Intelligent Control System for Multi-Source Data Quality of Marine Observation Station Network

Abstract

This invention discloses an intelligent control system for the quality of multi-source data from marine observation network, relating to the field of intelligent control technology. This invention improves the accuracy of data-sea state matching and quality assessment in the data quality control process of marine observation network. It can adapt to the distribution characteristics of different special marine events and the needs for identifying abnormal elements and predicting quality degradation under equipment contamination conditions. It achieves efficient and accurate solutions for abnormal element screening and dynamic adjustment of personalized maintenance parameters, significantly improving the accuracy of data quality assessment and the rationality of maintenance resource allocation in the marine observation network. It effectively prevents the insufficient response of fixed quality thresholds to differences in actual sea state characteristics and equipment status evolution, avoids quality assessment errors and waste of maintenance resources caused by data misjudgment, and ensures a high degree of alignment between quality control measures and actual marine observation needs.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to an intelligent control system for the quality of multi-source data from marine observation station networks. Background Technology

[0002] Intelligent control of multi-source data quality in marine observation network is a core technology for quality assurance in marine environmental monitoring system. Under long-term continuous observation conditions, it is necessary to achieve accurate matching between sensor anomaly characteristics and marine special event states, and dynamic coupling evaluation of data quality and real-time deviation. Therefore, it is essential to construct an intelligent data quality control system based on bidirectional matching analysis of multi-source data and sea state environment and feedback correction of deviation field between stations.

[0003] In existing technologies, the assessment of data quality in marine observation network relies on human experience and fixed threshold judgment rules, which cannot accurately distinguish between data anomalies caused by sensor malfunctions and normal fluctuations caused by special marine events. Although existing technologies have introduced basic data range checking devices and outlier detection mechanisms, they do not consider the physical coupling relationship between observation elements and the spatial correlation of marine environmental dynamics, and lack coupled modeling of sensor fouling accumulation and data quality degradation trends. Therefore, it is difficult to achieve accurate assessment of the true state of marine observation network data quality, the trajectory of personalized maintenance parameter adjustments, and future quality risks. When faced with complex sea state distribution or the multi-factor coupling effect caused by uneven equipment fouling, problems such as data misjudgment accidents, quality assessment defects, and missed detection of key equipment hazards are likely to occur.

[0004] Furthermore, existing technologies lack a similarity matching mechanism for the abnormal distribution characteristics and equipment fouling status differences of different marine observation stations during the quality assessment process. This makes it impossible to effectively achieve predictive adjustment of maintenance parameters for the next station and the formulation of differentiated quality control schemes based on the deviation field data of the current station. In addition, there are defects such as weak sea state impact identification capability, low equipment fouling compensation accuracy, lack of abnormal element correlation analysis, insufficient quality degradation prediction strategy, and unreasonable maintenance scheduling priority determination. As a result, the timeliness and accuracy of data quality early warning of marine observation station network cannot be effectively guaranteed. At the same time, there is a lack of adaptive control strategies for sequentially maintained stations within the same observation station network, and it is impossible to adaptively generate differentiated maintenance compensation schemes for the next station based on dynamic changes such as the current data deviation field and equipment fouling evolution. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent control system for the quality of multi-source data from a marine observation network, thereby solving the problems existing in the background technology.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a multi-source data quality intelligent control system for a marine observation station network, including: a marine environmental feature extraction module, used to acquire multi-source observation data of each marine observation station at each monitoring time point during the monitoring period, thereby analyzing the marine environmental dynamic characteristic parameters of each marine observation station at each monitoring time point.

[0007] The multi-element coupling verification module is used to construct the marine element coupling correlation matrix of each marine observation station within the monitoring period based on the multi-source observation data of each marine observation station at each monitoring time point, thereby identifying each coupling anomalous element of each marine observation station within the monitoring period.

[0008] The spatiotemporal gradient field quality control module is used to detect gradient field anomalies based on the coupled anomalies of each marine observation station within the monitoring period, thereby screening the target anomalies of each marine observation station within the monitoring period.

[0009] The sea state impact assessment module is used to identify the types and occurrence times of various marine special events at each marine observation station during the monitoring period based on the marine environmental dynamic characteristic parameters at each monitoring time point. It also performs sea state correlation analysis by combining the various target anomaly elements at each marine observation station during the monitoring period, thereby generating the quality degradation characteristic curves of each marine observation station.

[0010] The intelligent quality control and scheduling module is used to predict equipment fouling based on the quality degradation characteristic curves of each marine observation station, deduce the graded quality control and maintenance plan for the marine observation station, and send it to the observation station network management center.

[0011] The beneficial effects of this invention are as follows: This invention improves the accuracy of data-sea state matching and quality assessment in the data quality control process of marine observation station networks. It can adapt to the distribution characteristics of different marine special events and the needs for identifying abnormal elements and predicting quality degradation under equipment fouling conditions. Through bidirectional matching analysis and coupled computation, it accurately distinguishes between sensor malfunctions and influencing factors of marine special events, thereby achieving efficient and accurate solutions for abnormal element screening and dynamic adjustment of personalized maintenance parameters. This significantly improves the accuracy and rationality of maintenance resource allocation in the data quality assessment process of marine observation station networks. It can also be applied based on the temporal clustering degree of coupled abnormal elements and the overall fouling degree of equipment at each marine observation station. The system dynamically determines the data quality risk coefficient and generates maintenance level adjustment plans and quality control compensation distribution plans for the next site through real-time feedback from the deviation field. This effectively prevents the fixed quality threshold from being insufficient to respond to differences in actual sea conditions and equipment status evolution, avoids quality assessment errors and waste of maintenance resources caused by data misjudgment, and enables accurate calculation of maintenance time compensation values ​​based on quality degradation characteristic curves and the formulation of differentiated site maintenance priority plans. This effectively improves the pertinence and implementation effect of the final quality control scheduling plan, accurately identifies high-risk sites that threaten the quality of marine observation data and their maintenance adjustment windows, and ensures that quality control measures are highly consistent with actual marine observation needs. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Reference Figure 1 As shown, the present invention provides an intelligent control system for the quality of multi-source data from a marine observation network, comprising: a marine environmental feature extraction module, a multi-element coupling verification module, a spatiotemporal gradient field quality control module, a sea state impact assessment module, and an intelligent quality control scheduling module.

[0016] It should be noted that the marine environmental feature extraction module is connected to the multi-element coupling verification module, the multi-element coupling verification module is connected to the spatiotemporal gradient field quality control module, the spatiotemporal gradient field quality control module is connected to the sea state impact assessment module, and the sea state impact assessment module is connected to the intelligent quality control scheduling module.

[0017] The marine environmental feature extraction module is used to acquire multi-source observation data from each marine observation station at each monitoring time point during the monitoring period, thereby analyzing the marine environmental dynamic characteristic parameters of each marine observation station at each monitoring time point.

[0018] In one specific embodiment, the method for acquiring multi-source observation data of each marine observation station at each monitoring time point during the monitoring period is as follows: multi-source observation data are collected through sensor equipment deployed at each marine observation station, including water temperature sensors, salinity sensors, wind speed and direction sensors, current velocity and direction sensors, etc. These sensors automatically collect data at each monitoring time point according to a preset sampling frequency to form multi-source observation data of each marine observation station at each monitoring time point.

[0019] In a specific embodiment of the present invention, the marine environmental dynamic characteristic parameters of each marine observation station at each monitoring time point are analyzed. The specific method is as follows: the tidal harmonic constant of the sea area where each marine observation station is located is obtained from the local database, and the theoretical tide level value of each marine observation station at each monitoring time point is evaluated in combination with each monitoring time point.

[0020] It should be noted that the tidal harmonic constant is a set of parameters that describe the characteristics of tidal motion in a sea area. It can predict the theoretical tidal level at any time in that sea area and is an existing technology in the field of marine surveying.

[0021] In one specific embodiment, the theoretical tide level values ​​of each marine observation station at each monitoring time point are evaluated. The specific method is as follows: the theoretical tide level values ​​of each marine observation station at each monitoring time point can be obtained by using existing techniques for predicting theoretical tide level values ​​through tidal harmonic constants.

[0022] Based on multi-source observation data from each marine observation station at each monitoring time point, the seawater density of each marine observation station at each monitoring time point is calculated, and it is unified with the theoretical tide level value of each marine observation station at each monitoring time point as the marine environmental dynamic characteristic parameter of each marine observation station at each monitoring time point.

[0023] In one specific embodiment, the seawater density of each marine observation station at each monitoring time point is calculated by extracting the measured water temperature, measured salinity, and measuring point depth from the multi-source observation data of each marine observation station at each monitoring time point. These three parameters are then substituted into the general seawater state equation for calculation. The seawater state equation describes the functional relationship between seawater density and water temperature, salinity, and pressure. The pressure value is converted from the measuring point depth value. The theoretical value of seawater density of each marine observation station at each monitoring time point is calculated through the seawater state equation.

[0024] The multi-element coupling verification module is used to construct the marine element coupling correlation matrix of each marine observation station within the monitoring period based on the multi-source observation data of each marine observation station at each monitoring time point, thereby identifying each coupling anomalous element of each marine observation station within the monitoring period.

[0025] In a specific embodiment of the present invention, a coupled correlation matrix of marine elements for each marine observation station within the monitoring time period is constructed. The specific method is as follows: based on the multi-source observation data of each marine observation station at each monitoring time point, the measured water temperature, measured salinity, measured wind speed, measured wind direction, measured current velocity, and measured current direction of each marine observation station at each monitoring time point are extracted.

[0026] Historical observation data from various marine observation stations were obtained from local databases. The historical salinity distribution patterns of each marine observation station under the same water temperature range were statistically analyzed to construct a statistical model of temperature-salinity relationship for each marine observation station. The historical flow velocity and direction distribution patterns of each marine observation station under the same wind speed and direction conditions were also statistically analyzed to construct a statistical model of wind flow relationship for each marine observation station.

[0027] In one specific embodiment, the historical salinity distribution patterns of each marine observation station under the same water temperature range are statistically analyzed to construct a temperature-salinity statistical relationship model for each marine observation station. Similarly, the historical current velocity and direction distribution patterns of each marine observation station under the same wind speed and direction conditions are statistically analyzed to construct a wind-current statistical relationship model for each marine observation station. The specific method is as follows: historical observation data of each marine observation station is obtained from a local database; historical water temperature data is divided into multiple water temperature intervals at 0.5 degrees Celsius intervals; all historical salinity values ​​corresponding to each water temperature interval are statistically analyzed; the mean, median, and standard deviation of historical salinity values ​​within each water temperature interval are calculated; a correspondence table between water temperature intervals and salinity statistical characteristics is established; and a temperature-salinity statistical relationship model for each marine observation station is formed. Similarly, historical wind speed data is divided into multiple wind speed intervals at 1 meter per second intervals, and wind direction is divided into 16 directions at 22.5-degree intervals. All historical flow velocity and direction values ​​corresponding to each wind speed interval and wind direction combination are statistically analyzed. The mean and standard deviation of historical flow velocity and direction under each combination condition are calculated, and a correspondence table between wind speed and wind direction combinations and flow velocity and direction statistical characteristics is established to form a wind flow statistical relationship model for each marine observation station.

[0028] Based on the measured water temperature and salinity at each monitoring time point of each marine observation station, and combined with the temperature-salinity statistical relationship model of each marine observation station, the deviation was calculated to obtain the temperature-salinity coupling deviation coefficient of each marine observation station at each monitoring time point.

[0029] In one specific embodiment, deviation calculation is performed to obtain the temperature-salinity coupling deviation coefficient of each marine observation station at each monitoring time point. The specific method is as follows: based on the measured water temperature value of each marine observation station at each monitoring time point, the water temperature interval to which the measured water temperature value belongs is found in the temperature-salinity statistical relationship model. The historical salinity mean and standard deviation corresponding to the water temperature interval are extracted from the temperature-salinity statistical relationship model. The measured salinity value of each marine observation station at each monitoring time point is subtracted from the historical salinity mean corresponding to the water temperature interval to obtain the salinity deviation value. Then, the salinity deviation value is divided by the historical salinity standard deviation corresponding to the water temperature interval to obtain the standardized temperature-salinity coupling deviation coefficient. This coefficient reflects the degree to which the measured salinity value deviates from the historical statistical law under the current water temperature conditions.

[0030] Based on the measured wind speed and direction at each monitoring time point of each marine observation station, and combined with the wind flow statistical relationship model of each marine observation station, the deviation was calculated to obtain the wind flow coupling deviation coefficient of each marine observation station at each monitoring time point.

[0031] In one specific embodiment, deviation calculation is performed to obtain the wind-flow coupling deviation coefficient of each marine observation station at each monitoring time point. The specific method is as follows: based on the measured wind speed and measured wind direction values ​​of each marine observation station at each monitoring time point, the wind speed interval and wind direction combination to which the measured wind speed and wind direction belong are found in the wind-flow statistical relationship model. The historical mean and standard deviation of the current velocity corresponding to the combination conditions are extracted from the wind-flow statistical relationship model. The measured current velocity value of each marine observation station at each monitoring time point is subtracted from the historical mean current velocity corresponding to the combination conditions to obtain the current velocity deviation value. Then, the current velocity deviation value is divided by the historical standard deviation of the current velocity corresponding to the combination conditions to obtain the current velocity deviation target value. The angular deviation value between the measured current direction and the historical mean current direction is calculated. The dimensionless value of the current velocity deviation target value and the dimensionless value of the angular deviation value are added together to obtain the wind-flow coupling deviation coefficient.

[0032] The temperature-salinity coupling deviation coefficient and wind-current coupling deviation coefficient of each marine observation station at each monitoring time point are constructed into a matrix, thereby constructing the marine element coupling correlation matrix of each marine observation station within the monitoring time period.

[0033] In a specific embodiment of the present invention, the method for identifying each coupling anomaly element of each marine observation station within a monitoring period is as follows: obtaining the temperature-salinity coupling deviation threshold and the wind-current coupling deviation threshold from the local database; based on the temperature-salinity coupling deviation coefficient and the wind-current coupling deviation coefficient of each marine observation station at each monitoring time point; if the temperature-salinity coupling deviation coefficient of a certain marine observation station at a certain monitoring time point exceeds the temperature-salinity coupling deviation threshold, then the water temperature sensor and the salinity sensor are marked as coupling anomalies, and the monitoring time point is taken as the occurrence monitoring time point, thereby obtaining the occurrence monitoring time points of each water temperature coupling anomaly element and each occurrence monitoring time point of each salinity coupling anomaly element of each marine observation station.

[0034] If the wind-flow coupling deviation coefficient of a certain marine observation station exceeds the wind-flow coupling deviation threshold at a certain monitoring time point, then the wind speed and wind direction sensors and the current velocity and current direction sensors at that monitoring time point are marked as coupling anomalies, thereby obtaining the monitoring time points at which the wind speed and wind direction coupling anomalies and the current velocity and current direction coupling anomalies occur at each marine observation station.

[0035] The above data were compiled and used as the coupling anomaly elements of each marine observation station during the monitoring period.

[0036] The spatiotemporal gradient field quality control module is used to detect gradient field anomalies based on the coupled anomalies of each marine observation station within the monitoring period, thereby screening the target anomalies of each marine observation station within the monitoring period.

[0037] In a specific embodiment of the present invention, the method for screening the target anomaly elements of each marine observation station within the monitoring time period is as follows: based on the monitoring time points of each occurrence of each coupled anomaly element of each marine observation station within the monitoring time period, the frequency of anomalies of each coupled anomaly element of each marine observation station within the monitoring time period is counted, and the anomaly time interval sequence of each coupled anomaly element of each marine observation station within the monitoring time period is calculated.

[0038] In one specific embodiment, the frequency of occurrence of each coupled anomaly element at each marine observation station within the monitoring time period is counted, and the anomaly time interval sequence of each coupled anomaly element at each marine observation station within the monitoring time period is calculated. The specific method is as follows: Based on each coupled anomaly element at each marine observation station within the monitoring time period, the total number of monitoring time points for each occurrence of water temperature coupled anomaly element, salinity coupled anomaly element, wind speed and direction coupled anomaly element, and current speed and direction coupled anomaly element is counted. This number divided by the number of monitoring time points is the frequency of occurrence of each coupled anomaly element. For each coupled anomaly element, its occurrence monitoring time points are arranged in chronological order, and the time difference between two adjacent occurrence monitoring time points is calculated to form a time interval sequence. For example, if the water temperature coupled anomaly element occurs at the 2nd minute, 5th minute, and 9th minute, then its anomaly time interval sequence is 3 minutes and 4 minutes.

[0039] Based on the anomalous time interval sequence of each coupled anomalous element of each marine observation station within the monitoring period, the temporal gradient acceleration coefficient of each coupled anomalous element of each marine observation station within the monitoring period is evaluated. Combined with the frequency of anomalous occurrence of each coupled anomalous element of each marine observation station within the monitoring period, the comprehensive anomalous coefficient of each coupled anomalous element of each marine observation station within the monitoring period is calculated. Based on this, each target anomalous element of each marine observation station within the monitoring period is selected.

[0040] In one specific embodiment, the comprehensive anomaly coefficient of each coupled anomaly element of each marine observation station within the monitoring time period is calculated, and the target anomaly element of each marine observation station within the monitoring time period is screened accordingly. The specific method is as follows: based on the temporal gradient acceleration coefficient of each coupled anomaly element of each marine observation station within the monitoring time period. Where x represents the number of each ocean observation station, y is a positive integer greater than 2, and n represents the number of each coupling anomaly element. m is a positive integer greater than 2, and the frequency of occurrence of each coupled anomaly element at each marine observation station within the monitoring period is also considered. Calculate the comprehensive anomaly coefficient of each coupled anomaly element at each marine observation station within the monitoring period. Where e represents a natural constant, the comprehensive anomaly coefficient threshold is obtained from the local database, and the comprehensive anomaly coefficient of each coupled anomaly element is compared with the comprehensive anomaly coefficient threshold. If the comprehensive anomaly coefficient of a coupled anomaly element is greater than the comprehensive anomaly coefficient threshold, the coupled anomaly element is marked as a target anomaly element, thereby obtaining the target anomaly elements of each marine observation station within the monitoring period.

[0041] In a specific embodiment of the present invention, the temporal gradient acceleration coefficient of each coupled anomaly element of each marine observation station within the monitoring period is evaluated. The specific method is as follows: based on the anomaly time interval sequence of each coupled anomaly element of each marine observation station within the monitoring period, the time interval values ​​of each coupled anomaly element of each marine observation station within the monitoring period between two consecutive anomalies are extracted, and the time interval change rate sequence of each coupled anomaly element of each marine observation station within the monitoring period is summarized accordingly.

[0042] Based on the time interval change rate sequence of each coupled anomaly element of each marine observation station within the monitoring period, the anomalous acceleration trend fitting curve of each coupled anomaly element of each marine observation station within the monitoring period is fitted, and the slope values ​​of each coupled anomaly trend fitting curve of each coupled anomaly element of each marine observation station are extracted, thereby calculating the temporal gradient acceleration coefficient of each coupled anomaly element of each marine observation station within the monitoring period.

[0043] In one specific embodiment, fitting curves of the anomalous acceleration trends of each coupled anomaly element at each marine observation station within the monitoring period are fitted, and the slope values ​​of the fitting curves of the anomalous acceleration trends of each coupled anomaly element at each marine observation station are extracted. This allows for the calculation of the temporal gradient acceleration coefficient of each coupled anomaly element at each marine observation station within the monitoring period. Specifically, based on the time interval change rate sequence of each coupled anomaly element at each marine observation station within the monitoring period, the order of anomaly occurrence is used as the horizontal axis, and the time interval change rate is used as the vertical axis. A curve fitting method is used to fit the data points, obtaining the fitting curves of the anomalous acceleration trends of each coupled anomaly element at each marine observation station within the monitoring period. The derivative of these fitting curves is calculated at each point of anomaly occurrence to obtain the slope values ​​of the fitting curves of the anomalous acceleration trends of each coupled anomaly element at each marine observation station. If the slope value is negative, it indicates that the time interval is shortening, i.e., the anomaly is accelerating. The number and magnitude of all negative slope values ​​are counted, and the average absolute value of the negative slope values ​​is calculated as the temporal gradient acceleration coefficient of each coupled anomaly element at each marine observation station within the monitoring period.

[0044] The sea state impact assessment module is used to identify the types and occurrence times of various marine special events at each marine observation station during the monitoring period based on the marine environmental dynamic characteristic parameters at each monitoring time point. It also performs sea state correlation analysis by combining the various target anomaly elements at each marine observation station during the monitoring period, thereby generating the quality degradation characteristic curves of each marine observation station.

[0045] In a specific embodiment of the present invention, the method for identifying the type and occurrence time of each special marine event at each marine observation station within the monitoring period is as follows: historical tide level data and historical density data of each marine observation station are obtained from the local database, the historical tide level quantile distribution of each marine observation station in each month and time period is statistically analyzed to construct a tide level benchmark database for each marine observation station, and the historical density quantile distribution of each marine observation station in each month is statistically analyzed to construct a density benchmark database for each marine observation station.

[0046] In one specific embodiment, the historical tidal quantile distribution of each marine observation station in each month and time period is statistically analyzed to construct a tidal benchmark database for each marine observation station. Similarly, the historical density quantile distribution of each marine observation station in each month is statistically analyzed to construct a density benchmark database for each marine observation station. The specific method is as follows: historical tidal data is classified by month and time period. For each time period of each month, all historical tidal values ​​under that category are extracted. These historical tidal values ​​are statistically analyzed to calculate the representative value and the quantized dispersion value for that category. The representative value characterizes the normal tidal level for that month and time period, and the quantized dispersion value characterizes the tidal fluctuation range for that month and time period. The representative value and quantized dispersion value corresponding to each time period of each month are stored to form the tidal benchmark database for each marine observation station. Likewise, historical density data is classified by month and time period. For each time period of each month, all historical density values ​​are extracted, and the representative value and quantized dispersion value for each time period of each month are calculated. These are then summarized to form the density benchmark database for each marine observation station.

[0047] Based on the theoretical tide levels of each marine observation station at each monitoring time point, and in conjunction with the tide level reference database of each marine observation station, the tide level deviation index of each marine observation station at each monitoring time point is calculated.

[0048] In one specific embodiment, the tide level deviation index of each marine observation station at each monitoring time point is calculated. The specific method is as follows: based on the time information of each marine observation station at each monitoring time point, the month and time period to which the monitoring time point belongs are determined. The representative value of the tide level and the quantified value of the dispersion of the corresponding month and time period are queried from the tide level benchmark database. The theoretical tide level value of each marine observation station at each monitoring time point is subtracted from the representative value of the tide level to obtain the absolute value of the tide level deviation. Then, the absolute value of the tide level deviation is divided by the quantified value of the dispersion to obtain the tide level deviation index of each marine observation station at each monitoring time point.

[0049] Based on the seawater density at each monitoring time point of each marine observation station, and in conjunction with the density benchmark library of each marine observation station, the density deviation index of each marine observation station at each monitoring time point is calculated, and the seawater density change rate between each adjacent monitoring time point of each marine observation station is calculated.

[0050] In one specific embodiment, the density deviation index of each marine observation station at each monitoring time point is calculated, and the seawater density change rate between each marine observation station at each adjacent monitoring time point is calculated. The specific method is as follows: based on the time information of each marine observation station at each monitoring time point, the month and time period to which the monitoring time point belongs are determined. Following the method of calculating the tide deviation index of each marine observation station at each monitoring time point, the density deviation index of each marine observation station at each monitoring time point is calculated in the same way. The seawater density values ​​of two adjacent monitoring time points of each marine observation station are extracted. The seawater density value of the previous time point is subtracted from the seawater density value of the later time point to obtain the density change. The density change is then divided by the time interval between the two monitoring time points to obtain the seawater density change rate between each marine observation station at each adjacent monitoring time point.

[0051] The system retrieves a rule base for association of marine special events from a local database. Based on the tide level deviation index, density deviation index, and seawater density change rate between adjacent monitoring time points for each marine observation station, and cross-validates these rules with the rule base for association of marine special events, it identifies the types and time periods of marine special events for each marine observation station within the monitoring period.

[0052] It should be noted that the marine special event association rule base is a pre-established knowledge base containing identification rules for various marine special events. Marine special events include phenomena such as storm surges, ocean fronts, mesoscale eddies, internal waves, red tides, and sea ice. The rule base stores typical combinations of characteristic parameters for each marine special event, including the tide level deviation index threshold, density deviation index threshold, and density change rate threshold. For example, for certain regions, the identification rule for storm surge events is a tide level deviation index greater than 2 and a duration of more than 6 hours; the identification rule for ocean front events is an absolute value of density deviation index greater than 1.5 and an absolute value of density change rate greater than 0.5 kg / m³ / h; and the identification rule for mesoscale eddies is a density deviation index that is continuously positive or negative for more than 12 hours and a density change rate that shows periodic changes. The rule base contains combined logical judgment conditions for multiple characteristic parameters to distinguish different types of marine special events.

[0053] In one specific embodiment, the method for identifying the type and occurrence time of various marine special events at each marine observation station within a monitoring period is as follows: Identification rules for various types of marine special events are extracted from a marine special event association rule base. The tide level deviation index, density deviation index, and seawater density change rate at each monitoring time point of each marine observation station are compared with the threshold conditions in each identification rule. If the tide level deviation index, density deviation index, and density change rate at a certain monitoring time point simultaneously meet all the threshold conditions for a certain type of marine special event, it is preliminarily determined that such a marine special event may occur at that monitoring time point. Further analysis is conducted to determine whether multiple consecutive monitoring time points continuously meet the identification rules for the event. If the number of consecutively met monitoring time points exceeds the minimum duration required for the event type, the occurrence of the marine special event is confirmed. The first monitoring time point that meets the conditions is taken as the event start time, and the last monitoring time point that meets the conditions is taken as the event end time. These two constitute the occurrence time of the marine special event. The type and occurrence time of the marine special event are recorded, thereby obtaining the type and occurrence time of various marine special events at each marine observation station within a monitoring period.

[0054] In a specific embodiment of the present invention, the quality degradation characteristic curves of each marine observation station are generated by the following method: obtaining the element anomaly feature library corresponding to the type of each marine special event from the local database, and mapping the element anomaly feature library corresponding to each marine special event of each marine observation station according to the type of each marine special event within the monitoring period.

[0055] It should be noted that the element anomaly feature database is a pre-established database of sensor anomaly features for various marine special events. It records the normal and abnormal fluctuation characteristics that water temperature sensors, salinity sensors, wind speed and direction sensors, and current velocity and direction sensors may exhibit during various marine special events. For example, during storm surge events, current velocity and direction sensors usually exhibit abnormally violent fluctuations, with a normal range of 0 to 3.0 for their comprehensive anomaly coefficient. The normal range of the comprehensive anomaly coefficient for wind speed and direction sensors is 0 to 4.0. The element anomaly feature database defines a standard anomaly amplitude range for each type of sensor anomaly for each type of marine special event, which is used to determine whether the sensor anomaly is a normal anomaly caused by the marine special event or a real anomaly caused by sensor malfunction.

[0056] The monitoring time points of each target anomaly element within the monitoring period of each marine observation station are spatiotemporally matched with the occurrence time periods of each marine special event. If a certain monitoring time point of a target anomaly element falls within the occurrence time period of a marine special event, a correspondence is established between the target anomaly element and the marine special event, thereby obtaining the marine special events corresponding to each target anomaly element within the monitoring period of each marine observation station.

[0057] Based on the comprehensive anomaly coefficients of each target anomaly element at each marine observation station during the monitoring period and the corresponding element anomaly feature database of each marine special event, a matching analysis is performed to calculate the sea state correlation degree of each target anomaly element at each marine observation station during the monitoring period, and based on this, each target anomaly element that needs to be removed is selected.

[0058] The anomaly elements of each target retained by each marine observation station during the monitoring period are summarized as the final anomaly elements. Based on this, the data quality deviation rate of each marine observation station at each monitoring time point is calculated, and curve fitting is performed to generate the quality degradation characteristic curve of each marine observation station.

[0059] In one specific embodiment, the data quality deviation rate of each marine observation station at each monitoring time point is calculated, and curve fitting is performed to generate the quality degradation characteristic curve of each marine observation station. The specific method is as follows: Based on each final abnormal element of each marine observation station within the monitoring period, the comprehensive abnormal coefficient corresponding to each final abnormal element is extracted. For each monitoring time point, it is searched to see if there is a monitoring time point where a final abnormal element appears. If there is a final abnormal element at the monitoring time point, the comprehensive abnormal coefficient of all final abnormal elements at the monitoring time point is extracted. These comprehensive abnormal coefficients are summed and then divided by the total number of observed elements to obtain the data quality deviation rate of each marine observation station at each monitoring time point. If there is no final abnormal element at the monitoring time point, the data quality deviation rate of the monitoring time point is zero. With each monitoring time point within the monitoring period as the horizontal axis and the corresponding data quality deviation rate as the vertical axis, curve fitting is performed on each data point using an exponential function. The optimal fitting curve is calculated by the fitting algorithm, and this curve is the quality degradation characteristic curve of each marine observation station.

[0060] In a specific embodiment of the present invention, the correlation degree of sea state of each target anomaly element of each marine observation station within the monitoring period is calculated, and the target anomaly elements to be removed are screened accordingly. The specific method is as follows: based on each marine special event corresponding to each target anomaly element of each marine observation station within the monitoring period, the standard anomaly range of each marine special event is extracted from the element anomaly feature database.

[0061] Based on the comprehensive anomaly coefficients of each target anomaly element at each marine observation station during the monitoring period, they are compared with the standard anomaly amplitude range of each corresponding marine special event. If the comprehensive anomaly coefficient of a target anomaly element falls within the standard anomaly amplitude range of a marine special event, it is marked as amplitude matching. The number of amplitude matching times for each target anomaly element at each marine observation station is counted.

[0062] Based on the number of amplitude matching times of each target anomaly element at each marine observation station during the monitoring period, and combined with the number of marine special events corresponding to each target anomaly element, the sea state correlation degree of each target anomaly element at each marine observation station during the monitoring period is calculated.

[0063] In one specific embodiment, the sea state correlation degree of each target anomaly element of each marine observation station within the monitoring period is calculated by dividing the number of amplitude matching times of each target anomaly element of each marine observation station within the monitoring period by the number of occurrences of marine special events corresponding to each target anomaly element, thereby obtaining the sea state correlation degree of each target anomaly element of each marine observation station within the monitoring period.

[0064] The sea state correlation threshold is obtained from the local database. If the sea state correlation of a certain target anomalous element exceeds the sea state correlation threshold, the target anomalous element is marked as a target anomalous element that needs to be removed, thereby obtaining each target anomalous element that needs to be removed.

[0065] The intelligent quality control and scheduling module is used to predict equipment fouling based on the quality degradation characteristic curves of each marine observation station, deduce the graded quality control and maintenance plan for the marine observation station, and send it to the observation station network management center.

[0066] In a specific embodiment of the present invention, equipment fouling is predicted, and a graded quality control maintenance scheme for marine observation stations is derived. The specific method is as follows: the last maintenance and cleaning time of each marine observation station is obtained from the local database, and the time since the last maintenance of each marine observation station is calculated.

[0067] Based on the quality degradation characteristic curves of each marine observation station and the maintenance analysis of the time since the last maintenance, a graded quality control maintenance plan for marine observation stations is derived.

[0068] In one specific embodiment, maintenance levels are classified to deduce a graded quality control maintenance scheme for marine observation stations. The specific method is as follows: the quality degradation characteristic curves of each marine observation station are analyzed; the derivative is calculated at each monitoring time point on the curve to obtain the slope value of the quality degradation characteristic curve at each point; and the average value of all slope values ​​is calculated as the average quality degradation rate of each marine observation station. Based on the time elapsed since the last maintenance at each marine observation station .

[0069] Calculate the comprehensive maintenance requirement coefficient for each marine observation station. , where y represents the number of ocean observation stations.

[0070] The marine observation stations are arranged in descending order of comprehensive maintenance demand coefficient, and then maintained in sequence to form a graded quality control maintenance plan for the marine observation stations.

[0071] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0072] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. An intelligent control system for multi-source data quality of a network of ocean observing stations, characterized in that, include: The marine environmental feature extraction module is used to acquire multi-source observation data from each marine observation station at each monitoring time point during the monitoring period, thereby analyzing the marine environmental dynamic characteristic parameters of each marine observation station at each monitoring time point. The multi-element coupling verification module is used to construct the marine element coupling correlation matrix of each marine observation station within the monitoring time period based on the multi-source observation data of each marine observation station at each monitoring time point, thereby identifying each coupling anomalous element of each marine observation station within the monitoring time period. The spatiotemporal gradient field quality control module is used to detect gradient field anomalies based on the coupled anomalies of each marine observation station within the monitoring period, thereby screening the target anomalies of each marine observation station within the monitoring period. The sea state impact assessment module is used to identify the types and occurrence times of various special marine events at each marine observation station during the monitoring period based on the marine environmental dynamics characteristic parameters of each marine observation station at each monitoring time point, and to perform sea state correlation analysis in conjunction with the various target anomaly elements of each marine observation station during the monitoring period, thereby generating the quality degradation characteristic curves of each marine observation station. The specific method for identifying the type and time period of each special marine event at each marine observation station within the monitoring period is as follows: Historical tide level and density data of each marine observation station are obtained from the local database. The historical tide level quantile distribution of each marine observation station in each month and time period is statistically analyzed to construct a tide level benchmark database for each marine observation station. The historical density quantile distribution of each marine observation station in each month is statistically analyzed to construct a density benchmark database for each marine observation station. Based on the theoretical tide levels of each marine observation station at each monitoring time point, and in conjunction with the tide level reference database of each marine observation station, the tide level deviation index of each marine observation station at each monitoring time point is calculated. Based on the seawater density at each monitoring time point of each marine observation station, and in conjunction with the density benchmark library of each marine observation station, the density deviation index of each marine observation station at each monitoring time point is calculated, and the seawater density change rate between each adjacent monitoring time point of each marine observation station is calculated. The system obtains a rule base for the association of special marine events from the local database. Based on the tide level deviation index, density deviation index and seawater density change rate between adjacent monitoring time points of each marine observation station, and cross-validates the rule base for the association of special marine events, it identifies the types and time periods of special marine events of each marine observation station within the monitoring period. The specific method for generating the quality degradation characteristic curves for each marine observation station is as follows: The system retrieves the element anomaly time feature library and element anomaly amplitude feature library corresponding to the type of each marine special event from the local database. Based on the type of each marine special event within the monitoring period of each marine observation station, it maps the element anomaly feature library corresponding to each marine special event of each marine observation station. The monitoring time points of each target anomaly element within the monitoring period of each marine observation station are spatiotemporally matched with the occurrence time periods of each marine special event. If a certain monitoring time point of a target anomaly element falls within the occurrence time period of a marine special event, a correspondence is established between the target anomaly element and the marine special event, thereby obtaining the marine special events corresponding to each target anomaly element within the monitoring period of each marine observation station. Based on the comprehensive anomaly coefficient of each target anomaly element of each marine observation station within the monitoring period and the corresponding element anomaly feature database of each marine special event, a matching analysis is performed to calculate the sea state correlation degree of each target anomaly element of each marine observation station within the monitoring period, and the target anomaly elements that need to be removed are screened accordingly. The anomaly elements of each target retained by each marine observation station during the monitoring period are summarized as each final anomaly element. Based on this, the data quality deviation rate of each marine observation station at each monitoring time point is calculated, and curve fitting is performed to generate the quality degradation characteristic curve of each marine observation station. The intelligent quality control and scheduling module is used to predict equipment fouling based on the quality degradation characteristic curves of each marine observation station, deduce the graded quality control and maintenance plan for the marine observation station, and send it to the observation station network management center.

2. The net of ocean observing stations multi-source data quality intelligent control system according to claim 1, characterized in that, The specific method for analyzing the marine environmental dynamics characteristics parameters of each marine observation station at each monitoring time point is as follows: The tidal harmonic constants of the sea areas where each marine observation station is located are obtained from the local database, and the theoretical tide level values ​​of each marine observation station at each monitoring time point are evaluated in combination with each monitoring time point. Based on multi-source observation data from each marine observation station at each monitoring time point, the seawater density of each marine observation station at each monitoring time point is calculated, and it is unified with the theoretical tide level value of each marine observation station at each monitoring time point as the marine environmental dynamic characteristic parameter of each marine observation station at each monitoring time point.

3. The system according to claim 2, wherein, The specific method for constructing the coupling correlation matrix of marine elements for each marine observation station within the monitoring period is as follows: Based on multi-source observation data from various marine observation stations at various monitoring time points, the measured water temperature, measured salinity, measured wind speed, measured wind direction, measured current velocity, and measured current direction of each marine observation station at each monitoring time point were extracted. Historical observation data from each marine observation station were obtained from the local database. The historical salinity distribution patterns of each marine observation station under the same water temperature range were statistically analyzed to construct a statistical model of temperature-salinity relationship for each marine observation station. The historical flow velocity and flow direction distribution patterns of each marine observation station under the same wind speed and direction conditions were also statistically analyzed to construct a statistical model of wind flow relationship for each marine observation station. Based on the measured water temperature and salinity at each monitoring time point of each marine observation station, and combined with the temperature-salinity statistical relationship model of each marine observation station, the deviation was calculated to obtain the temperature-salinity coupling deviation coefficient of each marine observation station at each monitoring time point. Based on the measured wind speed and direction at each monitoring time point of each marine observation station, and combined with the wind flow statistical relationship model of each marine observation station, the deviation was calculated to obtain the wind flow coupling deviation coefficient of each marine observation station at each monitoring time point. The temperature-salinity coupling deviation coefficient and wind-current coupling deviation coefficient of each marine observation station at each monitoring time point are constructed into a matrix, thereby constructing the marine element coupling correlation matrix of each marine observation station within the monitoring time period.

4. The system according to claim 3, wherein, The specific method for identifying the coupled anomaly elements of each marine observation station within the monitoring period is as follows: The temperature-salinity coupling deviation threshold and the wind-current coupling deviation threshold are obtained from the local database. Based on the temperature-salinity coupling deviation coefficient and the wind-current coupling deviation coefficient of each marine observation station at each monitoring time point, if the temperature-salinity coupling deviation coefficient of a certain marine observation station at a certain monitoring time point exceeds the temperature-salinity coupling deviation threshold, the water temperature sensor and the salinity sensor are marked as coupling anomalies, and the monitoring time point is taken as the occurrence monitoring time point. Thus, the occurrence monitoring time points of water temperature coupling anomalies and salinity coupling anomalies of each marine observation station are obtained. If the wind-flow coupling deviation coefficient of a certain marine observation station exceeds the wind-flow coupling deviation threshold at a certain monitoring time point, then the wind speed and wind direction sensor and the current velocity and current direction sensor at that monitoring time point are marked as coupling anomalies, thereby obtaining the monitoring time points at which the wind speed and wind direction coupling anomalies and the current velocity and current direction coupling anomalies occur at each marine observation station. The above data were compiled and used as the coupling anomaly elements of each marine observation station during the monitoring period.

5. The system of claim 4, wherein, The specific method for screening the anomaly elements of each target at each marine observation station within the monitoring period is as follows: Based on the coupled anomaly elements of each marine observation station during the monitoring period, the frequency of anomalies of each coupled anomaly element of each marine observation station during the monitoring period was statistically analyzed. Based on the monitoring time points of each coupled anomaly element of each marine observation station within the monitoring period, calculate the anomaly time interval sequence of each coupled anomaly element of each marine observation station within the monitoring period. Based on the anomalous time interval sequence of each coupled anomalous element of each marine observation station within the monitoring period, the temporal gradient acceleration coefficient of each coupled anomalous element of each marine observation station within the monitoring period is evaluated. Combined with the frequency of anomalous occurrence of each coupled anomalous element of each marine observation station within the monitoring period, the comprehensive anomalous coefficient of each coupled anomalous element of each marine observation station within the monitoring period is calculated. Based on this, each target anomalous element of each marine observation station within the monitoring period is selected.

6. The net of ocean observing stations multi-source data quality intelligent control system according to claim 5, characterized in that, The specific method for evaluating the temporal gradient acceleration coefficients of various coupled anomaly elements at each marine observation station within the monitoring period is as follows: Based on the abnormal time interval sequence of each coupled anomaly element of each marine observation station within the monitoring period, the time interval values ​​of each coupled anomaly element of each marine observation station within the monitoring period between two consecutive anomalies are extracted, and the time interval change rate sequence of each coupled anomaly element of each marine observation station within the monitoring period is summarized accordingly. Based on the time interval change rate sequence of each coupled anomaly element of each marine observation station within the monitoring period, the anomalous acceleration trend fitting curve of each coupled anomaly element of each marine observation station within the monitoring period is fitted, and the slope values ​​of each coupled anomaly trend fitting curve of each coupled anomaly element of each marine observation station are extracted, thereby calculating the temporal gradient acceleration coefficient of each coupled anomaly element of each marine observation station within the monitoring period.

7. The netted ocean observing station multi-source data quality intelligent control system according to claim 1, characterized in that, The specific method for calculating the sea state correlation degree of each target anomaly element of each marine observation station within the monitoring period, and selecting target anomalies to be removed accordingly, is as follows: Based on the specific marine events corresponding to the anomalies of each target within the monitoring period of each marine observation station, the standard anomaly range of each specific marine event is extracted from the feature anomaly database. Based on the comprehensive anomaly coefficients of each target anomaly element at each marine observation station during the monitoring period, they are compared with the standard anomaly amplitude range of each corresponding marine special event. If the comprehensive anomaly coefficient of a target anomaly element falls within the standard anomaly amplitude range of a marine special event, it is marked as amplitude matching. The number of amplitude matching times for each target anomaly element at each marine observation station is counted. Based on the number of amplitude matching times of each target anomaly element at each marine observation station during the monitoring period, and combined with the number of occurrences of marine special events corresponding to each target anomaly element, the sea state correlation degree of each target anomaly element at each marine observation station during the monitoring period is calculated. The sea state correlation threshold is obtained from the local database. If the sea state correlation of a certain target anomalous element exceeds the sea state correlation threshold, the target anomalous element is marked as a target anomalous element that needs to be removed, thereby obtaining each target anomalous element that needs to be removed.

8. The net of ocean observing stations multi-source data quality intelligent control system according to claim 1, characterized in that, The specific method for predicting equipment fouling and deriving a graded quality control and maintenance scheme for marine observation stations is as follows: The last maintenance and cleaning time of each marine observation station is obtained from the local database, and the time since the last maintenance is calculated for each marine observation station. Based on the quality degradation characteristic curves of each marine observation station and the maintenance analysis of the time since the last maintenance, a graded quality control maintenance plan for marine observation stations is derived.

Images