A method for monitoring oil quantity data in ship oil tanks based on recursive average filtering algorithm

Abstract

The present invention discloses a method for monitoring the oil quantity data of a ship's oil tank based on a recursive average filtering algorithm, and relates to the technical field of oil quantity monitoring in oil tanks. If the frequency of occurrence of abnormal values ​​in the collected data exceeds expectations, the filter coefficient is adjusted, and the current real-time monitoring value is output based on the recursive average filtering algorithm; the updating frequency of the real-time monitoring value is constrained according to the navigation conditions of the ship, and if the oil quantity of the predicted value of the remaining oil exceeds the oil quantity threshold, an alarm instruction is issued to the outside; the density of receiving the alarm instruction is analyzed and obtained, and if the density exceeds expectations, a calibration instruction is issued to the outside; after periodically calibrating the sensor network, the error degree of the real-time monitoring value generated before calibration is judged, the filter coefficient is optimized again, and the real-time monitoring value is output again based on the recursive average filtering algorithm. When monitoring the oil quantity in the oil tank, the output real-time monitoring value is closer to the actual value, and the error degree is lower.

Description

Technical Field

[0001] The invention relates to the technical field of oil tank oil quantity monitoring, and in particular to a ship oil tank oil quantity data monitoring method based on a recursive average filtering algorithm. Background Art

[0002] Oil tank oil quantity monitoring is an indispensable key technology in the field of modern transportation and storage. It uses advanced sensor technology and data transmission systems to monitor the fuel or lubricating oil in the oil tank in real time 24 hours a day. The collected data is transmitted to the central control unit or remote monitoring platform through wireless or wired networks, realizing remote monitoring and management of oil quantity. On the monitoring interface, users can intuitively see key information such as the current oil quantity, oil level height and oil quantity change trend of the oil tank, helping users to grasp the status of the oil tank at any time. It can also predict the consumption rate of oil through historical data analysis, so as to arrange refueling plans in advance and avoid emergency situations where fuel is exhausted.

[0003] The oil tank oil quantity monitoring system is also equipped with an intelligent alarm function. Once the oil quantity is lower than the preset safety threshold, or there is an abnormal oil quantity fluctuation, the system will immediately trigger an alarm and notify relevant personnel through sound, light signals, text messages, emails, etc. This instant response mechanism ensures that problems can be quickly discovered and handled, effectively preventing risks such as fuel leakage and equipment downtime, and ensuring the safe operation of transportation and storage equipment.

[0004] In the Chinese invention patent with application publication number CN117666420A, a data monitoring system and method for fuel measurement is disclosed, the system includes a fuel measurement control unit, the fuel measurement control unit includes a control module and a data recording module, the control module is connected to the data recording module and the airborne equipment respectively; the input end of the control module is used to receive the oil pump switch signal and the sampling data of the dielectric constant sensor, the oil level sensor, the temperature sensor, the consumption sensor and the instruction data sent by the airborne equipment; the output end of the control module is used to send the external cross-linking data and the intermediate data in the measurement process to the data recording module through the UART port; the data recording module is used to record data and download data through the network port. The present invention stores the external cross-linking data and the intermediate data in the measurement process through the data recording module, realizes the monitoring of the entire operation of the fuel measurement control unit, and has good practicality.

[0005] Combined with the above application and the contents of the prior art:

[0006] When the ship is in a sailing state, it is necessary to pay attention to the fuel data in the cabin and the current fuel consumption data in real time, and adjust the navigation status of the ship according to the collected data. However, considering that the pre-deployed sensor network may often output abnormal values ​​when continuously outputting monitoring data due to its own performance limitations and the influence of external navigation conditions, in order to ensure the reliability and accuracy of the output monitoring values, it is usually necessary to smooth the collected monitoring values.

[0007] However, when the ship's navigation conditions are abnormal, especially when the wind and waves are strong, the ship will be more bumpy and the oil tank liquid level stability will be lower. This will easily lead to a larger deviation between the data obtained by the sensor and the actual data during data collection, a higher degree of error, and lower reliability of the output real-time oil data. Effective alarms cannot be issued for fuel consumption management, which may have a negative impact on the ship's navigation.

[0008] To this end, the present invention provides a method for monitoring ship oil tank oil quantity data based on a recursive average filtering algorithm. Summary of the invention

[0009] 1. Technical issues to be resolved

[0010] In view of the deficiencies of the prior art, the present invention provides a method for monitoring the oil quantity data of a ship's oil tank based on a recursive average filtering algorithm. By constraining the update frequency of the real-time monitoring value according to the ship's navigation conditions, if the predicted value of the remaining oil quantity exceeds the oil quantity threshold, an alarm instruction is issued to the outside; the density of receiving the alarm instruction is analyzed and obtained, and if the density exceeds the expectation, a calibration instruction is issued to the outside; after periodically calibrating the sensor network, the error degree of the real-time monitoring value generated before calibration is judged, the filter coefficient is optimized again, and the real-time monitoring value is output again based on the recursive average filtering algorithm. When monitoring the oil quantity in the oil tank, the output real-time monitoring value is closer to the actual value, and the error degree is lower, thereby solving the technical problems raised in the background.

[0011] (II) Technical solution

[0012] To achieve the above objectives, the present invention is implemented through the following technical solutions: a ship oil tank oil quantity data monitoring method based on a recursive average filtering algorithm, collecting navigation condition data within the ship's navigation area, generating a navigation condition data set after summarizing, generating a navigation condition coefficient Tsx from the navigation condition data set, and issuing a data collection instruction to the outside if the condition coefficient Tsx exceeds expectations;

[0013] A sensor network is arranged in the ship and the oil tank, and the sensor network collects data and generates a set of oil tank oil data. If the frequency of abnormal values ​​in the collected data exceeds expectations, the filter coefficient is adjusted, and the current real-time monitoring value is output based on the recursive average filter algorithm;

[0014] According to the ship's navigation conditions, the update frequency of the real-time monitoring value is constrained, the fuel remaining data is predicted and the corresponding fuel remaining prediction value is obtained. If the fuel remaining prediction value exceeds the fuel threshold [Qa, Qb], an alarm command is issued to the outside;

[0015] Analyze and obtain the density of received alarm instructions. If the density exceeds expectations, perform trend analysis on the fuel consumption under current navigation conditions, and construct the trend coefficient Qts of fuel consumption based on the trend analysis data. If the trend coefficient Qts exceeds the trend threshold, issue a calibration instruction to the outside.

[0016] After periodically calibrating the sensor network, the error degree of the real-time monitoring value generated before calibration is judged. If the error degree does not exceed the expectation, the filter coefficient is optimized again, and the real-time monitoring value is output again based on the recursive average filtering algorithm; wherein, the error value Pw is linearly normalized, and the corresponding data value is mapped to the interval [0, 1], and then according to the following method:

[0017]

[0018] Where n is the number of error values, Pw i is the ith error value, Pw a is the average value of the error, Pw b is the acceptable value of the error value, weight coefficient: 0≤K1≤1, 0≤K2≤1, and K1+K2=1.

[0019] Furthermore, in the ship's navigation area, wave data and wind speed data are collected at each monitoring node, and the collected navigation condition data are aggregated to generate a navigation condition data set;

[0020] The sailing condition coefficient Tsx is generated from the sailing condition data set, where the wind speed Fs and the wave height Bg are linearly normalized and the corresponding data values ​​are mapped to the interval [0, 1] in the following manner:

[0021]

[0022] Weight coefficient: 0≤F1≤1, 0≤F2≤1 and F2+F1=1; Bg i is the wave height of the sea wave at the ith monitoring node, Bg a is the mean wave height, Bg b is the acceptable value of wave height; Fs i is the wind speed of the ith monitoring node, Fs a is the mean wind speed, Fs b is the acceptable value for wind speed.

[0023] Further, after receiving the data collection instruction, a sensor network is arranged in the ship and the oil tank, the sensor data is read regularly, and the collected data is summarized into an oil tank oil data set;

[0024] Taking the real-time data collected by the sensor network as input, outlier monitoring is performed and after obtaining the outliers, the frequency of occurrence of outliers is analyzed and the outlier frequency is generated; if the outlier frequency exceeds expectations, the filter coefficient is adjusted according to the outlier frequency, and the adjustment range of the filter coefficient is positively correlated with the outlier frequency.

[0025] Furthermore, at the data update node that meets the constraint conditions, the real-time monitoring value and the predicted data are output. The constraint mode for the update frequency of the real-time monitoring value according to the navigation conditions of the ship is as follows:

[0026]

[0027] Where n is the number of data updates within the data update cycle, T ij is the time interval from the ith update to the jth update, T a is the average value of the time interval, Tsx is the conditional coefficient; weight coefficient: 0≤F1≤1, 0≤F2≤1, and F1+F2=1.

[0028] Furthermore, a neural convolutional network is trained with sample data to generate a fuel balance prediction model. Historical fuel data and ship navigation conditions are used as input to predict the fuel balance data and obtain the corresponding fuel balance prediction value. A fuel level threshold is set for the fuel in the oil tank. When the fuel level of the predicted fuel balance exceeds the set fuel level threshold [Qa, Qb], an alarm command is issued to the outside.

[0029] Furthermore, the time node of receiving the alarm instruction is recorded and used as the alert node, and the alert density of the alert node is calculated. If the acquired alert density exceeds the preset density threshold, a data analysis instruction is issued to the outside; wherein:

[0030]

[0031] Among them, p is the number of alert nodes, G ij is the time interval from the ith alert node to the jth alert node, G a is the average value of the time interval.

[0032] Furthermore, the collected oil quantity data is recorded in real time and summarized to generate a tank oil data set; after continuously obtaining the oil consumption in several sub-periods, a trend analysis is performed on the oil consumption under the current sailing conditions to generate a trend graph of oil quantity changes and corresponding trend analysis data.

[0033] Furthermore, the trend coefficient Qts of fuel consumption is constructed based on the trend analysis data of fuel consumption, wherein the fuel difference Yc and the fuel ratio Yb are linearly normalized and the corresponding data values ​​are mapped to the interval [0, 1] according to the following formula:

[0034]

[0035] Among them, ρ is the weight coefficient, Yc i is the fuel difference in the ith sub-period, Yc b is the target value of oil difference, Yb i is the oil-to-fuel ratio in the ith sub-cycle, Yb b is the target value of fuel-fuel ratio.

[0036] Further, after receiving the calibration instruction, the calibration cycle is set according to the data collection frequency of the sensor network and the current navigation conditions. After determining the calibration times and calibration nodes within the data calibration cycle, the sensor is calibrated. The setting method is as follows:

[0037]

[0038] Where n is the number of calibrations within the calibration cycle, T ij is the time interval from the i-th calibration to the j-th calibration, T a is the average value of the time interval, and Tsx is the conditional coefficient.

[0039] Furthermore, the real-time monitoring value before calibration is used as a reference value, and the difference between the monitoring data before and after calibration is used as an error value. After a number of error values ​​are obtained continuously, an error degree Td is generated from the error value; if the error degree Td does not exceed the error threshold, a coefficient optimization instruction is issued to the outside.

[0040] After receiving the optimization instruction, the calibrated sensor network continuously obtains several groups of monitoring data and the oil prediction data output by the neural convolutional network; taking reducing the error degree as the optimization goal, the trained coefficient optimization model is used to optimize the filter coefficient to obtain the optimized filter coefficient.

[0041] (III) Beneficial effects

[0042] The present invention provides a method for monitoring the oil quantity data of a ship oil tank based on a recursive average filtering algorithm, which has the following beneficial effects:

[0043] 1. The current ship navigation conditions can be judged based on the condition coefficient Tsx. If the current ship navigation conditions are poor, the fuel monitoring value can be corrected to eliminate the influence of external ship navigation conditions and improve the accuracy of fuel monitoring data. After the monitoring data is identified with the trained outlier recognition model, the outliers are screened out and the outlier frequency is generated. The accuracy of the current monitoring data can be evaluated and judged based on the frequency of occurrence of outliers.

[0044] 2. When the frequency of abnormal values ​​currently output is high, the filter coefficient is corrected and optimized, and the data smoothing effect is increased on the basis of the acquired monitoring values. After the abnormal values ​​are processed, the frequency of occurrence of abnormal values ​​is reduced, thereby improving the reliability and accuracy of the data.

[0045] 3. By constraining and limiting the data update frequency according to the current navigation conditions, the data update frequency is more adapted to the current navigation conditions. By constraining and adjusting the update frequency of the real-time monitoring value, data update omissions are avoided.

[0046] 4. After constructing the fuel threshold, determine whether the fuel remaining amount will be abnormal based on the relationship between the fuel remaining amount prediction value and the fuel remaining amount threshold. If an abnormality occurs, it can be handled in time to ensure the safety of the ship's navigation.

[0047] 5. According to the frequency of alarm instructions, the current abnormal degree of fuel consumption is judged. If the abnormal degree of fuel consumption is high, it is handled in time, and the actual consumption data is analyzed and judged to give priority to the navigation of the ship;

[0048] 6. The trend coefficient Qts is used to judge and evaluate the current fuel consumption trend. If the current fuel consumption is abnormal, the sensor network collects errors in the data as a starting point, and the sensor is calibrated to verify the authenticity of the current fuel consumption, thereby improving the accuracy of data monitoring when monitoring the oil volume in the fuel tank.

[0049] 7. Evaluate the accuracy of the sensor network output data based on the error degree Td. By calibrating the sensor, the collected data will be more accurate when the monitoring value is output in the future.

[0050] 8. When the filter coefficient is optimized using the coefficient optimization model and the real-time monitoring value is output based on the recursive average filtering algorithm, the smoothing effect of abnormal values ​​is better. Therefore, when monitoring the oil volume in the oil tank, the output real-time monitoring value is closer to the actual value, the error degree is lower, and the interference from external navigation conditions is smaller. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1The figure is a flow chart of the method for monitoring the oil quantity data in a ship oil tank of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] See also Figure 1 The present invention provides a method for monitoring the oil quantity data of a ship oil tank based on a recursive average filtering algorithm, comprising: step 1, collecting navigation condition data in a navigation area of ​​the ship, generating a navigation condition data set after summarizing, generating a navigation condition coefficient Tsx from the navigation condition data set, and issuing a data collection instruction to the outside if the condition coefficient Tsx exceeds expectations;

[0054] The step 1 includes the following contents:

[0055] Step 101: When the ship is in a sailing state, an environmental condition monitoring cycle including a plurality of monitoring nodes is set, and wave data and wind speed data are collected at each monitoring node in the ship's sailing area, and the collected sailing condition data are aggregated to generate a sailing condition data set;

[0056] Step 102: Generate the sailing condition coefficient Tsx from the sailing condition data set, wherein the wind speed Fs and the wave height Bg are linearly normalized and the corresponding data values ​​are mapped to the interval [0,1] in the following manner:

[0057]

[0058] Weight coefficient: 0≤F1≤1, 0≤F2≤1 and F2+F1=1; the weight coefficient can be obtained by referring to the hierarchical analysis method, Bg i is the wave height of the sea wave at the ith monitoring node, Bg a is the mean wave height, Bg b is the acceptable value of wave height; Fs i is the wind speed of the ith monitoring node, Fs a is the mean wind speed, Fs b is the acceptable value of wind speed;

[0059] According to historical data and management expectations of the oil status of the oil tank, the condition threshold is set in advance; if the condition coefficient Tsx during navigation exceeds the condition threshold, it means that under the current navigation conditions, the accuracy of the collection and estimation of the oil data in the tank may be affected. At this time, a data collection instruction is issued to the outside;

[0060] When using, combine the contents in steps 101 and 102:

[0061] By monitoring the navigation environment and navigation conditions of the ship, the condition coefficient Tsx is generated by the monitoring data. The current navigation conditions of the ship can be judged based on the condition coefficient Tsx. If the current navigation conditions of the ship are poor, the ship's fuel consumption will be faster. At the same time, under more abnormal navigation conditions, the ship's navigation posture will also be affected and constantly change. Therefore, the monitoring of the ship's fuel will also be affected. It is necessary to correct the monitoring value of the fuel to eliminate the influence of external ship navigation conditions and improve the accuracy of the fuel monitoring data.

[0062] However, when the ship's navigation conditions are abnormal, especially when the wind and waves are strong, the ship will be more bumpy and the oil tank liquid level stability will be lower. This will easily lead to a larger deviation between the data obtained by the sensor and the actual data during data collection, a higher degree of error, and lower reliability of the output real-time oil data. Effective alarms cannot be issued for fuel consumption management, which may have a negative impact on the ship's navigation.

[0063] Step 2: Arrange a sensor network in the ship and the oil tank, collect data through the sensor network and generate a set of oil tank oil data. If the frequency of abnormal values ​​in the collected data exceeds expectations, adjust the filter coefficient and output the current real-time monitoring value based on the recursive average filter algorithm;

[0064] The step 2 includes the following contents:

[0065] Step 201: after receiving the data collection instruction, a sensor network is arranged in the ship and the oil tank, such as a liquid level sensor, a pressure sensor, a temperature sensor, a density sensor and a flow sensor, etc., and the sensor data is read regularly, including the current oil level, temperature and density, etc.; the collected data is summarized into an oil tank oil data set;

[0066] Step 202: Generate an outlier recognition model after training the neural convolution network with sample data, use the real-time data collected by the sensor network as input, perform outlier monitoring and obtain outliers, analyze the frequency of occurrence of outliers and generate an abnormal frequency; if the abnormal frequency exceeds expectations, issue an adjustment instruction for the filter coefficient;

[0067] When in use, after the sensor network completes data collection, the monitoring data is identified using the trained outlier recognition model, outliers are screened out and outlier frequencies are generated. The accuracy of the current monitoring data can be evaluated and judged based on the frequency of occurrence of outliers.

[0068] Step 203: after receiving the adjustment instruction, the filter coefficient is adjusted according to the abnormal frequency, and the adjustment amplitude of the filter coefficient is positively correlated with the abnormal frequency; after continuously acquiring several items of data, the data is smoothed based on the recursive average filtering algorithm, and the current real-time monitoring value is output, so as to display the oil level, oil volume and total oil volume of each oil tank in real time;

[0069] When using, combine the contents in steps 201 to 203:

[0070] When the frequency of abnormal values ​​currently output is high, the filter coefficient is corrected and optimized, thereby increasing the data smoothing effect based on the acquired monitoring values. After the abnormal values ​​are processed, the frequency of occurrence of abnormal values ​​is reduced, thereby improving the reliability and accuracy of the data.

[0071] Step 3: Constrain the update frequency of the real-time monitoring value according to the ship's navigation conditions, predict the fuel remaining data and obtain the corresponding fuel remaining prediction value. If the fuel remaining prediction value exceeds the fuel threshold [Qa, Qb], issue an alarm command to the outside;

[0072] The step three includes the following contents:

[0073] Step 301: The updating frequency of the real-time monitoring value is constrained according to the navigation conditions of the ship as follows:

[0074]

[0075] Where n is the number of data updates within the data update cycle, T ij is the time interval from the ith update to the jth update, T a is the average value of the time interval, Tsx is the conditional coefficient; weight coefficient: 0≤F1≤1, 0≤F2≤1, and F1+F2=1, the weight coefficient is consistent with the previous value;

[0076] On the data update nodes that meet the constraint conditions, the real-time monitoring values ​​and predicted data are output;

[0077] When in use, the data update frequency is constrained and limited according to the current navigation conditions, so that the data update frequency is more adapted to the current navigation conditions. By constraining and adjusting the update frequency of the real-time monitoring value, omissions in data updates can be avoided.

[0078] Step 302: After training the neural convolution network with sample data, a fuel remaining quantity prediction model is generated. The historical fuel data and the ship's navigation conditions are used as input to predict the fuel remaining quantity data and obtain the corresponding fuel remaining quantity prediction value; the fuel quantity threshold [Qa, Qb] is set for the fuel in the fuel tank, in the following manner:

[0079]

[0080] Where i = 1, 2, ..., k, k is the number of remaining oil, Y avg is the mean value of oil residue, Y i is the i-th fuel residue;

[0081] When the predicted value of the remaining fuel quantity exceeds the set fuel quantity threshold [Oa, Qb], an alarm command is issued to the outside and a notification is sent to relevant personnel;

[0082] When used, combine the contents in steps 301 and 302:

[0083] After obtaining some historical data, the fuel remaining data is predicted and the corresponding fuel remaining prediction value is obtained. After obtaining the fuel remaining prediction value, if the fuel remaining prediction value is lower than expected, it can be processed in time; and after constructing the fuel threshold, based on the relationship between the fuel remaining prediction value and the fuel threshold, it is judged whether the fuel remaining will produce abnormalities. If an abnormality occurs, it can be processed in time to ensure the safety of the ship's navigation.

[0084] Step 4: Analyze and obtain the density of receiving alarm instructions. If the density exceeds expectations, perform trend analysis on the fuel consumption under current navigation conditions, and construct the trend coefficient Qts of fuel consumption from the trend analysis data. If the trend coefficient Qts exceeds the trend threshold, issue a calibration instruction to the outside.

[0085] The step 4 includes the following contents:

[0086] Step 401: Record the time node of receiving the alarm instruction and use it as the alert node, and calculate the alert density of the alert node in the following way:

[0087]

[0088] Among them, p is the number of alert nodes, G ij is the time interval from the ith alert node to the jth alert node, G a is the time interval average;

[0089] If the obtained alert density exceeds the preset density threshold, it means that the current fuel consumption is abnormal. At this time, a data analysis instruction is issued to the outside;

[0090] When in use, after receiving several alarm instructions in succession, the current abnormal fuel consumption level is judged according to the density of the alarm instructions. If the abnormal fuel consumption level is high, it needs to be processed in time, and the real consumption data is analyzed and judged to give priority to ensuring the navigation of the ship;

[0091] Step 402: Record the collected oil quantity data in real time, for example, after recording the timestamp, the oil tank number and the oil quantity related data, summarize and generate the oil tank oil data set; after continuously acquiring the oil consumption in several sub-periods, perform trend analysis on the oil consumption under the current sailing conditions, generate the oil quantity change trend chart and the corresponding trend analysis data;

[0092] Step 403: construct the trend coefficient Qts of fuel consumption based on the trend analysis data of fuel consumption, where:

[0093] The difference in fuel consumption between two adjacent sub-periods is taken as the fuel difference Yc, and the corresponding ratio is taken as the fuel ratio Yb. After linear normalization, the corresponding data values ​​are mapped to the interval [0, 1] according to the following formula:

[0094]

[0095] Where ρ is the weight coefficient, and its value falls within [0.1, 0.9], which can be obtained through simulation software analysis or by referring to the hierarchical analysis method; Yc i is the fuel difference in the ith sub-period, Yc b is the target value of oil difference, Yb i is the oil-to-fuel ratio in the ith sub-cycle, Yb b is the target value of fuel-to-fuel ratio;

[0096] Based on historical data and management expectations for fuel consumption, a trend threshold is pre-set; if the trend coefficient Qts exceeds the trend threshold, it means that the current fuel consumption rate is too fast and fuel needs to be saved under the current navigation conditions, or it can be said that when collecting fuel data, the accuracy of the fuel data obtained is insufficient. At this time, a calibration instruction is issued to the outside;

[0097] When used, combine the contents in steps 401 to 403:

[0098] After continuously collecting the fuel consumption data during navigation, the fuel consumption data is trend analyzed, and the trend coefficient Qts is calculated based on the trend analysis data. The trend coefficient Qts is used to judge and evaluate the current fuel consumption trend. If the current fuel consumption is abnormal, the error in the data collected by the sensor network is taken as the starting point. The sensor is calibrated to verify the authenticity of the current fuel consumption, thereby improving the accuracy of monitoring when monitoring the oil amount in the fuel tank.

[0099] Step 5: After periodically calibrating the sensor network, the error degree of the real-time monitoring value generated before calibration is judged. If the error degree does not exceed the expectation, the filter coefficient is optimized again, and the real-time monitoring value is output again based on the recursive average filtering algorithm;

[0100] The step five includes the following contents:

[0101] Step 501: After receiving the calibration instruction, the calibration period is set according to the data collection frequency of the sensor network and the current navigation conditions. The setting method is as follows:

[0102]

[0103] Where n is the number of calibrations within the calibration cycle, T ij is the time interval from the i-th calibration to the j-th calibration, T a is the average value of the time interval, Tsx is the conditional coefficient;

[0104] After determining the calibration times and calibration nodes within the data calibration cycle, calibrate the sensor, for example, using a standard level measurement tool, to ensure measurement accuracy, thereby improving the accuracy of tank oil monitoring;

[0105] Step 502: After completing the data calibration of the sensor network, the real-time monitoring value before calibration is used as the reference value, and the difference between the monitoring data before and after calibration is used as the error value. After continuously obtaining several error values, the error degree Td is generated from the error value:

[0106] Among them, the error value Pw is linearly normalized, and the corresponding data value is mapped to the interval [0, 1], and then according to the following method:

[0107]

[0108] Where n is the number of error values, Pw i is the ith error value, Pw a is the average value of the error, Pw b is the acceptable value of the error value, the weight coefficient is: 0≤K1≤1, 0≤K2≤1, and K1+K2=1; the weight coefficient can be obtained by referring to the hierarchical analysis method;

[0109] When the error degree obtained exceeds the preset error threshold, it means that the output real-time monitoring value may be wrong and the accuracy is insufficient. On the contrary, if the error degree Td does not exceed the error threshold, it means that the accuracy of the output real-time monitoring value may be insufficient. At this time, a coefficient optimization instruction is sent to the outside. By optimizing the filter coefficient, the accuracy of the output real-time monitoring value can be improved.

[0110] When in use, after completing the calibration of the sensor network, compare the monitoring values ​​before and after calibration, construct the error degree Td based on the difference between the two, and evaluate the accuracy of the sensor network output data based on the error degree Td. At the same time, by calibrating the sensor, when continuing to output monitoring values ​​in the future, the collected data will be more accurate.

[0111] Step 503: after receiving the optimization instruction, the coefficient optimization model is obtained by training the genetic algorithm with the sample data, and several groups of monitoring data are continuously obtained by the calibrated sensor network, as well as the oil prediction data output by the neural convolution network;

[0112] Taking the reduction of error as the optimization goal, the filter coefficient is optimized using the trained coefficient optimization model to obtain the optimized filter coefficient; after continuously obtaining several groups of monitoring data, the real-time monitoring value is output based on the recursive average filtering algorithm in combination with the optimized filter coefficient;

[0113] When using, combine the contents in steps 501 to 503:

[0114] On the basis of sensor calibration, the filter coefficient is optimized using the coefficient optimization model, and then the real-time monitoring value is output based on the recursive average filtering algorithm. The smoothing effect of abnormal values ​​is better. Therefore, when monitoring the oil amount in the oil tank, the output real-time monitoring value is closer to the actual value, the error degree is lower, and the interference from external navigation conditions is smaller.

[0115] The Analytic Hierarchy Process (AHP) is a decision-making method that decomposes decision-related elements into levels such as goals, criteria, and plans, and conducts qualitative and quantitative analysis on this basis. It is particularly suitable for dealing with target systems with hierarchical and staggered evaluation indicators, and when the target value is difficult to describe quantitatively, the AHP is an effective decision-making tool.

[0116] The core of the hierarchical analysis method is to decompose the decision problem into multiple levels to form a hierarchical structure, which usually includes the target level, the criterion level, the sub-criterion level and the solution level. By solving the eigenvector of the judgment matrix, the priority weight of each element of each level to an element of the previous level is obtained, and the final weight of each alternative solution to the total goal is recursively merged by the weighted sum method, so as to find the optimal solution.

[0117] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0118] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0119] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0120] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0121] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0122] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0123] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0124] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for monitoring the oil quantity data of a ship oil tank based on a recursive average filtering algorithm, characterized in that: include, Collect the navigation condition data in the ship's navigation area, summarize them to generate a navigation condition data set, and generate the navigation condition coefficient Tsx from the navigation condition data set. If the condition coefficient Tsx exceeds the expectation, send a data collection instruction to the outside; A sensor network is arranged in the ship and the oil tank, and the sensor network collects data and generates a set of oil tank oil data. If the frequency of abnormal values ​​in the collected data exceeds expectations, the filter coefficient is adjusted, and the current real-time monitoring value is output based on the recursive average filter algorithm; According to the ship's navigation conditions, the update frequency of the real-time monitoring value is constrained, the fuel remaining data is predicted and the corresponding fuel remaining prediction value is obtained. If the fuel remaining prediction value exceeds the fuel threshold [Qa, Qb], an alarm command is issued to the outside; Analyze and obtain the density of received alarm instructions. If the density exceeds expectations, perform trend analysis on the fuel consumption under current navigation conditions, and construct the trend coefficient Qts of fuel consumption based on the trend analysis data. If the trend coefficient Qts exceeds the trend threshold, issue a calibration instruction to the outside. After periodically calibrating the sensor network, the error degree of the real-time monitoring value generated before calibration is judged. If the error degree does not exceed the expectation, the filter coefficient is optimized again, and the real-time monitoring value is output again based on the recursive average filtering algorithm; wherein, the error value Pw is linearly normalized, the corresponding data value is mapped to the interval [0,1], and then the error degree Td is generated from the error value in the following way: Where n is the number of error values, Pw i is the ith error value, Pw a is the average value of the error, Pw b is the acceptable value of the error value, weight coefficient: 0≤K1≤1, 0≤K2≤1, and K1+K2=1.

2. The method for monitoring the oil quantity data of a ship oil tank according to claim 1, characterized in that: In the ship's navigation area, wave data and wind speed data are collected at each monitoring node, and the collected navigation condition data are summarized to generate a navigation condition data set; The sailing condition coefficient Tsx is generated from the sailing condition data set, where the wind speed Fs and the wave height Bg are linearly normalized and the corresponding data values ​​are mapped to the interval [0,1] in the following manner: Weight coefficient: 0≤F1≤1, 0≤F2≤1 and F2+F1=1; Bg i is the wave height of the sea wave at the ith monitoring node, Bg a is the mean wave height, Bg b is the acceptable value of wave height; Fs i is the wind speed of the ith monitoring node, Fs a is the mean wind speed, Fs b is the acceptable value for wind speed.

3. The method for monitoring the oil quantity data of a ship oil tank according to claim 2, characterized in that: After receiving the data collection instruction, a sensor network is arranged in the ship and the oil tank, the sensor data is read regularly, and the collected data is summarized into the oil tank oil data set; Taking the real-time data collected by the sensor network as input, outlier monitoring is performed and after obtaining the outliers, the frequency of occurrence of outliers is analyzed and the outlier frequency is generated; if the outlier frequency exceeds expectations, the filter coefficient is adjusted according to the outlier frequency, and the adjustment range of the filter coefficient is positively correlated with the outlier frequency.

4. The method for monitoring the oil quantity data of a ship oil tank according to claim 3 is characterized in that: At the data update node that meets the constraint conditions, the real-time monitoring value and the predicted data are output. The constraint method for the update frequency of the real-time monitoring value according to the navigation conditions of the ship is as follows: Where n is the number of data updates within the data update cycle, T ij is the time interval from the i-th update to the j-th update; T a is the average value of the time interval, Tsx is the conditional coefficient; weight coefficient: 0≤F1≤1, 0≤F3≤1, and F1+F2=1.

5. The method for monitoring the oil quantity data of a ship oil tank according to claim 4 is characterized in that: After training the neural convolutional network with sample data, a fuel balance prediction model is generated. Taking historical fuel data and ship navigation conditions as input, the fuel balance data is predicted and the corresponding fuel balance prediction value is obtained, and the fuel level threshold is set for the fuel in the tank.

6. The method for monitoring the oil quantity data of a ship oil tank according to claim 5, characterized in that: After recording the time node of receiving the alarm instruction, it is used as the alert node, and the alert density Moj of the alert node is calculated. If the obtained alert density Moj exceeds the preset density threshold, a data analysis instruction is issued to the outside. The method of generating the alert density Moj is as follows: Among them, p is the number of alert nodes, G ij is the time interval from the ith alert node to the jth alert node, G a is the average value of the time interval.

7. The method for monitoring the oil quantity data of a ship oil tank according to claim 6, characterized in that: The collected oil quantity data is recorded in real time and summarized to generate a set of oil tank oil data; after continuously obtaining the oil consumption in several sub-periods, a trend analysis is performed on the oil consumption under the current sailing conditions to generate a trend chart of oil quantity changes and corresponding trend analysis data.

8. The method for monitoring the oil quantity data of a ship oil tank according to claim 7, characterized in that: The trend coefficient Qts of fuel consumption is constructed based on the trend analysis data of fuel consumption. After linear normalization of the fuel difference Yc and the fuel ratio Yb, the corresponding data values ​​are mapped to the interval [0,1] according to the following formula: Among them, ρ is the weight coefficient, Yc i is the fuel difference in the ith sub-period, Yc b is the target value of oil difference, Yb i is the oil-to-fuel ratio in the ith sub-cycle, Yb b is the target value of fuel-fuel ratio.

9. The method for monitoring the oil quantity data of a ship oil tank according to claim 8, characterized in that: After receiving the calibration instruction, the calibration cycle is set according to the data collection frequency of the sensor network and the current navigation conditions. After determining the calibration times and calibration nodes within the data calibration cycle, the sensor is calibrated. The calibration cycle is set as follows: Where n is the number of calibrations within the calibration cycle, T ij is the time interval from the i-th calibration to the j-th calibration, T a is the average value of the time interval, and Tsx is the conditional coefficient.

10. The method for monitoring the oil quantity data of a ship oil tank according to claim 9, characterized in that: The real-time monitoring value before calibration is used as the reference value, and the difference between the monitoring data before and after calibration is used as the error value. After obtaining several error values ​​continuously, the error degree Td is generated from the error value. If the error degree Td does not exceed the error threshold, several groups of monitoring data are continuously obtained by the calibrated sensor network, as well as the oil prediction data output by the neural convolutional network. The error degree reduction is taken as the optimization goal, and the filter coefficient is optimized using the trained coefficient optimization model to obtain the optimized filter coefficient.

Images