based on key factors in the initial condition parameters, such as the operating power, workload, and fuel consumption rate of the equipment, and combined with the emission factor (pollutant emission per unit time or unit workload) of this type of equipment, the pollutant emission per unit time, i.e., the emission rate, is estimated. Simultaneously, based on the work plan or equipment operating status, the possible duration of pollution emissions is determined. By integrating information such as emission rate and duration, initial intensity characterization information of pollution emissions is derived. This initial intensity characterization information reflects the strength and persistence characteristics of pollutants emitted by the pollution source.

[0104] Step S1533: Convert the wind direction and wind speed information in the environmental impact parameters into vector representations, determine the wind field vector data for different time periods, and dynamically adjust the wind field vector data according to the changes in time periods.

[0105] Wind direction information indicates the direction from which the wind is coming, which is converted into a vector azimuth angle; wind speed information indicates the wind magnitude, which is converted into a vector magnitude. Combining wind direction and wind speed forms wind field vector data, which represents the direction and distance of air movement per unit time. Since meteorological conditions change over time, it is necessary to determine wind field vector data for different time periods (such as the next 1 hour, 2 hours, 4 hours, 8 hours, etc.) based on meteorological forecast data or short-term weather trend predictions, and to dynamically adjust the wind field vector data over time to reflect the impact of wind changes on pollution diffusion.

[0106] Step S1534: Convert the atmospheric stability information in the environmental impact parameters into atmospheric diffusion capacity characterization parameters. Different atmospheric stability corresponds to different atmospheric diffusion capacity characterization parameters, which are used to reflect the atmosphere's ability to diffuse pollutants.

[0107] Atmospheric stability is typically categorized into extremely unstable, unstable, weakly unstable, neutral, relatively stable, and extremely stable. Different atmospheric stability categories are converted into corresponding atmospheric diffusion capacity parameters, such as horizontal and vertical diffusion parameters. Generally, the lower the atmospheric stability, the stronger the atmospheric diffusion capacity and the larger the diffusion parameter value; conversely, the higher the atmospheric stability, the weaker the atmospheric diffusion capacity and the smaller the diffusion parameter value. These atmospheric diffusion capacity parameters are used to quantify the atmosphere's ability to diffuse and dilute pollutants in the horizontal and vertical directions.

[0108] Step S1535: Based on the set of starting coordinate points, initial intensity characterization information, wind field vector data and atmospheric diffusion capacity characterization parameters, a core framework for deducing the pollution diffusion trend is established. This core framework is designed based on the transport and diffusion laws of pollutants in the atmosphere.

[0109] By integrating the initial coordinate point set, initial intensity characterization information, wind field vector data, and atmospheric diffusion capacity characterization parameters, a core framework for deducing pollution diffusion trends is constructed. This core framework is based on the transport and diffusion laws of pollutants in the atmosphere, such as the basic principles of turbulent diffusion theory and Gaussian diffusion models. The framework considers the advection transport (moving with airflow) of pollutants after emission from the initial coordinate point under the influence of the wind field, the diffusion and dilution (horizontal and vertical diffusion) under the influence of atmospheric diffusion capacity, and possible processes such as deposition and chemical reactions (selectively considered according to the type of pollutant). Through simulation of these processes, the pollutant concentration distribution at different times and locations is calculated.

[0110] Step S1536: Set multiple consecutive time intervals, and simulate the range of port areas that may be covered by pollution at the end of each time interval based on the core framework.

[0111] Multiple consecutive time intervals, such as one hour, are set to simulate pollution diffusion over the next six hours. For each time interval, based on the core framework and incorporating wind field vector data and atmospheric diffusion capacity characterization parameters, the transport and diffusion process of pollutants is simulated starting from the initial coordinate point. At the end of each time interval, the area where pollutant concentration exceeds a certain threshold (such as the hourly average concentration limit in national air quality standards or the port's internal warning threshold) is calculated and recorded. This area represents the port area that pollution may cover at the end of that time interval. Through simulations of multiple consecutive time intervals, the dynamic process of pollutant diffusion over time can be obtained.

[0112] Step S154: Based on the pollution coverage area obtained from the simulation within different time periods, extract the core features of pollution diffusion. The core features include the main direction of pollution diffusion, the expansion of the pollution coverage area, and the changing trend of pollution concentration. Generate a pollution diffusion trend description based on the core features.

[0113] The simulated pollution coverage areas over different time periods were analyzed to extract core characteristics of pollution diffusion. Regarding the main direction, the main direction of pollution diffusion was determined based on the movement trajectory of the center of gravity of the pollution coverage area over different time periods, such as diffusion towards the northeast as a whole, or diffusion first towards the east and then towards the southeast. Regarding the expansion of the coverage area, the changes in the area of ​​the pollution coverage area over time were analyzed: whether it expanded rapidly, slowly, or remained relatively stable; and how the shape of the coverage area changed, whether it exhibited a fan shape, a band shape, or other shapes. Regarding the concentration trend, the changes in pollutant concentrations at various monitoring points or representative locations within the pollution coverage area over time were analyzed: whether it continuously increased, gradually decreased, or increased first and then decreased. Based on these core characteristics, a description of the pollution diffusion trend was generated in textual form, clearly showing the diffusion pattern of pollutants over a future period.

[0114] Step S155: Obtain the current port air quality monitoring status, which includes real-time gas component concentration information at each monitoring point and air quality identification information for each port area.

[0115] The current port air quality monitoring status is obtained through a real-time monitoring system, including real-time gas component concentration information from gas monitoring points distributed throughout the port, such as the current concentration values ​​of sulfur dioxide, nitrogen oxides, and particulate matter; and air quality labeling information for each port area, determined according to certain evaluation standards (such as the national air quality index (AQI) evaluation method or the port's internal evaluation standards), such as labels for excellent, lightly polluted, moderately polluted, and heavily polluted, as well as corresponding health recommendations or control requirements.

[0116] Step S156: Integrate the pollution diffusion trend description with the current port air quality monitoring status to identify the port monitoring areas that need to be focused on and the monitoring frequency that needs to be adjusted.

[0117] This study integrates and analyzes the predicted pollution coverage and concentration trends for different future time periods as described in the pollution diffusion trend analysis, along with real-time concentration information from current monitoring stations and air quality labeling information for each region. Areas currently polluted or predicted to be affected by pollution in the near future, especially those potentially reaching or exceeding severe pollution levels, are identified as key port monitoring areas requiring immediate attention. Simultaneously, the monitoring frequency for these key areas is adjusted based on the speed of pollution diffusion and the severity of concentration changes; for example, the monitoring frequency may be increased from once per hour to once every half hour or more frequently to more promptly grasp pollution changes. For non-key areas, the regular monitoring frequency can be maintained or appropriately reduced to optimize monitoring resource allocation.

[0118] Step S157: Based on the key port monitoring areas and the adjusted monitoring frequency, generate a monitoring equipment allocation plan, integrate the key area information, the adjusted monitoring frequency and the monitoring equipment allocation plan, and generate a port air quality monitoring response command.

[0119] Based on the number, location, and area of ​​the identified key port monitoring areas, and the adjusted monitoring frequency requirements, assess whether the coverage and frequency of existing monitoring equipment meet the needs. If the number of existing monitoring points in some key areas is insufficient or the monitoring equipment is malfunctioning, a monitoring equipment deployment plan needs to be developed, including temporarily adding portable monitoring equipment, adjusting the patrol routes and docking locations of mobile monitoring vehicles, or dispatching backup monitoring equipment from other areas to key areas. Integrate the information on key areas (name, location range, current air quality status), the adjusted monitoring frequency (specific monitoring time intervals for each area), and the monitoring equipment deployment plan (equipment type, quantity, deployment route, installation location) to form a complete and clearly defined port air quality monitoring response instruction.

[0120] Step S158: Send the port air quality monitoring response command to the port monitoring and control terminal.

[0121] The generated port air quality monitoring response commands are sent to the port monitoring and control terminal via the port's internal data communication network. The port monitoring and control terminal can be a central monitoring platform located in the port control room, or a sub-terminal responsible for managing monitoring equipment in various areas. Upon receiving the command, the monitoring and control terminal will, according to the command content, execute operations such as allocating monitoring equipment and adjusting monitoring frequencies to strengthen monitoring of key areas and obtain timely information on changes in air quality.

[0122] Figure 2 The illustration shows exemplary hardware and software components of a port air quality monitoring system 100 based on multi-source data analysis, which can implement the ideas of this application, according to some embodiments of this application. For example, a processor 120 can be used in the port air quality monitoring system 100 based on multi-source data analysis and to perform the functions in this application.

[0123] The port air quality monitoring system 100 based on multi-source data analysis can be a general-purpose server or a special-purpose server; both can be used to implement the port air quality monitoring method based on multi-source data analysis of this application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0124] For example, a port air quality monitoring system 100 based on multi-source data analysis may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the port air quality monitoring system 100 based on multi-source data analysis may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The port air quality monitoring system 100 based on multi-source data analysis also includes an I / O interface 150 between the computer and other input / output devices.

[0125] For ease of explanation, only one processor is described in the port air quality monitoring system 100 based on multi-source data analysis. However, it should be noted that the port air quality monitoring system 100 based on multi-source data analysis in this application may also include multiple processors. Therefore, the steps performed by one processor as described in this application may also be performed jointly or individually by multiple processors. For example, if the processor of the port air quality monitoring system 100 based on multi-source data analysis performs steps A and B, it should be understood that steps A and B may also be performed jointly by two different processors or individually by one processor. For example, the first processor performs step A, the second processor performs step B, or the first processor and the second processor jointly perform steps A and B.

[0126] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the port air quality monitoring method based on multi-source data analysis described above is implemented.

[0127] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A port air quality monitoring method based on multi-source data analysis, characterized in that, The method includes: A multi-source monitoring data set for the port is obtained, which includes port gas monitoring data, port meteorological monitoring data, and port operation activity monitoring data. The port gas monitoring data records gas composition monitoring information at different monitoring points in the port, the port meteorological monitoring data records meteorological status monitoring information for the port area, and the port operation activity monitoring data records operational status monitoring information for port operation equipment. Time synchronization and coordination are performed on the port multi-source monitoring data set to determine the core correlation items for cross-type data association, a correlation representation method is set for each core correlation item, and numerical mapping is used to represent the degree of correlation between different types of monitoring data. Based on the defined association representation method, the association degree value of each core association item is calculated respectively; Determine the row and column identifiers of the multi-source data association mapping table. The row identifiers correspond to the type classification of port multi-source monitoring data, and the column identifiers correspond to the classification of core association items. The calculated correlation values ​​of each core correlation item are filled into the corresponding positions of the multi-source data correlation mapping table according to the correspondence between row identifiers and column identifiers to form an initial multi-source data correlation mapping table. The indirect correlation relationships between each core correlation item are supplemented, the correlation transmission coefficient between different core correlation items is calculated, and the correlation transmission coefficient is integrated into the initial multi-source data correlation mapping table to form a complete multi-source data correlation mapping table. Based on the multi-source data association mapping table, abnormal association patterns of port air quality are mined, and association combinations that exceed the preset association threshold range are identified to obtain a set of abnormal association patterns. Calculate the feature similarity between each abnormal association pattern in the abnormal association pattern set and the historical multi-source data association pattern in the port historical pollution association record set, and extract the pollution source type and diffusion path information of the historical pollution events corresponding to the historical multi-source data association patterns that exceed the preset similarity threshold, as reference source tracing information; The spatiotemporal characteristics of the current port multi-source monitoring data set corresponding to the abnormal correlation pattern are analyzed. The reference source tracing information is correlated and fused with the spatiotemporal characteristics to determine the location range of the potential pollution source corresponding to the current abnormal correlation pattern and the possible pollution emission mode of the potential pollution source in order to analyze the key factors affecting pollution generation. The location range of the potential pollution source, the possible pollution emission mode and the key factors are integrated to generate port air quality pollution source tracing clues containing pollution source location clues, pollution emission mode clues and key influencing factors. The key factors include the operating parameters of the operating equipment related to the potential pollution source and the meteorological state parameters related to pollution diffusion. Based on the source tracing clues of port air quality pollution, a description of pollution diffusion trend is derived. The description of pollution diffusion trend is integrated with the current port air quality monitoring status to generate a port air quality monitoring response command, which is then sent to the port monitoring and control terminal.

2. The port air quality monitoring method based on multi-source data analysis according to claim 1, characterized in that, The method involves mining abnormal correlation patterns in port air quality based on the multi-source data association mapping table, identifying correlation combinations that exceed a preset correlation threshold range, and obtaining a set of abnormal correlation patterns, including: Determine the association dimensions in the multi-source data association mapping table. The association dimensions include the association dimension between gas composition in port gas monitoring data and meteorological status in port meteorological monitoring data, the association dimension between gas composition in port gas monitoring data and operating status of operating equipment in port operation activity monitoring data, and the association dimension between meteorological status in port meteorological monitoring data and operating status of operating equipment in port operation activity monitoring data. The correlation strength of each correlation dimension is assessed, and the assessment priority of each correlation dimension is adjusted according to the port air quality monitoring needs to generate the correlation dimension priority allocation result. The multi-source data association mapping table is matched with a preset port normal air quality association pattern library. The matching degree between each association combination in the multi-source data association mapping table and the corresponding association pattern in the port normal air quality association pattern library is calculated to obtain the association pattern matching degree sequence. Filter the association combinations in the association pattern matching degree sequence whose matching degree is lower than a preset matching degree threshold, and mark the association combinations whose matching degree is lower than the preset matching degree threshold as candidate abnormal association combinations; Analyze the frequency and duration of the candidate anomaly association combinations during the continuous monitoring period, and retain candidate anomaly association combinations whose frequency of occurrence exceeds a preset frequency threshold and whose duration of occurrence exceeds a preset duration threshold; The retained candidate anomaly association combinations are supplemented with complete association dimension information to supplement the missing association dimension information in the retained candidate anomaly association combinations and generate complete anomaly association patterns. All complete anomaly association patterns are categorized and integrated according to association dimensions to obtain an anomaly association pattern set.

3. The port air quality monitoring method based on multi-source data analysis according to claim 1, characterized in that, The calculation of the feature similarity between each abnormal association pattern in the set of abnormal association patterns and the historical multi-source data association patterns in the set of port historical pollution association records includes: Extract the feature dimensions of each abnormal correlation pattern. The feature dimensions include the gas composition type involved in the abnormal correlation pattern, the meteorological state parameters involved in the abnormal correlation pattern, the operating parameters of the operating equipment involved in the abnormal correlation pattern, and the correlation strength characteristics between each parameter. Extract the corresponding feature dimension of each historical multi-source data association pattern in the port historical pollution association record set, characterize the parameters under each feature dimension, convert different types of parameters into a unified representation form, calculate the degree of parameter difference between the abnormal association pattern and the historical multi-source data association pattern under each feature dimension, and assign corresponding dimension weights to each feature dimension according to the degree of influence of each feature dimension on the port air quality pollution source tracing. Based on the degree of parameter difference and corresponding dimension weight of each feature dimension, the comprehensive feature similarity between the abnormal association pattern and the historical multi-source data association pattern is calculated. The comprehensive feature similarity is used to reflect the overall matching situation between the abnormal association pattern and the historical multi-source data association pattern. Record the comprehensive feature similarity between each abnormal association pattern and the association patterns of each historical multi-source data, and form a feature similarity comparison table.

4. The port air quality monitoring method based on multi-source data analysis according to claim 1, characterized in that, The description of pollution diffusion trend derived from the source tracing clues of port air quality pollution, combined with the current port air quality monitoring status, generates a port air quality monitoring response command, including: Extract the potential pollution source location range and key factors affecting pollution generation from the port air quality pollution source tracing clues, and determine the initial condition parameters required to deduce the pollution diffusion trend; Obtain wind direction, wind speed, and atmospheric stability information from current port meteorological monitoring data, and use these information as environmental impact parameters affecting pollution diffusion. Based on the initial condition parameters and the environmental impact parameters, a logic for deducing the pollution diffusion trend is established to simulate the range of port areas that may be covered by pollution in different time periods. Based on the pollution coverage area obtained from the simulation within different time periods, the core features of pollution diffusion are extracted. These core features include the main direction of pollution diffusion, the expansion of the pollution coverage area, and the changing trend of pollution concentration. A description of pollution diffusion trend is generated based on these core features. Obtain the current port air quality monitoring status, which includes real-time gas component concentration information at each monitoring point and air quality identification information for each port area; By integrating the pollution spread trend description with the current port air quality monitoring status, the port monitoring areas that need to be focused on and the monitoring frequencies that need to be adjusted can be identified. Based on the key port monitoring areas and the adjusted monitoring frequencies, a monitoring equipment allocation plan is generated. By integrating the information on key monitoring areas, the adjusted monitoring frequencies, and the monitoring equipment allocation plan, a port air quality monitoring response command is generated.

5. The port air quality monitoring method based on multi-source data analysis according to claim 4, characterized in that, Based on the initial condition parameters and the environmental impact parameters, a logic for deducing the pollution diffusion trend is established to simulate the port area that may be covered by pollution in different time periods, including: The potential pollution source location range in the initial condition parameters is converted into a spatial coordinate representation to determine the set of starting coordinate points for pollution emissions; Based on the key factors in the initial condition parameters, the initial intensity characterization information of pollution emissions is derived. The initial intensity characterization information is derived based on the operating parameters of the equipment corresponding to the potential pollution source and the possible emission persistence. The wind direction and wind speed information in the environmental impact parameters are converted into vector representations to determine the wind field vector data for different time periods. The wind field vector data will be dynamically adjusted according to the changes in time periods. The atmospheric stability information in the environmental impact parameters is converted into atmospheric diffusion capacity characterization parameters. Different atmospheric stability corresponds to different atmospheric diffusion capacity characterization parameters, which are used to reflect the atmosphere's ability to diffuse pollutants. Based on the set of starting coordinate points, initial intensity characterization information, wind field vector data and atmospheric diffusion capacity characterization parameters, a core framework for deducing the pollution diffusion trend is constructed. This core framework is designed based on the transport and diffusion laws of pollutants in the atmosphere. Multiple consecutive time intervals are set, and the spatial distribution of pollution in the port area at the end of each time interval is simulated based on the core framework. The simulation results of the spatial distribution of pollution for all time intervals were collected, and the range of port areas that may be covered by pollution in different time periods was obtained.

6. The port air quality monitoring method based on multi-source data analysis according to claim 1, characterized in that, Based on the defined association representation method, the association degree value of each core association item is calculated and recorded, including: For each core related item, extract the numerical sequence of the two types of monitoring data corresponding to the same time node, wherein the numerical sequence contains monitoring data values ​​of multiple consecutive time nodes; Analyze the changing trends of the two types of monitoring data numerical sequences, determine whether the directions of increase and decrease of the two types of numerical sequences are consistent, calculate the matching of the increase and decrease magnitudes, and derive the consistency coefficient of the changing trends based on the consistency of the increase and decrease directions and the matching of the increase and decrease magnitudes. The consistency coefficient of the trend of change is converted into a correlation degree value. The representation range of the correlation degree value corresponds to the derivation range of the consistency coefficient of the trend of change. The correlation degree value corresponding to each core correlation item is recorded. At the same time, the time node range and the length of the numerical sequence on which the correlation degree value is calculated are also recorded. Compile the effective correlation values ​​of all core related items to form a correlation comparison table of core related items.

7. The port air quality monitoring method based on multi-source data analysis according to claim 2, characterized in that, The analysis examines the frequency and duration of the candidate anomaly association combinations within a continuous monitoring period, retaining candidate anomaly association combinations whose frequency exceeds a preset frequency threshold and whose duration exceeds a preset duration threshold, including: The time range for continuous monitoring is set, which is determined based on the regular monitoring cycle of port air quality and the possible duration of pollution events. Within a set time frame, the frequency of occurrence of each candidate anomaly association combination at different time points is counted to obtain the occurrence frequency of each candidate anomaly association combination. For each candidate anomaly association combination, identify the time interval in which it occurs consecutively within a set time range, and each consecutive time interval contains multiple consecutive time nodes; Calculate the time difference between the start and end time points of each consecutive time interval; this time difference is the duration of the corresponding consecutive time interval. For each candidate anomaly association combination, compare the duration of all continuous time intervals, select the continuous time interval with the longest duration, and determine the duration of the candidate anomaly association combination as the duration of the longest duration. The occurrence frequency of each candidate anomaly association combination is compared with a preset frequency threshold, and its duration is compared with a preset duration threshold. Candidate abnormal association combinations that appear more frequently than a preset frequency threshold and last for more than a preset duration threshold are selected and used as abnormal association combinations to be supplemented in the association dimension. Record the occurrence time of abnormal association combinations of the dimensions to be supplemented, as well as the start and end time of the longest continuous time interval.

8. A port air quality monitoring system based on multi-source data analysis, characterized in that, The port air quality monitoring system based on multi-source data analysis includes a processor and a memory, the memory and the processor being connected. The memory is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the memory to implement the port air quality monitoring method based on multi-source data analysis as described in any one of claims 1-7.

Images