Ship fault prediction method and system based on artificial intelligence

Abstract

The invention discloses a ship fault prediction method and system based on artificial intelligence, and relates to the technical field of fault prediction, and the method comprises the steps: recording the driving path of a ship based on a ship system, and obtaining the navigation data; working data of a steering engine of a ship are recorded based on a ship system, and rudder direction data are obtained; the navigation data is compared with the rudder direction data, if the actual driving course of the ship is different from the guiding course of the steering engine for the ship, it is judged that the rudder of the ship is in an abnormal working state, a steering engine system reset signal is output, and a steering engine system reset confirmation signal of a user is received; the steering engine system is reset based on the steering engine system reset confirmation signal, the ship running course after resetting is recorded, course data obtained through recording are judged, and course deviation factors are determined; and determining a fault source based on the course deviation factor, and outputting a corresponding alarm signal according to the fault source. The method has the effect of improving the fault judgment accuracy.

Description

Technical Field

[0001] The present application relates to the technical field of fault prediction, and in particular to a method and system for predicting ship faults based on artificial intelligence. Background Art

[0002] With the continuous development of global trade, ships play a vital role in the field of transportation. However, during the long-term operation of ships, various faults are inevitable, such as engine failure, mechanical parts wear, electrical system failure, etc. Among them, rudder failure will not only affect the normal operation of the ship and cause transportation delays, but may also cause serious safety accidents, causing huge losses to the life safety of the crew and property.

[0003] Among the related technologies, machine learning algorithms are often used to train ship operation data and establish fault prediction models. Various ship parameters, such as temperature, pressure, vibration, etc., are collected through sensors, and these parameters are analyzed using deep learning technology to predict the occurrence of faults. Fuzzy logic and other methods are used to process the uncertainty of the ship system to improve the accuracy of fault prediction.

[0004] Regarding the above-mentioned related technologies, since different ships have different equipment information configurations and are subject to different environmental influences during driving, the trained fault prediction model has a low degree of adaptability to the ship itself, which in turn causes the accuracy of the system's fault judgment to change with the adaptation situation, thereby reducing the accuracy of fault judgment. There is room for improvement. Summary of the invention

[0005] In order to improve the accuracy of fault judgment, the present application provides a ship fault prediction method and system based on artificial intelligence.

[0006] In the first aspect, the present application provides a ship fault prediction method based on artificial intelligence, which adopts the following technical solutions:

[0007] A ship fault prediction method based on artificial intelligence, comprising:

[0008] The ship's travel path is recorded based on the ship system to obtain navigation data;

[0009] Based on the ship system, the working data of the ship's steering gear is recorded to obtain the rudder direction data;

[0010] Compare the navigation data with the steering data to determine whether the actual heading of the ship is the same as the heading guided by the steering gear;

[0011] If the actual heading of the ship is the same as the heading guided by the steering gear, it is determined that the rudder of the ship is in normal working condition;

[0012] If the actual heading of the ship is different from the heading guided by the steering gear, it is determined that the rudder of the ship is in an abnormal working state, and the abnormal working state is analyzed;

[0013] When it is determined that the ship's rudder is in an abnormal working state, a steering gear system reset signal is output, and a steering gear system reset confirmation signal from the user is received;

[0014] Reset the steering gear system based on the steering gear system reset confirmation signal, record the ship's heading after the reset, judge the recorded heading data, and determine the heading deviation factor;

[0015] Based on the heading deviation factor, the fault source is determined, and the corresponding alarm signal is output according to the fault source.

[0016] Preferably, based on the built-in ship simulation model, the working data of the steering servo, the working time of the corresponding working data and the speed data are input into the ship simulation model, and the driving course simulation is performed to obtain a simulated route map;

[0017] Draw the navigation data into a driving route map based on the time data;

[0018] Based on the time data, the driving route map under the same time axis is compared with the simulation route map;

[0019] If the similarity between the driving route map and the simulated route map is greater than a preset similarity threshold, it is determined that the actual driving direction of the ship is the same as the guiding direction of the ship by the steering gear;

[0020] If the similarity between the driving route map and the simulated route map is less than a preset similarity threshold, it is determined that the actual driving heading of the ship is different from the guiding heading of the ship by the steering gear.

[0021] Preferably, the initial travel direction of the ship after reset is used as the basic reference direction, and the travel direction of the ship is recorded based on the built-in time interval to obtain the direction data;

[0022] Compare the recorded heading data with the basic reference direction to determine whether the heading of the ship after reset has changed within the built-in time interval;

[0023] If the ship's heading does not change within the built-in time interval, the abnormal working state of the ship's rudder is determined to be a system failure;

[0024] If the ship's heading changes within the built-in time interval, the two adjacent heading data are compared based on the collection time of the heading data to determine the position offset relationship between the heading data collected later and the heading data collected earlier, and the position offset relationship is statistically analyzed to obtain the offset relationship data;

[0025] Judge the offset relationship data and determine the offset factor.

[0026] Preferably, the heading data is processed based on built-in rule information to obtain the offset attribute of the offset relationship data; the rule information is that, according to the collected time data, the heading data collected earlier is used as the reference heading of the heading data collected later, the heading data that changes counterclockwise relative to the reference heading is marked as a positive value, and the heading data that changes clockwise relative to the reference heading is marked as a negative value;

[0027] Based on the offset attributes in the offset relationship data, the offset relationship data are compared for unity. If the offset attributes corresponding to the respective offset relationship data are all positive values ​​or all negative values, it is determined that the offset relationship data are unified.

[0028] If the offset attributes corresponding to each offset relationship data are not all positive or negative, it is determined that the offset relationship data is not unique;

[0029] If the offset relationship data is unique, the offset in the offset relationship data is judged to determine a first offset factor; the first offset factor includes external force and rudder structure failure;

[0030] If the offset relationship data is not unique, a change rule of the offset attribute of the offset relationship data is determined to determine a second offset factor; the second offset factor includes a rudder structure fault and a rudder connection fault;

[0031] The first offset factor and the second offset factor are collectively labeled as offset factors.

[0032] Preferably, mathematical statistics are performed on the offsets of the offset relationship data to determine whether there is clustering between the offset relationship data;

[0033] If it is determined that there is a clustering between the offset relationship data, then it is determined that the first offset factor is a rudder structure failure;

[0034] If it is determined that there is no clustering between the offset relationship data, then the offset of the offset relationship data is judged according to the acquisition time sequence to determine whether the offset of the offset relationship data gradually decreases with time;

[0035] If it is determined that the offset of the offset relationship data gradually decreases with time, a base value judgment of the value change of the offset of the offset relationship data is performed according to the acquisition time sequence to determine whether the offset of the offset relationship data has a base value;

[0036] If it is determined that the base value of the offset relationship data is greater than the built-in fluctuation base value, then the first offset factor is determined to be a rudder structure failure;

[0037] If it is determined that the base value of the offset relationship data is less than the built-in fluctuation base value, it is determined that the first offset factor is an external force.

[0038] Preferably, based on the offset attribute of the offset relationship data, the offset amounts of the offset relationship data with the same offset attribute are counted to obtain two single attribute offset data;

[0039] Performing mathematical statistical judgment on the two single attribute offset data respectively to determine whether at least one of the two single attribute offset data has clustering properties;

[0040] If at least one of the two single attribute offset data has a clustering property, it is determined that the second offset factor is a rudder structure failure;

[0041] If the two single attribute offset data are not clustered, it is determined that the second offset factor is a rudder connection failure.

[0042] In the second aspect, the present application provides a ship fault prediction system based on artificial intelligence, which adopts the following technical solutions:

[0043] A ship fault prediction system based on artificial intelligence, comprising: a data acquisition module, a first fault judgment module, a second fault judgment module, and a fault alarm module;

[0044] The data acquisition module records the ship's travel path based on the ship system to obtain navigation data; it records the working data of the ship's steering gear based on the ship system to obtain steering direction data;

[0045] The first fault judgment module compares the navigation data with the rudder data to determine whether the actual driving direction of the ship is the same as the guiding direction of the ship by the steering rudder; if the actual driving direction of the ship is the same as the guiding direction of the ship by the steering rudder, it is determined that the rudder of the ship is in a normal working state; if the actual driving direction of the ship is different from the guiding direction of the ship by the steering rudder, it is determined that the rudder of the ship is in an abnormal working state, and the abnormal working state is analyzed;

[0046] The second fault judgment module, when determining that the rudder of the ship is in an abnormal working state, outputs a steering gear system reset signal, and receives a steering gear system reset confirmation signal from the user; resets the steering gear system based on the steering gear system reset confirmation signal, and records the ship's heading after the reset, judges the recorded heading data, and determines the heading deviation factor;

[0047] The fault alarm module determines the fault source based on the heading deviation factor and outputs a corresponding alarm signal according to the fault source.

[0048] In summary, the present application includes at least one of the following beneficial technical effects:

[0049] 1. By comparing the ship's driving path with the working data of the steering servo, it is determined whether the ship changes its course according to the working data of the steering servo in the current working state, and then the working state of the ship is judged simply and quickly, which improves the recognition rate of rudder failure. Then, by resetting the steering system and analyzing and judging the reset heading data, the system interference of the steering system is reduced, and the accuracy of identifying the cause of the rudder failure is improved;

[0050] 2. By judging the heading after reset again, it is clear whether the cause of the fault has been eliminated. When the heading data is the same as the initial heading of the ship, it indicates that the working state of the ship's rudder is normal, that is, the fault has been eliminated, so no alarm reminder is issued. When the heading data is different from the initial heading of the ship, it indicates that the cause of the ship's fault has not been eliminated, so further judgment is needed. By comparing the heading data of two adjacent collection times, the heading change of the ship within the built-in time interval can be clarified. By analyzing the heading change, the cause of the heading change can be clarified, thereby improving the accuracy of ship fault judgment;

[0051] 3. Comprehensively utilize the offset attributes and offset amounts of the offset relationship data to perform singleness judgment, clustering judgment, base value judgment and statistical judgment of the same offset attribute on the offset relationship data. By performing singleness judgment on the offset relationship data, further offset factor judgment methods can be determined. By performing clustering judgment and base value judgment on the offset relationship data with singleness, it can be further clarified that the cause of the ship's heading deviation is a rudder structure failure or external force. By performing statistics on the same offset attribute of the offset relationship data that does not have singleness, the clustering of the offset amounts corresponding to the same offset attribute can be determined, and then it can be determined that the cause of the ship's heading deviation is a rudder structure failure or a rudder connection failure, making the ship's fault judgment more accurate and effective. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1This is a flowchart of the steps of the ship fault prediction method based on artificial intelligence in this embodiment;

[0053] Figure 2 This is a module block diagram of the ship fault prediction system based on artificial intelligence in this embodiment.

[0054] Figure numerals: 1. Data acquisition module; 2. First fault judgment module; 3. Second fault judgment module; 4. Fault alarm module. DETAILED DESCRIPTION

[0055] The following is combined with Figure 1-Figure 2 This application is described in further detail.

[0056] The embodiments of the present application disclose a ship fault prediction method and system based on artificial intelligence.

[0057] Example: Figure 1 As shown, the present invention provides a ship fault prediction method based on artificial intelligence, comprising:

[0058] S100, recording the travel path of the ship based on the ship system to obtain navigation data;

[0059] S200, recording working data of a steering servo of the ship based on a ship system to obtain rudder direction data;

[0060] S300, comparing the navigation data with the steering data to determine whether the actual sailing direction of the ship is the same as the guiding direction of the ship by the steering rudder;

[0061] S400, if the actual traveling heading of the ship is the same as the guiding heading of the ship by the steering gear, it is determined that the rudder of the ship is in a normal working state;

[0062] S500, if the actual driving direction of the ship is different from the guiding direction of the steering servo for the ship, it is determined that the rudder of the ship is in an abnormal working state, and the abnormal working state is analyzed; in the previous part, a simple judgment is made to quickly determine whether the rudder is abnormal, and the abnormality is affected by external factors and other conditions. For example, wind blowing will accelerate the turning of the ship, and the subsequent reset, on the one hand, is to simply and quickly determine whether it is a steering gear system failure through restart, so as to quickly repair it. On the other hand, since the rudder in the initial state corresponding to the restarted steering gear system is in a non-deflected state, in this state, the ship's driving path is a straight line, so by judging whether it is a straight line, it is further determined whether the rudder is normal.

[0063] S600, when it is determined that the rudder of the ship is in an abnormal working state, outputting a steering gear system reset signal, and receiving a steering gear system reset confirmation signal from the user;

[0064] S700, resetting the steering gear system based on the steering gear system reset confirmation signal, recording the ship's heading after the reset, judging the recorded heading data, and determining the heading deviation factor;

[0065] S800, based on the heading deviation factor, determine the fault source, and output a corresponding alarm signal according to the fault source.

[0066] In this embodiment, by comparing the ship's driving path with the working data of the steering servo, it is determined whether the ship changes its heading according to the working data of the steering servo in the current working state, and then the working state of the ship is judged simply and quickly. When it is determined that the working state of the ship is abnormal, the steering servo failure caused by the system failure is eliminated by automatically resetting the steering gear system. The ship's driving heading after the reset is recorded, analyzed and judged to identify the source of the fault, and output a corresponding alarm signal according to the source of the fault to remind the crew to perform corresponding maintenance.

[0067] By utilizing the ship's travel path and the working data of the steering servo, the working status of the rudder can be judged simply and quickly, thereby improving the recognition rate of rudder failures. Then, by resetting the steering gear system and analyzing and judging the reset heading data, the system interference of the steering gear system is reduced, thereby improving the accuracy of identifying the cause of the rudder failure.

[0068] For example, when a ship is turning, water flow hits the rudder at different angles, causing the ship to be subjected to a certain steering torque, thereby enabling the ship to turn normally. For example, the working data of the ship is that the ship deviates 15° to the right within a unit mileage, but in fact the ship deviates 15° to the left within a unit mileage. It is determined that the actual driving path of the ship is different from the action path of the rudder, and then it is determined that the rudder of the ship has failed. The fault may be a software fault, that is, a steering gear system fault, or it may be a rudder hardware fault. Therefore, the steering gear system is reset to eliminate the possibility of a steering gear system fault.

[0069] If the steering gear system is abnormal, the abnormality of the steering gear system is eliminated after the steering gear system is reset, and the rudder is in the initial state in the current state, that is, the ship's driving path is a straight line. When the actual driving path is not a straight line, it is determined to be a rudder hardware failure. Therefore, the specific cause of the rudder failure is determined by analyzing and judging the ship's driving data after the reset.

[0070] In step S300, the navigation data is compared with the steering data to determine whether the actual sailing direction of the ship is the same as the guiding direction of the ship by the steering servo, including the following steps:

[0071] S310, based on the built-in ship simulation model, inputting the working data of the steering servo, the working time of the corresponding working data and the speed data into the ship simulation model, and performing a driving course simulation to obtain a simulation route map;

[0072] S320, drawing the navigation data into a driving route map based on the time data;

[0073] S330, comparing the driving route map under the same time axis with the simulation route map based on the time data;

[0074] S340, if the similarity between the driving route map and the simulated route map is greater than a preset similarity threshold, it is determined that the actual driving heading of the ship is the same as the guiding heading of the ship by the steering gear;

[0075] S350: If the similarity between the driving route map and the simulation route map is less than a preset similarity threshold, it is determined that the actual driving heading of the ship is different from the guiding heading of the ship by the steering servo.

[0076] In this embodiment, since the navigation data and the rudder data of the ship are not of the same data type, the rudder data needs to be converted. By utilizing the built-in ship simulation model and relevant navigation information on the ship system, the navigation route map under the rudder data is simulated to obtain a simulated route map. Similarly, the actual navigation data is path-drawn to obtain a driving route map corresponding to the actual ship's travel. Then, by performing graphic matching on the simulated route map and the driving route map under the same time scale, it is clear whether the simulated route of the ship is the same as the actual driving route, thereby making the judgment of the working status of the rudder more accurate.

[0077] In step S700, the steering gear system is reset based on the steering gear system reset confirmation signal, and the ship's heading after the reset is recorded, and the recorded heading data is judged to determine the heading deviation factor, including the following steps:

[0078] S710, taking the initial travel direction of the ship after reset as a basic reference direction, based on the built-in time interval, recording the travel direction of the ship to obtain the direction data;

[0079] S720, comparing the recorded heading data with the basic reference direction to determine whether the heading of the ship after reset has changed within the built-in time interval;

[0080] S730, if the ship's heading does not change within the built-in time interval, the abnormal working state of the ship's rudder is determined to be a system failure; since the current heading has not changed, it indicates that the state of the current ship's rudder is normal, so no alarm reminder is issued.

[0081] S740, if the ship's heading changes within the built-in time interval, then based on the collection time of the heading data, two adjacent heading data are compared to determine the position offset relationship between the heading data collected later and the heading data collected earlier, and the position offset relationship is statistically analyzed to obtain offset relationship data; wherein the offset relationship data is the offset data value and offset attribute; wherein the offset attribute, for example: offset to the left or offset to the right.

[0082] S750, judging the offset relationship data and determining the offset factor. The offset cause may be an external factor or equipment failure. For example, when a ship is traveling in a straight line, the wind exerts a force on the ship from the side, causing the ship to turn under the action of the wind, thereby changing the ship's heading. Therefore, judging the offset relationship data further clarifies the yaw factor, thereby ensuring the accuracy of the ship fault judgment.

[0083] In this embodiment, the heading after reset is judged again to determine whether the cause of the fault has been eliminated. When the heading data is the same as the initial heading of the ship, it indicates that the rudder of the ship is in a normal working state, that is, the fault has been eliminated, so no alarm is given. When the heading data is different from the initial heading of the ship, it indicates that the cause of the fault of the ship has not been eliminated, so further judgment is needed. By comparing the heading data of two adjacent collection times, the heading change of the ship within the built-in time interval is determined. By analyzing the heading change, the cause of the heading change is determined, thereby improving the accuracy of ship fault judgment.

[0084] In step S750, the offset relationship data is judged to determine the offset factor, including the following steps:

[0085] S751, based on the built-in rule information, the heading data is processed to obtain the offset attribute of the offset relationship data; the rule information is that according to the collected time data, the heading data collected earlier is used as the reference heading of the heading data collected later, the heading data that changes counterclockwise relative to the reference heading is marked as a positive value, and the heading data that changes clockwise relative to the reference heading is marked as a negative value;

[0086] S752, based on the offset attributes in the offset relationship data, performing a uniqueness comparison on the offset relationship data, if the offset attributes corresponding to each offset relationship data are all positive values ​​or all negative values, then it is determined that the offset relationship data is unique;

[0087] S753, if the offset attributes corresponding to the offset relationship data are not all positive or negative, it is determined that the offset relationship data is not unique;

[0088] S754, if the offset relationship data is unique, the offset in the offset relationship data is judged to determine a first offset factor; the first offset factor includes external force and rudder structure failure; wherein the rudder structure failure may be deformation of the rudder.

[0089] S755, if the offset relationship data is not unique, the change rule of the offset attribute of the offset relationship data is judged to determine the second offset factor; the second offset factor includes a rudder structure failure and a rudder connection failure; the rudder connection failure may be that the steering gear cannot fix the position of the rudder.

[0090] S756: Mark the first offset factor and the second offset factor as an offset factor.

[0091] In this embodiment, the offset attributes of the offset relationship data are confirmed, and then the uniqueness of the offset attributes in each offset relationship data is judged, and then the first screening and judgment of the offset cause of the ship's deviation is made according to the uniqueness of the offset attributes. Then, by utilizing the uniqueness of the offset attributes, corresponding judgment methods are adopted for different offset situations, thereby making the judgment result of the offset cause more accurate and improving the accuracy of fault judgment.

[0092] Exemplarily, assuming that the heading of the initial position is due north, the heading obtained in the next collection is 1° north-east, the heading obtained in the second collection is 2° north-east, the heading obtained in the third collection is 3° north-east, and so on. Then all headings collected later are counterclockwise compared to headings collected earlier, so it is determined that the offset relationship data are all negative values. Conversely, if the heading obtained in the first collection is 1° north-west, the heading obtained in the second collection is 2° north-west, the heading obtained in the third collection is 3° north-west, and so on, then it is determined that the offset relationship data are all positive values.

[0093] When both are positive or negative, it indicates that the deviation of the ship is unidirectional. Therefore, when the rudder is in the initial state, the reasons for the ship's unidirectional deflection are: 1. The actual position of the rudder is not in the initial state or the rudder is deformed, resulting in uneven water pressure on both sides of the rudder; 2. The actual position of the rudder is in the initial state, but is affected by external forces, causing the ship to deviate from its course.

[0094] If the heading obtained in the first collection is 1° north-west, the heading obtained in the second collection is due north, the heading obtained in the third collection is 3° north-west, and the heading obtained in the fourth collection is 2° north-west, although they are still in the north-west azimuth, the azimuth angle of the north-west changes back and forth, so there are both positive and negative values ​​in the offset relationship data, which shows that the deviation of the ship does not have a single direction. Therefore, when the rudder is in the initial state, the reasons for the back-and-forth changes in the ship's heading are: 1. The actual position of the rudder is not in the initial state or the rudder is deformed and is acted upon by opposite external forces, causing the ship's heading to change with the change of external forces; 2. The rudder itself is not stable, causing it to swing back and forth in the water flow, which in turn causes the ship's heading to be unstable.

[0095] In step S754, if the offset relationship data is unique, the offset in the offset relationship data is judged to determine a first offset factor; the first offset factor includes external force and rudder structure failure, including the following steps:

[0096] S7541, perform mathematical and statistical judgment on the offset of the offset relationship data to determine whether there is clustering between the offset relationship data; when the offset of most of the offset relationship data is 1°, it is determined that the offset relationship data has clustering and clusters toward 1°, otherwise, it is determined that the offset relationship data does not have clustering.

[0097] S7542, if it is determined that there is convergence between the various offset relationship data, then the first offset factor is determined to be a rudder structure failure; when there is convergence between the various offset relationship data, for example, they all converge to 1°, it means that each time the data is collected, the heading of the ship is offset by 1° based on the previous collection, which further indicates that the magnitude and direction of the steering torque acting on the ship remain unchanged, so the offset factor is determined to be a rudder structure failure, resulting in the rudder still being subjected to a fixed steering torque when it is in the initial state.

[0098] S7543, if it is determined that there is no clustering between the offset relationship data, then according to the acquisition time sequence, the offset amount of the offset relationship data is judged for numerical change, and whether the offset amount of the offset relationship data gradually decreases with time;

[0099] S7544, if it is determined that the offset of the offset relationship data gradually decreases with time, the offset of the offset relationship data is judged based on the acquisition time sequence to determine whether the offset of the offset relationship data has a base value; for example, the effect of wind on the ship is 5 degrees per kilometer, and the effect of rudder structure failure on the ship is 5 degrees per kilometer; since the force of wind on the ship will gradually decrease with the change of the ship's heading, it is assumed that the offset of the ship by wind decreases by 1 degree for every 10 degrees of deflection of the ship. At one kilometer, the ship is offset by 10 degrees compared with the initial state, and at two kilometers, the ship is offset by 9 degrees compared with one kilometer. By analogy, the subsequent results are 8, 7, 6, 5, 5, and 5 respectively. The base value of the corresponding data change is 5. If there is no rudder structure failure, the results are 5, 4, 3, 2, 1, 0, 0, and 0. The base value of the corresponding data change is 0.

[0100] S7545, if it is determined that the base value of the deviation relationship data is greater than the built-in fluctuation base value, it is determined that the first deviation factor is a rudder structure failure; wherein the fluctuation base value is the degree of heading deviation caused by other unknown factors under normal conditions of the rudder and other equipment, which can be obtained through scientific experiments.

[0101] S7546: If it is determined that the base value of the offset relationship data is less than the built-in fluctuation base value, it is determined that the first offset factor is an external force.

[0102] In this embodiment, the offset of the offset relationship data is judged to be convergent, so as to determine whether it is a rudder structure failure. When it is determined that there is no convergence, a second judgment is made on the numerical change of the offset relationship data to further determine the cause of the ship's deviation. If there is a base value in the offset of the offset relationship data, it indicates that the heading deviation of the ship is changed on the basis of a fixed offset, so that the offset of the ship does not have convergence. Therefore, after eliminating the interference, the offset of the ship has convergence, and then it is determined that the first offset factor is a rudder structure failure. On the contrary, when it is determined that there is no base value, it indicates that the offset of the ship is a heading deviation caused by external interference, so the first offset factor is determined to be an external force.

[0103] By making convergence judgments and value change judgments on the offset of the offset relationship data, a first judgment on the rudder structure fault is made according to the convergence, and whether the offset of the offset relationship data has a base value is determined according to the value change, so as to make a second judgment on the rudder structure fault, thereby making the judgment result of the rudder structure fault more accurate, thereby improving the accuracy of ship fault judgment.

[0104] In step S755, if the offset relationship data is not unique, the change rule of the offset attribute of the offset relationship data is judged to determine the second offset factor; the second offset factor includes a rudder structure fault and a rudder connection fault, including the following steps:

[0105] S7551, based on the offset attribute of the offset relationship data, counting the offset amounts of the offset relationship data with the same offset attribute to obtain two single attribute offset data;

[0106] S7552, performing mathematical statistics on the two single attribute offset data to determine whether at least one of the two single attribute offset data has clustering properties;

[0107] S7553, if at least one of the two single attribute offset data has a clustering property, it is determined that the second offset factor is a rudder structure failure;

[0108] S7554: If the two single attribute offset data are not clustered, it is determined that the second offset factor is a rudder connection failure.

[0109] In this embodiment, when the offset relationship data is not singular, if there is an external force and a rudder structure failure, it indicates that the external force acting on the ship is greater than the force between the rudder structure failures, and the directions are opposite. At the same time, due to the discontinuity of the external force, the ship's driving path is similar to the driving path when it is unstable, so further judgment is needed. Since the intermittent time of the external force is not fixed, and the force of the structural failure is fixed, the deflection angle of the ship is fixed during the intermittent period of the external force. Therefore, a convergent judgment is made on the deflection angle in a single direction to determine whether it is a rudder structure failure.

[0110] By utilizing the offset attributes of each offset in the offset relationship data, the ship offset when external force is present and the ship offset when no external force is present can be screened out. Then, by judging the single attribute offset data, it can be determined whether the offset of the ship is fixed when no external force is present, thereby determining whether the ship offset is caused by a rudder structure failure. When the single attribute offset data is clustered, the clustered single attribute offset data is the offset caused by the rudder structure failure, thereby making the judgment of the rudder structure failure more accurate.

[0111] Based on the description of the embodiment of the ship fault prediction method based on artificial intelligence, the embodiment of the present invention also discloses a ship fault prediction system based on artificial intelligence:

[0112] like Figure 2As shown, a ship fault prediction system based on artificial intelligence, by applying the above-mentioned ship fault prediction method based on artificial intelligence, includes: a data acquisition module 1, a first fault judgment module 2, a second fault judgment module 3, and a fault alarm module 4;

[0113] The data acquisition module 1 records the travel path of the ship based on the ship system to obtain navigation data; and records the working data of the ship's steering gear based on the ship system to obtain steering direction data;

[0114] The first fault judgment module 2 compares the navigation data with the rudder data to determine whether the actual driving direction of the ship is the same as the guiding direction of the ship by the steering steering gear; if the actual driving direction of the ship is the same as the guiding direction of the ship by the steering steering gear, it is determined that the rudder of the ship is in a normal working state; if the actual driving direction of the ship is different from the guiding direction of the ship by the steering steering gear, it is determined that the rudder of the ship is in an abnormal working state, and the abnormal working state is analyzed;

[0115] The second fault judgment module 3 outputs a steering gear system reset signal and receives a steering gear system reset confirmation signal from a user when it is determined that the rudder of the ship is in an abnormal working state; resets the steering gear system based on the steering gear system reset confirmation signal, and records the ship's heading after the reset, judges the recorded heading data, and determines the heading deviation factor;

[0116] The fault alarm module 4 determines the fault source based on the heading deviation factor and outputs a corresponding alarm signal according to the fault source.

[0117] Compared with the existing ship fault prediction method and system based on artificial intelligence, the present invention improves the accuracy of fault judgment.

[0118] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A ship fault prediction method based on artificial intelligence, characterized in that: include: The ship's travel path is recorded based on the ship system to obtain navigation data; Based on the ship system, the working data of the ship's steering gear is recorded to obtain the rudder direction data; Compare the navigation data with the steering data to determine whether the actual heading of the ship is the same as the heading guided by the steering gear; If the actual heading of the ship is the same as the heading guided by the steering gear, it is determined that the rudder of the ship is in normal working condition; If the actual heading of the ship is different from the heading guided by the steering gear, it is determined that the rudder of the ship is in an abnormal working state, and the abnormal working state is analyzed; When it is determined that the ship's rudder is in an abnormal working state, a steering gear system reset signal is output, and a steering gear system reset confirmation signal from the user is received; Reset the steering gear system based on the steering gear system reset confirmation signal, record the ship's heading after the reset, judge the recorded heading data, and determine the heading deviation factor; Based on the heading deviation factor, the fault source is determined, and the corresponding alarm signal is output according to the fault source.

2. The method for predicting ship failures based on artificial intelligence according to claim 1, characterized in that: The navigation data is compared with the steering data to determine whether the actual heading of the ship is the same as the heading guided by the steering gear to the ship, specifically: Based on the built-in ship simulation model, the working data of the steering servo, the working time of the corresponding working data and the speed data are input into the ship simulation model, and the driving course simulation is performed to obtain a simulated route map; Draw the navigation data into a driving route map based on the time data; Based on the time data, the driving route map under the same time axis is compared with the simulation route map; If the similarity between the driving route map and the simulated route map is greater than a preset similarity threshold, it is determined that the actual driving direction of the ship is the same as the guiding direction of the ship by the steering gear; If the similarity between the driving route map and the simulated route map is less than a preset similarity threshold, it is determined that the actual driving heading of the ship is different from the guiding heading of the ship by the steering gear.

3. The method for predicting ship failures based on artificial intelligence according to claim 1, characterized in that: The method of resetting the steering gear system based on the steering gear system reset confirmation signal, and recording the ship's heading after the reset, judging the recorded heading data, and determining the heading deviation factor is specifically as follows: The initial travel direction of the ship after reset is used as the basic reference direction, and the travel direction of the ship is recorded based on the built-in time interval to obtain the direction data; Compare the recorded heading data with the basic reference direction to determine whether the heading of the ship after reset has changed within the built-in time interval; If the ship's heading does not change within the built-in time interval, the abnormal working state of the ship's rudder is determined to be a system failure; If the ship's heading changes within the built-in time interval, the two adjacent heading data are compared based on the collection time of the heading data to determine the position offset relationship between the heading data collected later and the heading data collected earlier, and the position offset relationship is statistically analyzed to obtain the offset relationship data; Judge the offset relationship data and determine the offset factor.

4. The method for predicting ship failures based on artificial intelligence according to claim 3 is characterized in that: The offset relationship data is judged to determine the offset factor, specifically: Based on the built-in rule information, the heading data is processed to obtain the offset attribute of the offset relationship data; the rule information is that according to the collected time data, the heading data collected earlier is used as the reference heading of the heading data collected later, the heading data that changes counterclockwise relative to the reference heading is marked as a positive value, and the heading data that changes clockwise relative to the reference heading is marked as a negative value; Based on the offset attributes in the offset relationship data, the offset relationship data are compared for unity. If the offset attributes corresponding to the respective offset relationship data are all positive values ​​or all negative values, it is determined that the offset relationship data are unified. If the offset attributes corresponding to each offset relationship data are not all positive or negative, it is determined that the offset relationship data is not unique; If the offset relationship data is unique, the offset in the offset relationship data is judged to determine a first offset factor; the first offset factor includes external force and rudder structure failure; If the offset relationship data is not unique, a change rule of the offset attribute of the offset relationship data is determined to determine a second offset factor; the second offset factor includes a rudder structure fault and a rudder connection fault; The first offset factor and the second offset factor are collectively labeled as offset factors.

5. The method for predicting ship failures based on artificial intelligence according to claim 4, characterized in that: If the offset relationship data is unique, the offset in the offset relationship data is judged to determine the first offset factor; the first offset factor includes external force and rudder structure failure, specifically: Perform mathematical and statistical judgment on the offset of the offset relationship data to determine whether there is clustering between the offset relationship data; If it is determined that there is a clustering between the offset relationship data, then it is determined that the first offset factor is a rudder structure failure; If it is determined that there is no clustering between the offset relationship data, then the offset of the offset relationship data is judged according to the acquisition time sequence to determine whether the offset of the offset relationship data gradually decreases with time; If it is determined that the offset of the offset relationship data gradually decreases with time, a base value judgment of the value change of the offset of the offset relationship data is performed according to the acquisition time sequence to determine whether the offset of the offset relationship data has a base value; If it is determined that the base value of the offset relationship data is greater than the built-in fluctuation base value, then the first offset factor is determined to be a rudder structure failure; If it is determined that the base value of the offset relationship data is less than the built-in fluctuation base value, it is determined that the first offset factor is an external force.

6. The method for predicting ship failures based on artificial intelligence according to claim 5 is characterized in that If the offset relationship data is not unique, the change rule of the offset attribute of the offset relationship data is judged to determine the second offset factor; the second offset factor includes a rudder structure fault and a rudder connection fault, specifically: Based on the offset attribute of the offset relationship data, the offset amount of the offset relationship data with the same offset attribute is counted to obtain two single attribute offset data; Performing mathematical statistical judgment on the two single attribute offset data respectively to determine whether at least one of the two single attribute offset data has clustering properties; If at least one of the two single attribute offset data has a clustering property, it is determined that the second offset factor is a rudder structure failure; If the two single attribute offset data are not clustered, it is determined that the second offset factor is a rudder connection failure.

7. A ship fault prediction system based on artificial intelligence, characterized in that: The system is used to implement a ship fault prediction method based on artificial intelligence as described in any one of claims 1 to 6: comprising: a data acquisition module, a first fault judgment module, a second fault judgment module, and a fault alarm module; The data acquisition module records the travel path of the ship based on the ship system to obtain navigation data; and records the working data of the steering gear of the ship based on the ship system to obtain steering direction data; The first fault judgment module compares the navigation data with the rudder data to determine whether the actual driving direction of the ship is the same as the guiding direction of the ship by the steering steering gear; if the actual driving direction of the ship is the same as the guiding direction of the ship by the steering steering gear, it is determined that the rudder of the ship is in a normal working state; if the actual driving direction of the ship is different from the guiding direction of the ship by the steering steering gear, it is determined that the rudder of the ship is in an abnormal working state, and the abnormal working state is analyzed; The second fault judgment module, when determining that the rudder of the ship is in an abnormal working state, outputs a steering gear system reset signal, and receives a steering gear system reset confirmation signal from a user; resets the steering gear system based on the steering gear system reset confirmation signal, and records the ship's heading after the reset, judges the recorded heading data, and determines the heading deviation factor; The fault alarm module determines the fault source based on the heading deviation factor and outputs a corresponding alarm signal according to the fault source.

Images