A method, medium and system for real-time observation data alarm of ocean station business

Abstract

The present invention provides a method, medium and system for real-time observation data alarm of ocean stations, which belongs to the field of ocean data detection technology, including: first, according to the basic data of different observation elements (such as sea temperature, waves, tides, etc.), a coupling equation group is established to describe the relationship between them. Then, the abnormal data is identified by the threshold method, and it is substituted into the coupling equation group to calculate the deviation between the theoretical value and the measured value. Next, the suspected abnormal data is further screened based on the deviation rate analysis, and the abnormal pattern and systematic error are identified by cluster analysis. Finally, based on the above analysis results, each observation element is scored, and the data below the standard is identified as unqualified data. The system adopts multidimensional data analysis means, which can effectively identify and eliminate abnormal values ​​and systematic errors in ocean observation data, improve data quality, provide a reliable basis for subsequent marine environment monitoring and analysis, and solve the technical problem of low efficiency of manual review.

Description

Technical Field

[0001] The present invention belongs to the technical field of ocean data detection, and in particular, relates to an ocean station business real-time observation data alarm method, medium and system. Background Art

[0002] With the rapid improvement of ocean observation and simulation methods and major breakthroughs in data science, modern ocean science has undergone three paradigm changes: theory-driven, technology-driven, and data-driven. In-depth exploration and rational use of massive data in the ocean have become the only way for humans to move from understanding the ocean to managing the ocean. Therefore, building a strong maritime nation and developing marine science and technology are increasingly dependent on reliable and accurate observation data.

[0003] As one of the important components of the ocean observation network, the ocean station is important because of its timeliness, accuracy, continuity and long-term nature. However, due to the influence of environment, equipment, transmission, etc., there will be certain errors in the observation data, which will be accumulated and amplified in the time dimension. Therefore, before the ocean station observation data is used, it must be subjected to necessary quality control and manual review. At present, the real-time observation data of the ocean station is mainly manually reviewed, and the delayed data is quality controlled.

[0004] With the increase in the frequency of real-time observation data from ocean stations, conventional manual review and delayed quality control can no longer meet normal work needs. Real-time quality control management of ocean station observation data, timely detection of erroneous data, suspicious data or abnormal data information in the data, and verification and modification of the data according to prompts can minimize errors in the data and ensure the reliability, representativeness and comparability of the data. Summary of the invention

[0005] In view of this, the present invention provides a method, medium and system for real-time observation data alarm of an ocean station, which can solve the technical problems of low efficiency and difficulty in timely detection of erroneous data, suspicious data or abnormal data information in the current conventional manual review.

[0006] The present invention is achieved in that:

[0007] A first aspect of the present invention provides a method for alarming real-time observation data of an ocean station, which comprises the following steps:

[0008] S10, continuously acquiring multiple groups of operational real-time observation data of ocean stations, including basic data and derived data, both of which include multiple different types of observation elements, the basic data include sea surface temperature, air temperature, one-tenth wave height of waves, tide data, and wind field data, the derived data include wave height variability, wave height accumulation rate, wind-wave relationship, tide accumulation rate, tide change accumulation, water increase and decrease, water increase and decrease accumulation rate, and water increase and decrease change accumulation;

[0009] S20. Based on the basic data and the derived data, a coupling equation group describing the relationship between different types of observation elements is established, including a temperature coupling equation, a wave coupling equation, a tide coupling equation, a wind-wave relationship equation, and a tide-increase-decrease relationship equation; optionally, the coupling equation group also includes a temperature-wave relationship equation and a temperature-tide relationship equation;

[0010] S30, using the threshold method to mark abnormal data for different types of observation elements;

[0011] S40, bringing the data marked with abnormalities and other data at corresponding moments into the coupling equation group, and solving to obtain the deviation between the theoretical value and the actual observed value of each observation element;

[0012] S50, calculating the deviation rate of each observation element, and marking the data whose deviation rate exceeds a preset threshold as suspected abnormal data;

[0013] S60. Perform cluster analysis on suspected abnormal data to identify abnormal patterns and potential systematic errors;

[0014] S70. Combine the threshold method and coupled equation analysis results to give a comprehensive score to each observation element, and the data with a score lower than the set standard will be ultimately marked as unqualified data.

[0015] Wherein, the temperature coupling equation is specifically expressed as:

[0016] ;

[0017] In the formula, is the sea surface temperature (°C); is the air temperature (°C); is the daily temperature variation amplitude (°C), obtained by fitting historical data; is the angular frequency, equal to (rad / h); is time (h); is the wave influence coefficient, obtained through regression analysis; is one tenth of the wave height (m); is the average value of one tenth of the wave height (m); is the wind speed influence coefficient, obtained through regression analysis; is the wind speed (m / s); is the average wind speed (m / s); is the air pressure influence coefficient, obtained through regression analysis; is atmospheric pressure (hPa); is the average value of atmospheric pressure (hPa); is the error term of the temperature coupling equation, which obeys the normal distribution .

[0018] The wave coupling equation is specifically expressed as:

[0019] ;

[0020] In the formula, is one tenth of the wave height (m); is the wind speed (m / s); is the length of the wind zone (m); is the water depth (m); is the acceleration due to gravity (m / s 2 ); is the sea surface temperature (°C); is the mean sea surface temperature (°C); is atmospheric pressure (hPa); is the average value of atmospheric pressure (hPa); is the relative humidity (%); is the tidal angular frequency, equal to (rad / h); is time (h); is the fitting coefficient, obtained by nonlinear regression; is the error term of the wave coupling equation, which obeys the normal distribution .

[0021] The tidal coupling equation is specifically expressed as:

[0022] ;

[0023] In the formula, is the tide level (m); is the mean sea level (m); is the amplitude of the i-th partial tide (m); is the slow-changing factor of the i-th tidal component, which is a function of time; is the angular frequency of the i-th tidal component (rad / h); is the phase of the i-th tidal component (rad); is the number of tidal subdivisions considered; is atmospheric pressure (hPa); is the average value of atmospheric pressure (hPa); is the wind speed (m / s); is the angle between wind direction and coastline (rad); is one tenth of the wave height (m); is the sea surface temperature (°C); is the mean sea surface temperature (°C); is the fitting coefficient, obtained by multiple linear regression; is the error term of the tidal coupling equation, which obeys the normal distribution .

[0024] The wind-wave relationship equation is specifically expressed as:

[0025] ;

[0026] In the formula, is one tenth of the wave height (m); is the wind speed (m / s); is the length of the wind zone (m); is the water depth (m); is the wavelength (m), which can be calculated by the dispersion relation as ; is the wave period (s); is the fitting coefficient, obtained by nonlinear regression; is the error term of the wind-wave relationship equation, which obeys the normal distribution .

[0027] The tide level-increase and decrease water relationship equation is specifically expressed as:

[0028] ;

[0029] In the formula, To increase or decrease the water height (m); is atmospheric pressure (hPa); is standard atmospheric pressure (1013.25hPa); is the wind speed (m / s); is the angle between wind direction and coastline (rad); is the tide level (m); is time (h); is one tenth of the wave height (m); is the sea surface temperature (°C); is the reference temperature (usually 15°C); is the seawater salinity (psu); is the fitting coefficient, obtained by multiple linear regression; is the error term of the tide level-increase and decrease water relationship equation, which obeys the normal distribution .

[0030] The temperature-wave relationship equation is specifically expressed as:

[0031] ;

[0032] In the formula, is the sea surface temperature (°C); is the air temperature (°C); is the dew point temperature (°C), which can be calculated from the relative humidity and air temperature; is one tenth of the wave height (m); is the wind speed (m / s); is the angular frequency, equal to (rad / h); is time (h); is atmospheric pressure (hPa); is standard atmospheric pressure ( 1013.25 hPa); is the fitting coefficient, obtained by nonlinear regression; is the error term of the temperature-wave relationship equation, which obeys the normal distribution .

[0033] The temperature-tide relationship equation is specifically expressed as:

[0034] ;

[0035] In the formula, is the sea surface temperature (°C); is the reference temperature (°C), obtained by averaging long-term observation data; is the daily temperature variation amplitude (°C); is the angular frequency of daily temperature variation, equal to (rad / h); is time (h); is the diurnal temperature variation phase (rad); is the tide level (m); is the wind speed (m / s); is atmospheric pressure (hPa); is standard atmospheric pressure ( 1013.25 hPa); is the net heat flux (W / m 2), which can be calculated by the radiation balance equation; is the fitting coefficient, obtained by multiple linear regression; is the error term of the temperature-tidal relationship equation, which obeys the normal distribution .

[0036] The threshold method is used to mark abnormal data for different types of observation elements, including at least the following four data abnormality markings:

[0037] (1) Tidal level accumulation rate

[0038] The cumulative probability statistics of the historical tide levels observed at the ocean station are:

[0039] ;

[0040] in For a set of tide data The tide level corresponding to the cumulative probability is, for Cumulative probability: select the tide level corresponding to 1% cumulative probability as the marking threshold, and mark it when the observed tide level exceeds this threshold;

[0041] (2) Cumulative rate of tidal level change

[0042] The cumulative probability statistics of the hourly changes of historical tide levels observed at the ocean station are:

[0043] ;

[0044] in is the hourly tidal level change value, , and are the tide levels at time t and time t-1 respectively; is a set of hourly tidal level change data The hourly tide level change value corresponding to the cumulative probability, for Cumulative probability: select the hourly tidal level change value corresponding to 1% cumulative probability as the marking threshold, and mark it when the hourly tidal level change value obtained by analysis exceeds the threshold;

[0045] (3) Tidal level accumulation rate

[0046] The cumulative probability statistics of the historical increase and decrease of water at the ocean station are:

[0047] ;

[0048] in For a set of water increase and decrease data The increase or decrease in water value corresponding to the cumulative probability, for Cumulative probability: select the water increase or decrease value corresponding to 1% cumulative probability as the marking threshold, and mark it when the water increase or decrease value exceeds this threshold;

[0049] (4) Cumulative rate of water increase and decrease

[0050] The cumulative probability statistics of the hourly changes in the historical increase and decrease of water at the ocean station are:

[0051] ;

[0052] in To increase or decrease the water change value hour by hour, , and are the increase and decrease water values ​​at time t and time t-1 respectively; is a set of hourly water change data The hourly increase or decrease in water value corresponding to the cumulative probability, for Cumulative probability: select the hourly increase or decrease water change value corresponding to 1% cumulative probability as the marking threshold, and mark it when the hourly increase or decrease water change value obtained by analysis exceeds the threshold.

[0053] Specifically, the step S10 includes:

[0054] Step 101: Establish a real-time data collection system to continuously collect multiple observation element data of the ocean station through a sensor network;

[0055] Step 102: Perform preliminary processing on the collected data, including unit conversion and outlier removal;

[0056] Step 103: store the processed data in a database according to a predefined data structure, where each record contains a timestamp, an observation element type, and a corresponding value;

[0057] Step 104: Establish a data backup and synchronization mechanism to ensure data security and reliability;

[0058] Step 105: Implement a data query interface to facilitate subsequent analysis programs to call required data.

[0059] Wherein, the step S20 specifically includes:

[0060] Step 201: Calculate derived data based on the collected basic data, and use statistical analysis methods to explore the relationship between different observation elements;

[0061] Step 202: Based on the theoretical knowledge of physical oceanography and atmospheric science and combined with the data analysis results, a coupled equation group describing the relationship between different observation elements is constructed;

[0062] Step 203, using nonlinear least squares method, multivariate linear regression and other methods to determine the parameters in the coupled equations;

[0063] Step 204: Perform model diagnosis, including residual analysis and multicollinearity test, to ensure the validity and reliability of the model.

[0064] Wherein, the step S30 specifically includes:

[0065] Step 301: Calculate the statistical characteristics of each observation element based on historical data, including mean, standard deviation and quantile;

[0066] Step 302: setting a reasonable threshold range for each observation element according to statistical characteristics and professional knowledge;

[0067] Step 303: Implement a dynamic adjustment mechanism for the threshold value, capable of updating the threshold value according to new observation data and analysis results;

[0068] Step 304: Compare the observed data with the set threshold, and mark the data points that exceed the threshold range as abnormal data.

[0069] Wherein, the step S40 specifically includes:

[0070] Step 401: extract all observation element data at the corresponding time from the data points marked as abnormal;

[0071] Step 402, substituting the extracted data into the established coupled equations;

[0072] Step 403: Use numerical methods to solve the coupled equations to obtain theoretical values ​​of each observation element;

[0073] Step 404: Compare the theoretical value with the actual observed value and calculate the deviation of each observed element.

[0074] Wherein, the step S50 specifically includes:

[0075] Step 501: Calculate the relative deviation rate of each observation element based on the calculated deviation;

[0076] Step 502: setting a preset threshold value of deviation rate for each observation element;

[0077] Step 503: compare the deviation rate of each observation element with its corresponding preset threshold;

[0078] Step 504: Mark the data points whose deviation rate exceeds a preset threshold as suspected abnormal data;

[0079] Step 505: Store the marking results in the database, including the timestamp of the data point, the type of observation element, the actual observation value, the theoretical calculation value, the deviation rate and the abnormal mark.

[0080] Wherein, the step S60 specifically includes:

[0081] Step 601: extract features from suspected abnormal data points, and perform data standardization and missing value processing;

[0082] Step 602: Select a density clustering algorithm to perform cluster analysis on the processed data;

[0083] Step 603: Analyze the clustering results to identify major abnormal patterns and isolated abnormal events;

[0084] Step 604: Combine domain knowledge to explain the physical meaning of each cluster and infer possible abnormal causes.

[0085] Wherein, the step S70 specifically includes:

[0086] Step 701, establish a comprehensive scoring system for each observation element, taking into account the results of the threshold method, the coupled equation analysis results and the cluster analysis results;

[0087] Step 702: Calculate a comprehensive score for each data point;

[0088] Step 703: Setting a scoring threshold to ultimately determine whether the data is qualified;

[0089] Step 704: Compare the score of each data point with the threshold, and mark the data points with scores below the threshold as unqualified data;

[0090] Step 705: Process the unqualified data, including complete elimination, interpolation replacement, model estimation or marking retention;

[0091] Step 706: Store the scoring results and processed data in the database, including the timestamp, observation element type, original value, score, quality mark and processed value of each data point.

[0092] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are executed, they are used to execute the above-mentioned method for real-time observation data alarm for ocean station operations.

[0093] A third aspect of the present invention provides a marine station operational real-time observation data alarm system, which includes the above-mentioned computer-readable storage medium.

[0094] Compared with the prior art, the beneficial effects of the ocean station business real-time observation data alarm method, medium and system provided by the present invention are:

[0095] 1. Comprehensiveness: The method of the present invention not only considers the threshold detection of a single observation element, but also establishes a set of coupled equations that describe the relationship between different types of observation elements. This multi-factor coupling analysis method can fully capture the complex interactions between marine environmental elements, greatly improving the accuracy and reliability of anomaly detection.

[0096] 2. Adaptability: By introducing a dynamic threshold adjustment mechanism and coupled equation analysis, the method of the present invention can better adapt to the dynamic changes of the marine environment. Compared with the fixed threshold method, this method can dynamically adjust the judgment criteria according to real-time data and historical statistical characteristics, significantly reducing the probability of misjudgment and missed judgment.

[0097] 3. Efficiency: The method of the present invention uses advanced data processing and analysis technologies, such as density clustering algorithm and multidimensional feature analysis, which can efficiently process large-scale real-time observation data. This not only improves the speed of data processing, but also can timely discover and identify abnormal patterns, providing strong support for real-time monitoring and early warning.

[0098] In summary, the solution of the present invention solves the technical problems of low efficiency and difficulty in timely discovering erroneous data, suspicious data or abnormal data information in the current conventional manual review. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] Figure 1 A flow chart of the method provided by the present invention. DETAILED DESCRIPTION

[0100] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0101] like Figure 1 As shown, it is a flow chart of a method for real-time observation data alarm of an ocean station provided by the present invention. The method comprises the following steps:

[0102] S10, continuously acquiring multiple groups of operational real-time observation data of ocean stations, including basic data and derived data, both of which include multiple different types of observation elements, the basic data include sea surface temperature, air temperature, one-tenth wave height of waves, tide data, and wind field data, the derived data include wave height variability, wave height accumulation rate, wind-wave relationship, tide accumulation rate, tide change accumulation, water increase and decrease, water increase and decrease accumulation rate, and water increase and decrease change accumulation;

[0103] S20, based on the basic data and the derived data, establishing a coupling equation group describing the relationship between different types of observation elements, including a temperature coupling equation, a wave coupling equation, a tide coupling equation, a wind-wave relationship equation, and a tide-increase-decrease relationship equation. Optionally, the coupling equation group also includes a temperature-wave relationship equation and a temperature-tide relationship equation;

[0104] S30, using the threshold method to mark abnormal data for different types of observation elements;

[0105] S40, bringing the data marked with abnormalities and other data at corresponding moments into the coupled equation group, and solving to obtain the deviation between the theoretical value and the actual observed value of each observation element;

[0106] S50, calculating the deviation rate of each observation element, and marking the data whose deviation rate exceeds a preset threshold as suspected abnormal data;

[0107] S60. Perform cluster analysis on suspected abnormal data to identify abnormal patterns and potential systematic errors;

[0108] S70. Combine the threshold method and coupled equation analysis results to give a comprehensive score to each observation element, and the data with a score lower than the set standard will be ultimately marked as unqualified data.

[0109] The specific implementation methods of the above steps are described in detail below:

[0110] The specific implementation method of step S10 is: first, establish a real-time data acquisition system to continuously collect various observation element data of the ocean station through a sensor network. These sensors include but are not limited to temperature sensors, wave meters, tide meters, etc. The data acquisition frequency is set according to the changing characteristics of different elements, usually once an hour or more frequently. Secondly, the collected data is preliminarily processed, including unit conversion, outlier removal, etc. Then, the processed data is stored in the database according to a predefined data structure, and each record contains a timestamp, an observation element type, and a corresponding numerical value. In addition, a data backup and synchronization mechanism needs to be established to ensure the security and reliability of the data. Finally, a data query interface is implemented to facilitate the subsequent analysis program to call the required data. The main purpose of this step is to establish a stable and reliable data foundation to provide necessary input for subsequent data analysis and model building.

[0111] The specific implementation method of step S20 is: first, based on the basic data and derived data collected in S10, statistical analysis methods such as correlation analysis and principal component analysis are used to preliminarily explore the relationship between different observation elements. Then, according to the theoretical knowledge of physical oceanography and atmospheric science, combined with the data analysis results, a coupling equation group describing the relationship between different observation elements is constructed. These equation groups include temperature coupling equations, wave coupling equations, tidal coupling equations, wind-wave relationship equations, tide-increase and decrease water relationship equations, temperature-wave relationship equations, and temperature-tidal relationship equations. When constructing these equations, nonlinear least squares methods, multiple linear regression methods, and other methods are used to determine the parameters in the equations. For example, for the temperature coupling equation, the nonlinear least squares method is used to fit the daily temperature variation amplitude α, the wave influence coefficient β, the wind speed influence coefficient γ, and the pressure influence coefficient δ. For the wave coupling equation, the nonlinear regression method is used to determine the coefficients a, b, c, d, e, f, g, and h. For the tidal coupling equation, the harmonic analysis method is first used to determine the amplitude and phase of the main tidal components, and then the multiple linear regression is used to determine the coefficients k, l, m, and n. The coefficients αw, βw, γw, and δw in the wind-wave relationship equation are obtained through nonlinear regression. The coefficients in the tide-increase and decrease water relationship equation and the temperature-tide relationship equation are determined by multiple linear regression. In the parameter estimation process, it is necessary to pay attention to the distribution characteristics of the data, and perform data conversion or use robust regression methods when necessary. In addition, model diagnosis is also required, including residual analysis, multicollinearity test, etc., to ensure the validity and reliability of the model. The main purpose of this step is to establish a mathematical model that can describe the complex interactions between marine environmental elements and provide a theoretical basis for subsequent anomaly detection.

[0112] The specific implementation of step S30 is as follows: For each observation element, first calculate its statistical characteristics based on historical data, including mean, standard deviation, quantiles, etc. Then, set a reasonable threshold range according to these statistical characteristics and professional knowledge. For example, for sea surface temperature and air temperature, the following method can be adopted: Collect historical data for the past 10 years and calculate the daily maximum temperature and minimum temperature. For any given day, take the maximum and minimum temperatures of the same day in the previous and next three days (a total of 7 days) over the years as samples, and calculate the mean and standard deviation of the maximum and minimum temperatures respectively. Set the daily maximum temperature threshold as the mean of the maximum temperature plus 2 times the standard deviation, and the minimum threshold as the mean of the minimum temperature minus 2 times the standard deviation. For wave data, thresholds can be set based on wave height variability, wave height accumulation rate, and wind-wave relationship. For example, divide the wave height into three intervals (0 ≤ H ≤ 0.5, 0.5 < H ≤ 1.0, and H > 1.0), and calculate the hourly wave height variability thresholds respectively so that the monthly alarm rate does not exceed 1%. Specifically, the hourly wave height variability alarm thresholds for these three intervals can be set to 120%, 60%, and 50% respectively. For tide data, thresholds can be set based on tide level accumulation rate, tide level change accumulation, water level rise and fall accumulation rate, and water level rise and fall change accumulation. For example, calculate the tide level corresponding to a 1% cumulative probability as the alarm threshold. During the threshold setting process, the influence of factors such as seasonal changes and extreme weather events needs to be considered, and different thresholds may need to be set for different periods. In addition, a flexible threshold adjustment mechanism needs to be implemented to dynamically update the thresholds according to new observation data and analysis results. Finally, compare the observation data with the set thresholds and mark the data points outside the threshold range as abnormal data. The main purpose of this step is to quickly identify obvious abnormal data points and provide preliminary screening results for subsequent detailed analysis.

[0113] The following takes sea surface temperature, air temperature, waves, and tides as examples to describe the threshold method annotation in detail:

[0114] (I) Threshold method annotation for sea surface temperature and air temperature

[0115] Considering the specific climate state characteristics, statistically analyze the daily maximum temperature and minimum temperature of this station in the past 10 years, and use the maximum and minimum temperatures of the same day in the previous three days and the next three days over the years as statistical samples to calculate the standard deviation of the maximum and minimum temperatures respectively.

[0116] The daily maximum temperature threshold is designed as the average of the maximum temperatures in the previous three days and the next three days of the current day + 2 * standard deviation, and the daily minimum temperature threshold is designed as the average of the minimum temperatures in the previous three days and the next three days of the current day - 2 * standard deviation.

[0117] Standard deviation calculation: ;

[0118] Among them, is the maximum / minimum value of the temperature in the previous and next three days, It is the average of the highest / lowest temperature values.

[0119] 2. Threshold Method for Wave Labeling

[0120] The annotation of wave observations combines the wave height variability, wave height accumulation rate and wind-wave relationship, takes into account the local climate characteristics, and uses the collected long-term wave observation data to set the annotation threshold. The observation elements are one-tenth of the wave height and the wind speed at 10 meters.

[0121] (1) Wave height variability

[0122] Statistical hourly wave height variation rate of one tenth of wave height observation data:

[0123] ;

[0124] in is the hourly wave height variation rate, and are the wave heights at time t and time t-1 respectively.

[0125] When the hourly wave height variation rate exceeds a certain threshold, the system starts marking. Since the variation rates of different wave height ranges vary greatly, the wave height must be divided into , and The thresholds of hourly wave height variability are calculated for each of the three intervals. The thresholds are set so that the monthly annotation rate does not exceed 1%.

[0126] For the ocean stations in the North Sea, the annotation thresholds for the hourly variation rate of wave height in the three intervals can be set to 120%, 60% and 50% respectively.

[0127] (2) Wave height accumulation rate

[0128] Perform cumulative probability statistics on historical observed wave heights:

[0129] ;

[0130] in is the wave height corresponding to the x% cumulative probability in a set of wave height data, The wave height corresponding to the cumulative probability of 1% is selected as the annotation threshold, and an annotation is performed when the observed wave height exceeds the threshold.

[0131] The wave height data for the whole year can be counted to obtain a universal annotated wave height, or the wave heights for each month can be counted separately to obtain the annotated wave heights for each month. Since the wave height corresponding to the 1% cumulative probability of each ocean station is different, statistics should be performed separately for each ocean station.

[0132] (3) Wind and wave relations

[0133] Considering that the wave height is generally well correlated with the wind speed, the relationship between the design wave height and wind speed of each ocean station is marked:

[0134] ;

[0135] Where H is the wave height, U is the wind speed, a and b are unknown coefficients. For wind direction, and They are the lower and upper wind direction thresholds for which the relationship applies. When the wave height and wind speed meet the above relationship, it is marked.

[0136] To determine the unknown coefficients a and b, the least squares method is first used to fit the following relationship between the observed wave height and wind speed:

[0137] ;

[0138] Thus, the unknown coefficient a is obtained. Then iterate b with a step size of -0.1 and calculate When the probability is equal to 1%, stop the iteration and determine the b value.

[0139] for and To determine the wind direction, the wind direction range corresponding to the seaward side of the ocean station should be selected.

[0140] 3. Tidal threshold marking

[0141] Tidal annotation combines the tidal level accumulation rate, tidal level change accumulation, water increase and decrease accumulation rate, and water increase and decrease change accumulation. Considering the local climate characteristics, the annotation threshold is set using the collected long-term tidal level observation data and the calculated water increase and decrease analysis data. Tidal observation annotation mainly targets two elements: tidal level and water increase and decrease. The observation element is the tidal level; the analysis element is the water increase and decrease. The astronomical tide is obtained by reconciling and analyzing the observed tidal level data, and the difference is obtained from the tidal level. The tide has a long-term cycle of 18.61 years. The selection of historical data is recommended to be 19 consecutive years of tidal level observation.

[0142] The threshold method is used to mark abnormal data for different types of observation elements, including at least the following four data abnormality markings:

[0143] (1) Tidal level accumulation rate

[0144] The cumulative probability statistics of the historical tide levels observed at the ocean station are:

[0145] ;

[0146] in For a set of tide data The tide level corresponding to the cumulative probability is, for Cumulative probability: select the tide level corresponding to 1% cumulative probability as the marking threshold, and mark it when the observed tide level exceeds this threshold;

[0147] (2) Cumulative rate of tidal level change

[0148] The cumulative probability statistics of the hourly changes of historical tide levels observed at the ocean station are:

[0149] ;

[0150] in is the hourly tidal level change value, , and are the tide levels at time t and time t-1 respectively; is a set of hourly tidal level change data The hourly tide level change value corresponding to the cumulative probability, for Cumulative probability: select the hourly tidal level change value corresponding to 1% cumulative probability as the marking threshold, and mark it when the hourly tidal level change value obtained by analysis exceeds the threshold;

[0151] (3) Tidal level accumulation rate

[0152] The cumulative probability statistics of the historical increase and decrease of water at the ocean station are:

[0153] ;

[0154] in For a set of water increase and decrease data The increase or decrease in water value corresponding to the cumulative probability, for Cumulative probability: select the water increase or decrease value corresponding to 1% cumulative probability as the marking threshold, and mark it when the water increase or decrease value exceeds this threshold;

[0155] (4) Cumulative rate of water increase and decrease

[0156] The cumulative probability statistics of the hourly changes in the historical increase and decrease of water at the ocean station are:

[0157] ;

[0158] in To increase or decrease the water change value hour by hour, , and are the increase and decrease water values ​​at time t and time t-1 respectively; is a set of hourly water change data The hourly increase or decrease in water value corresponding to the cumulative probability, for Cumulative probability: select the hourly increase or decrease water change value corresponding to 1% cumulative probability as the marking threshold, and mark it when the hourly increase or decrease water change value obtained by analysis exceeds the threshold.

[0159] The threshold annotation principles of the remaining parameters are the same as the above three items. In the specific implementation process, they can be used according to actual conditions.

[0160] The specific implementation method of step S40 is: first, obtain the data points marked as abnormal and their corresponding timestamps from S30. Then, for each abnormal data point, extract all the observation element data at the corresponding time. Next, substitute these data into the coupled equations established in S20. When solving these equations, it may be necessary to use numerical methods, such as the Newton-Raphson method or other iterative algorithms. After the solution is completed, the obtained theoretical value is compared with the actual observation value, and the deviation of each observation element is calculated. The deviation calculation formula is: Deviation = |Theoretical value - actual observation value|. The main purpose of this step is to conduct a more in-depth analysis of the abnormal data through the coupled equations, taking into account the mutual influence between the various observation elements, so as to more accurately evaluate the degree of abnormality of the data.

[0161] The specific implementation of step S50 is: first, for each observation element, based on the deviation calculated in S40, the relative deviation rate is calculated. The calculation formula of the deviation rate is: deviation rate = deviation / actual observation value × 100%. Then, a preset threshold is set to determine whether the deviation rate is too large. The setting of this threshold needs to take into account the characteristics of each observation element and the actual application requirements. For example, for sea temperature, the deviation rate threshold can be set to 5%; for one-tenth wave height, the deviation rate threshold can be set to 10%; for tide level, the deviation rate threshold can be set to 3%. These thresholds can be determined by expert experience or statistical analysis methods and should be adjustable. Next, the deviation rate of each observation element is compared with its corresponding preset threshold. If the deviation rate exceeds the preset threshold, the data point is marked as suspected abnormal data. In this process, it should be noted that different observation elements may have different tolerances, so the setting of the threshold should be performed separately for each observation element. In addition, it is also possible to consider introducing fuzzy logic or probabilistic methods to deal with situations where the deviation rate is close to the threshold, so as to avoid misjudgments that may be caused by hard thresholds. Finally, the labeling results are stored in the database, including the timestamp of the data point, the type of observation element, the actual observation value, the theoretical calculated value, the deviation rate, and the abnormal label. The main purpose of this step is to further screen out data points that may have problems through deviation rate analysis, and provide input for subsequent clustering analysis and comprehensive scoring.

[0162] The specific implementation of step S60 is: first, extract all data points marked as suspected anomalies from S50. Then, preprocess these data points, including data standardization and missing value processing. Data standardization can use the Z-score method, that is, subtract the mean of each feature and divide it by the standard deviation. For missing values, mean filling or more complex interpolation methods can be used. Next, select a suitable clustering algorithm for analysis. Considering the complexity of marine data and possible nonlinear relationships, the density clustering algorithm DBSCAN (Density-Based Spatial Clustering ofApplications with Noise) can be used. The advantage of the DBSCAN algorithm is that it does not require the number of clusters to be specified in advance and can identify clusters of any shape. The key parameters of the algorithm include ε (neighborhood radius) and MinPts (minimum number of points). These parameters can be determined by k-distance graph method or grid search. For example, ε can be set to 0.1 (normalized unit) and MinPts can be set to 5. When running the DBSCAN algorithm, a multidimensional feature space needs to be considered. Features can include observations, deviation rates, time information (such as time of day, season), etc. The algorithm will return the clustering results, including the cluster label and noise point markers for each data point. Then, the clustering results are analyzed and interpreted. First, the main abnormal patterns are identified, which correspond to larger clusters. For example, systematic deviations of certain observation elements may be found in a specific time period or under specific weather conditions. Secondly, the noise points are analyzed, which may represent isolated abnormal events or instrument failures. Finally, combined with domain knowledge, try to explain the physical meaning of each cluster and speculate on the possible causes of the anomaly. For example, a cluster may correspond to sensor drift, while another cluster may reflect the impact of extreme weather events. The main purpose of this step is to extract meaningful patterns from a large amount of suspected abnormal data through data mining technology, help understand the causes of the anomaly, and provide a basis for subsequent data quality improvements.

[0163] The specific implementation of step S70 is: first, establish a comprehensive scoring system for each observation element. This system should consider multiple factors, including the threshold method results in S30, the coupled equation analysis results in S40-S50, and the cluster analysis results in S60. The scoring system can adopt a weighted average method, and the specific weight can be determined by expert experience or machine learning methods (such as the feature importance of random forests). For example, the threshold method result can be given a weight of 30%, the coupled equation analysis result can be given a weight of 50%, and the cluster analysis result can be given a weight of 20%. Then, a comprehensive score is calculated for each data point. The score calculation can be as follows: Score = 0.3 × (1 - I_threshold) + 0.5 × (1- DR) + 0.2 × (1 - I_cluster), where I_threshold is the indicator function of the threshold method (1 if it exceeds the threshold, otherwise it is 0), DR is the standardized deviation rate, and I_cluster is the indicator function of the cluster analysis (1 if it belongs to an abnormal cluster, otherwise it is 0). The score ranges from 0 to 1, and the closer to 1, the better the data quality. Next, set a score threshold to finally determine whether the data is qualified. This threshold can be determined by analyzing the score distribution of historical data. For example, the 5% quantile of the score distribution can be selected as the threshold. Assume that after analysis, the threshold is set to 0.7. Then, the score of each data point is compared with the threshold. If the score is lower than the threshold, the data point is marked as unqualified data. For data marked as unqualified, further processing is required. The following solutions can be considered: (1) Complete elimination: directly delete unqualified data from the data set. (2) Interpolation replacement: use time series interpolation methods such as linear interpolation or spline interpolation to replace unqualified data. (3) Model estimation: use the coupled equations established in S20 to estimate the theoretical value of unqualified data based on the data of other observation elements, and replace the original value with this theoretical value. (4) Mark retention: retain the original data, but add a quality mark for reference in subsequent analysis. The choice of which processing solution should be determined according to the specific application scenario and data characteristics. Finally, the scoring results and processed data are stored in the database, including the timestamp, observation feature type, original value, score, quality flag, and processed value (if applicable) for each data point.

[0164] The main purpose of this step is to give a final quality assessment of each data point by comprehensively considering multiple analysis results, and to process the data appropriately accordingly. This approach can balance the consideration of different types of anomaly detection results and provide a more reliable data quality assessment.

[0165] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are executed, they are used to execute the above-mentioned method for real-time observation data alarm for ocean station operations.

[0166] A third aspect of the present invention provides a marine station operational real-time observation data alarm system, which includes the above-mentioned computer-readable storage medium.

[0167] Specifically, the principle of the present invention is:

[0168] Based on multi-factor coupling analysis and multi-dimensional data mining technology, a comprehensive, intelligent and efficient data quality control system is constructed. This method can effectively solve the problems existing in the existing technology, mainly based on the following key principles:

[0169] 1. Multi-factor coupling analysis: The various elements in the marine environment (such as sea temperature, air temperature, waves, tides, etc.) do not exist independently, but influence and restrict each other. The method of the present invention makes full use of these physical relationships by establishing a set of coupling equations that describe the relationship between different types of observation elements. For example, the heat exchange relationship between sea temperature and air temperature, the correlation between wind speed and wave height, the periodicity of tides and celestial motion, etc. This coupling analysis method can verify the rationality of data from multiple dimensions and greatly improve the accuracy of anomaly detection.

[0170] 2. Dynamic threshold adjustment: The marine environment has significant temporal and spatial variability, and fixed thresholds are often difficult to adapt to such dynamic changes. The method of the present invention dynamically adjusts the threshold range by analyzing the statistical characteristics of historical data and combining real-time observation data. This adaptive mechanism can better adapt to changes in the marine environment and reduce misjudgments and missed judgments.

[0171] 3. Density clustering algorithm: Traditional anomaly detection methods often have difficulty identifying complex anomaly patterns. This paper introduces a density clustering algorithm (such as DBSCAN), which can identify clusters of arbitrary shapes in high-dimensional feature space. This method is particularly suitable for processing various anomaly patterns that may exist in marine data, such as instrument failures, extreme weather events, etc.

[0172] 4. Multidimensional feature analysis: The method of the present invention not only considers the observation value itself, but also introduces multidimensional features such as time information and environmental factors. By analyzing in a high-dimensional feature space, the characteristics of the data can be more comprehensively characterized, and the accuracy of anomaly detection can be improved.

[0173] 5. Comprehensive scoring mechanism: By combining the results of threshold method, coupled equation analysis and cluster analysis, the method of the present invention establishes a comprehensive scoring system. This multi-angle and multi-level scoring mechanism can evaluate data quality more comprehensively and accurately.

[0174] 6. Modular design: The method of the present invention adopts modular design, including modules such as data acquisition, preprocessing, multidimensional analysis, and comprehensive scoring. This design makes the system have good scalability and adaptability, and can be flexibly adjusted according to the needs of different sea areas and observation tasks.

[0175] 7. Big data processing technology: Considering the large-scale characteristics of real-time observation data from ocean stations, the method of the present invention adopts efficient data processing and storage technologies, such as parallel computing, optimized database query, etc., to ensure that the system can process large amounts of data in real time.

[0176] The organic combination of these principles enables the method of the present invention to comprehensively, accurately and efficiently detect anomalies in the operational real-time observation data of ocean stations. Through multi-factor coupling analysis, the system can capture anomalies that may be overlooked by single-factor analysis; through dynamic thresholds and density clustering, the system can adapt to environmental changes and identify complex abnormal patterns; through a comprehensive scoring mechanism, the system can provide a more reliable data quality assessment.

[0177] In addition, the design of the method of the present invention meets the characteristics of ocean observation data and the requirements of quality control. It not only takes into account the statistical characteristics of the data, but also incorporates the basic principles of physical oceanography, so that the anomaly detection results have good physical interpretability. This idea of ​​combining data-driven methods with physical models not only ensures the scientific nature of the method, but also improves its practicality.

[0178] In order to better understand and implement the present invention, an embodiment of a specific application scenario of the present invention is provided below: This embodiment takes an ocean station in the Yellow Sea as an example to illustrate the specific implementation process of the ocean station business real-time observation data alarm method of the present invention. The ocean station is located at 36°03′ north latitude and 120°24′ east longitude, and is equipped with advanced ocean observation equipment, which can monitor multiple marine meteorological elements such as sea temperature, air temperature, waves, tides, etc. in real time.

[0179] Step S10: Continuously obtain multiple sets of ocean station operational real-time observation data

[0180] The ocean station is equipped with an automated observation system that collects data every hour. This example selects data from 00:00 on July 1, 2024 to 23:00 on July 31, 2024 for analysis, a total of 744 sets of data. The observation elements include: sea surface temperature ( , unit: ℃), air temperature ( , unit: ℃), effective wave height ( , unit: m), wave period ( , unit: s), wind speed ( , unit: m / s ), wind direction ( , unit: rad), tide level ( , unit: m), air pressure ( , unit: hPa), relative humidity ( , unit: %), seawater salinity ( , unit: psu). After data collection, preliminary processing is first performed, including unit conversion and removal of obvious outliers. For example, wind direction is converted from degrees to radians, and data points with sea surface temperatures exceeding 40°C or below 0°C are removed. The processed data is stored in a PostgreSQL database according to a predefined structure, and each record contains a timestamp and the value of each observation element.

[0181] Step S20: Establish a coupled equation group describing the relationship between different types of observation elements, and use the nonlinear least squares method to fit the historical data. The fitted equation is as follows:

[0182] Temperature coupling equation:

[0183] ;

[0184] Ocean wave coupling equations:

[0185] ;

[0186] Tidal coupling equations:

[0187] ;

[0188] Wind-wave relationship equation:

[0189] ;

[0190] Tide level-increase and decrease of water relationship equation:

[0191] ;

[0192] Temperature-wave relationship equation:

[0193] ;

[0194] Temperature-tidal relationship equation:

[0195] ;

[0196] Step S30: Using the threshold method to mark abnormal data for different types of observation elements

[0197] Based on the statistical characteristics of historical data (data from the same period of 2018 to 2022), the threshold ranges for each observation element are set as follows:

[0198] : 20~35℃;

[0199] : 18~38℃;

[0200] :0.1~5m;

[0201] :0~30m / s;

[0202] :-2~2m;

[0203] :980~1030hPa;

[0204] : 30%~100%;

[0205] For each observed data point, if a feature's value is outside its threshold range, it is marked as an anomaly.

[0206] Step S40: Bring the data with abnormal marking and other data at the corresponding time into the coupled equation group

[0207] Taking a set of abnormal data at 10:00 on July 15, 2024 as an example, the observed values ​​are as follows:

[0208] : 36.5℃ (exceeding the threshold);

[0209] : 32.0℃;

[0210] : 2.1m;

[0211] :12.5m / s;

[0212] :π / 4rad;

[0213] : 0.8m;

[0214] :1008 hPa;

[0215] : 85%;

[0216] : 35psu;

[0217] Substituting this set of data into the coupled equations, we can calculate:

[0218] theory : 30.8℃;

[0219] theory : 1.7m;

[0220] Calculate the deviation:

[0221] Deviation: 2.7℃;

[0222] Deviation: 0.1m;

[0223] Step S50: Calculate the deviation rate of each observation element

[0224] Calculate the deviation rate:

[0225] Deviation rate: 2.7 / 30.8 × 100% = 8.77%;

[0226] Deviation rate: 0.1 / 1.7 × 100% = 5.88%;

[0227] To set the deviation rate threshold: :5%, :10%;

[0228] According to the calculation results, If the deviation rate exceeds the preset threshold, the data of this element at that moment is marked as suspected abnormal data.

[0229] Step S60: Perform cluster analysis on suspected abnormal data

[0230] All data points marked as suspected anomalies in July were collected, totaling 156. These data points were preprocessed, including standardization and missing value filling. Then, the DBSCAN algorithm was used for density clustering analysis, with parameters set to ε = 0.1 and MinPts = 5.

[0231] The clustering results are as follows:

[0232] Cluster 1: 78 data points, main features It is high, mostly occurring during the daytime in mid-July;

[0233] Cluster 2: 45 data points, the main features are On the high side, it often appears around the full moon;

[0234] Cluster 3: 25 data points, the main features are Low, mostly occurs when the wind speed is high;

[0235] Noise points: 8 data points

[0236] The analysis found that cluster 1 may be related to heat wave events, cluster 2 may be related to astronomical tides, and cluster 3 may be related to wave instrument failures.

[0237] Step S70: Combine the threshold method and the coupled equation analysis results to give a comprehensive score to each observation element

[0238] A comprehensive scoring system is established with the following weight distribution:

[0239] Threshold method result: 30%;

[0240] Coupled equation analysis results: 50%;

[0241] Cluster analysis results: 20%;

[0242] Calculate the comprehensive score for each observation element. For example:

[0243] Threshold method score: 0 (outside the threshold range);

[0244] Coupled Equation Analysis Score: ;

[0245] Cluster analysis score: 0.7 (assuming the point belongs to a small cluster);

[0246] Overall Rating ;

[0247] The scoring threshold is set to 0.7. The data was eventually marked as unqualified.

[0248] For this unqualified data, the model estimation method is used to deal with it. Using the temperature coupling equation, the data of other observation elements are estimated The theoretical value (33.2℃) is used to replace the original observed value.

[0249] The processing results were stored in the database, including the original value (36.5°C), the score (0.5903), the quality mark (unqualified), and the processed value (33.2°C).

[0250] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for real-time observation data alarm of an ocean station, characterized in that: The following steps are involved: S10, continuously acquiring multiple sets of operational real-time observation data from ocean stations, including basic data and derived data, both of which include multiple different types of observation elements, the basic data include sea surface temperature, air temperature, one-tenth wave height of waves, tide data, and wind field data, the derived data include wave height variability, wave height accumulation rate, wind-wave relationship, tide accumulation rate, tide change accumulation, water increase and decrease, water increase and decrease accumulation rate, and water increase and decrease change accumulation; S20, based on the basic data and derived data, establish a set of coupling equations describing the relationship between different types of observation elements, including temperature coupling equations, wave coupling equations, tide coupling equations, wind-wave relationship equations, and tide-increase-decrease relationship equations; S30, using the threshold method to mark abnormal data for different types of observation elements; S40, bringing the data marked with abnormalities and other data at corresponding moments into the coupling equation group, and solving to obtain the deviation between the theoretical value and the actual observed value of each observation element; S50, calculating the deviation rate of each observation element, and marking the data whose deviation rate exceeds a preset threshold as suspected abnormal data; S60. Perform cluster analysis on suspected abnormal data to identify abnormal patterns and potential systematic errors; S70, combining the analysis results of the threshold method and the coupling equation, and comprehensively scoring each observation element. Data with a score lower than the set standard is ultimately marked as unqualified data, and an alarm is issued; Wherein, the step S20 specifically includes: Step 201: Calculate derived data based on the collected basic data, and use statistical analysis methods to explore the relationship between different observation elements; Step 202: Based on the theoretical knowledge of physical oceanography and atmospheric science and combined with the data analysis results, a coupled equation group describing the relationship between different observation elements is constructed; Step 203, using nonlinear least squares method, multivariate linear regression and other methods to determine the parameters in the coupled equations; Step 204: Perform model diagnosis, including residual analysis and multicollinearity test, to ensure the validity and reliability of the model; Wherein, the step S40 specifically includes: Step 401: extract all observation element data at the corresponding time from the data points marked as abnormal; Step 402, substituting the extracted data into the established coupled equations; Step 403: Use numerical methods to solve the coupled equations to obtain theoretical values ​​of each observation element; Step 404: Compare the theoretical value with the actual observed value and calculate the deviation of each observed element.

2. A method for alarming real-time observation data of an ocean station according to claim 1, characterized in that: The threshold method is used to mark abnormal data for different types of observation elements, including at least the following four data abnormality markings: (1) Tidal level accumulation rate The cumulative probability statistics of the historical tide levels observed at the ocean station are: ； in For a set of tide data The tide level corresponding to the cumulative probability is, for Cumulative probability: select the tide level corresponding to 1% cumulative probability as the marking threshold, and mark it when the observed tide level exceeds this threshold; (2) Cumulative rate of tidal level change The cumulative probability statistics of the hourly changes of historical tide levels observed at the ocean station are: ； in is the hourly tidal level change value, , and are the tide levels at time t and time t-1 respectively; is a set of hourly tidal level change data The hourly tide level change value corresponding to the cumulative probability, for Cumulative probability: select the hourly tidal level change value corresponding to 1% cumulative probability as the marking threshold, and mark it when the hourly tidal level change value obtained by analysis exceeds the threshold; (3) Tidal level accumulation rate The cumulative probability statistics of the historical increase and decrease of water at the ocean station are: ； in For a set of water increase and decrease data The increase or decrease in water value corresponding to the cumulative probability, for Cumulative probability: select the water increase or decrease value corresponding to 1% cumulative probability as the marking threshold, and mark it when the water increase or decrease value exceeds this threshold; (4) Cumulative rate of water increase and decrease The cumulative probability statistics of the hourly changes in the historical increase and decrease of water at the ocean station are: ； in To increase or decrease the water change value hour by hour, , and are the increase and decrease water values ​​at time t and time t-1 respectively; is a set of hourly water change data The hourly increase or decrease in water value corresponding to the cumulative probability, for Cumulative probability: select the hourly increase or decrease water change value corresponding to 1% cumulative probability as the marking threshold, and mark it when the hourly increase or decrease water change value obtained by analysis exceeds the threshold.

3. The method for real-time observation data alarm of an ocean station according to claim 1, characterized in that: The temperature coupling equation is specifically expressed as: ； In the formula, is the sea surface temperature; is the temperature; is the amplitude of daily temperature variation, obtained by fitting historical data; is the angular frequency; For time; is the wave influence coefficient, obtained through regression analysis; is one tenth of the wave height; It is the average value of one tenth wave height; is the wind speed influence coefficient, obtained through regression analysis; is the wind speed; is the average wind speed; is the air pressure influence coefficient, obtained through regression analysis; is atmospheric pressure; is the average value of atmospheric pressure; is the error term of the temperature coupling equation.

4. The method for real-time observation data alarm of an ocean station according to claim 3, characterized in that: The wave coupling equation is specifically expressed as: ； In the formula, is one tenth of the wave height; is the wind speed; is the length of the wind zone; for water depth; is the acceleration due to gravity; is the sea surface temperature; is the average sea surface temperature; is atmospheric pressure; is the average value of atmospheric pressure; is the relative humidity; is the tidal angular frequency; For time; is the fitting coefficient, obtained by nonlinear regression; is the error term of the ocean-wave coupling equation.

5. The method for real-time observation data alarm of an ocean station according to claim 4, characterized in that: The tidal coupling equation is specifically expressed as: ； In the formula, is the tide level; is the mean sea level; is the amplitude of the i-th partial tide; is the slow-changing factor of the i-th tidal component, which is a function of time; is the angular frequency of the i-th tidal component; is the phase of the i-th partial tide; is the number of tidal subdivisions considered; is atmospheric pressure; is the average value of atmospheric pressure; is the wind speed; is the angle between the wind direction and the coastline; is one tenth of the wave height; is the sea surface temperature; is the average sea surface temperature; is the fitting coefficient, obtained by multiple linear regression; is the error term of the tidal coupling equation.

6. A method for alarming operational real-time observation data of an ocean station according to claim 5, characterized in that: The wind-wave relationship equation is specifically expressed as: ； In the formula, is one tenth of the wave height; is the wind speed; is the length of the wind zone; for water depth; is the wavelength, calculated by the dispersion relation: ; is the wave period; is the fitting coefficient, obtained by nonlinear regression; is the error term of the wind-wave relationship equation.

7. A method for alarming operational real-time observation data of an ocean station according to claim 6, characterized in that: The tide level-increase and decrease water relationship equation is specifically expressed as: ； In the formula, To increase or decrease water height; is atmospheric pressure; is standard atmospheric pressure; is the wind speed; is the angle between the wind direction and the coastline; is the tide level; For time; is one tenth of the wave height; is the sea surface temperature; is the reference temperature; is the salinity of seawater; is the fitting coefficient, obtained by multiple linear regression; It is the error term of the tide level-increase and decrease water relationship equation.

8. The method for real-time observation data alarm of an ocean station according to claim 7, characterized in that: The coupled equations also include a temperature-wave relationship equation and a temperature-tide relationship equation. The temperature-wave relationship equation is specifically expressed as: ； In the formula, is the sea surface temperature; is the temperature; is the dew point temperature; is one tenth of the wave height; is the wind speed; is the angular frequency; For time; is atmospheric pressure; is standard atmospheric pressure; is the fitting coefficient, obtained by nonlinear regression; is the error term of the temperature-wave relationship equation; The temperature-tide relationship equation is specifically expressed as: ； In the formula, is the sea surface temperature; is the reference temperature, obtained by averaging long-term observation data; is the amplitude of daily temperature variation; is the angular frequency of daily temperature variation; For time; is the phase of daily temperature variation; is the tide level; is the wind speed; is atmospheric pressure; is standard atmospheric pressure; is the net heat flux; is the fitting coefficient, obtained by multiple linear regression; is the error term in the temperature-tide relationship equation.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program instructions, which, when executed, are used to execute the method for real-time observation data alarm for ocean station operations as described in any one of claims 1 to 8.

10. A real-time observation data alarm system for ocean stations, characterized in that: A computer-readable storage medium comprising the computer-readable storage medium of claim 9.

Images