Large ship lock ship lockage early warning system based on historical database analysis

Abstract

A large ship lock ship lockage early warning system based on historical database analysis comprises the following steps: S1, collecting ship lock operation historical data and fault historical data of a ship, and carrying out distribution similarity processing and common induction processing on the historical data to obtain a processed historical database; s2, performing time series data comparison analysis on the data set of each stage of real-time ship lockage and the database, and performing real-time early warning on abnormal vectors; and S3, prompting a ship party or ship lock personnel to check and analyze, thereby ensuring safe and efficient lockage of the ship and minimizing the accident probability and influence. According to the large ship lock ship lockage early warning system based on historical database analysis, data statistics and vectorization analysis are carried out on ship lockage conditions, and an application basis is established for a subsequent ship database.

Description

Technical Field

[0001] The invention relates to the technical field of ship lock control, in particular to a large ship lock ship passing early warning system based on historical database analysis. Background Art

[0002] In order to strengthen the monitoring and analysis of the operation status of ships in locks, it is necessary to accurately grasp the possible fault conditions. As the number of ships passing through the locks increases and the types become more complex, higher requirements are placed on the safe operation of the locks. However, there are relatively few studies on the real-time operation data collection of ships passing through the locks and the analysis of abnormal warning information of ships passing through the locks. To solve the above problems, this paper proposes a large-scale ship lock passing warning system based on historical database analysis. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a large-scale ship lock ship passage early warning system based on historical database analysis, by collecting historical data of ship operation and fault history through the lock, performing distribution similarity processing and common induction processing on the historical data, and performing time series data comparison and analysis on the data sets of each stage of real-time ship passage and the database, and giving timely alarm prompts for abnormal situations, so as to inform the ship or lock personnel to conduct inspection and analysis, ensure the safe and efficient passage of the ship and minimize the probability and impact of accidents.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is: a large ship lock ship passing through the lock early warning system based on historical database analysis, comprising the following steps: S1: By collecting the historical data of ship operation and fault history through the lock, the historical data are processed for distribution similarity and common induction to obtain the processed historical database; S2: Compare and analyze the real-time ship passing through the lock data set at each stage with the database, and issue real-time warning for abnormal vectors; S3: Prompt the ship or lock personnel to conduct inspection and analysis to ensure the safe and efficient passage of the ship through the lock and minimize the probability and impact of accidents.

[0005] Preferably, step S1 comprises the following steps: S1-1: Collect historical data of ships passing through the lock. When a ship passes through the lock, the logical process of waiting for the lock, entering the lock, lock servo, and exiting the lock can be decomposed into the ship's navigation process, dispatching and commanding process, lock equipment and facility state conversion process, and lock water delivery process according to the time and space sequence. Each process contains lock equipment operation status data, lock control system data, dispatching system business data, gear command process data, and ship automatic identification system data; S1-2: Collect historical data on ship lock failures, record the type, time, location, frequency and level of failures; and collect data on the ship that experienced the failure, including ship type, load and cargo; S1-3: Based on step 1-1, the historical data is processed to calibrate the tasks that should be sampled in the current iteration step. For tasks that require further sequential sampling, , , through the maximization formula: ; Find the current target task ; where C represents the number of samples of the sampling task data, C' represents the number of data points of a certain sampling task, and Q(c) represents the prediction quality of the multi-task Gaussian process model for each task. To describe, x is the sampling data point (data of ships passing through a certain lock), D is the data description, and the prediction variance is The task association information is implicitly included, and the coefficient It can help to explicitly consider the impact of related information and task computational efficiency in asymmetric task scenarios; S1-4: Based on steps 1-3, for the selected tasks , construct an adaptive sampling criterion from the perspective of model bias-variance decomposition by maximizing: ; Finding new points , the variance term in the formula It represents the uncertainty of the model prediction at point x, tends to be a global search, and can be based on the sample Euclidean distance metric; the deviation term in the formula It represents the deviation between the model prediction and the actual observation value at point x, and describes the local characteristics of the task output. This item can be measured by the leave-one-out cross-validation error. The coefficient It is a trade-off between local and global search. Reasonable dynamic trade-off can effectively improve sampling quality. S1-5: Based on steps 1-4, establish an automatic matching and fusion relationship between the planning information in the ship dispatching system, the historical fault information of the ship and the lock control system; establish a connection between the ship's lock number information and the lock operation process control system data in the order of time and space position, and associate and fuse the ship's lock number with the lock control system operation data and fault data according to the time tag.

[0006] Preferably, step S2 comprises the following steps: S2-1: Each sample collected and processed during the basic time sequence of ship navigation, dispatching and commanding (human-computer interaction), lock equipment and facility state conversion, and lock water delivery, as well as each information sample of the faulty ship, is represented as a vector. The vectors generated by different process models include the embedding of spatial position information and time information. The data processing model is: ; The goal of the data and model after the fusion of each information is to obtain the best model f, so that the actual time of the ship passing through the hub lock and the model prediction calculation time are minimized: ; In the formula Indicates the actual time taken to pass the gate. Indicates the time required for model prediction calculation; the maximum value of the actual time of a ship passing through the hub lock and the model prediction calculation time is: ; S2-2: Decompose the process of a ship passing through a lock into subtask sets according to the direction of the ship passing through the lock; S2-3: Aggregate the manipulation data corresponding to each local model according to a preset time period, and perform standardization on the aggregated manipulation data to obtain a historical manipulation behavior record and a fault history record; S2-4: Mark the historical operation behavior records with identification information, using the identification information, fault information and historical operation behavior records as input parameters. When the time a ship passes through the lock or the current equipment operation time exceeds the maximum value in the database, or is less than the minimum value in the archived database, the system will list the current lock vector as an abnormal variable and issue an alarm.

[0007] Preferably, the step S2-2 comprises the following steps: S2-2-1: Ship navigation process through the lock. The time series data of ship navigation (including entry and exit) is mainly derived from the ship supervision system data. The coordinate curve of the cumulative distance and time of ship navigation is established: ; In the formula, the ship's navigation track point time The track position status Convert to Accumulated distance over time The relationship between the trajectory point moment Convert to curve time , the trajectory position state The accumulated distance between the initial trajectory points is associated with the average speed of the ship passing through the lock, and then the average entry and exit time of the lock passing unit of a group of ships bundled through the lock is obtained; S2-2-2: The ship lock servo process, i.e. the operation of the ship lock equipment, includes the opening and closing time of the upstream and downstream gate valve equipment, and the key technical parameters of the operation process. Since the equipment and control system of each gate are not completely consistent, the operation time sequence of each gate equipment is embedded with the data number. The operation time sequence of the ship lock equipment is Through multi-task learning sampling and model training, the training data set of each gate task is divided into independent sub-datasets according to the gate. , train the multi-task Gaussian process expert model of the corresponding lock head, extract the equipment operation data characteristics and fault characteristics; perform feature fusion to obtain the expert model prediction value, and then derive the operation time results and fault characteristic results of the lock equipment; S2-2-3: Human-computer interaction process of ship passing through the lock. The process information comes from the business data of the navigation dispatching system of the ship lock, the ship supervision information and the operation information of the lock monitoring and management facility system; the business data of the navigation dispatching system includes the ship passing through the lock declaration information and the dispatching plan; among them, the lock declaration information includes the declaration method, declaration time, ship information, and ship passing through the lock declaration items; the dispatching plan information includes the lock number, lock number, course, number of barges, rated passenger and cargo capacity, cargo type, actual passenger and cargo capacity, formation, lock passing method, and lock passing start and end time. According to the lock operation history data collected by multi-task learning, the historical data of the same type of ships with faults are recorded at the same time to establish a data model , the data points are assigned to multiple local models corresponding to the locks and trained separately.

[0008] Preferably, step S2-4 comprises the following steps: 2-4-1: The time series data of the ship passing through the lock is normalized and the clock is synchronized. The adaptive sampling data Oi is normalized and standardized into a unified format, and the drift abnormal data and error records are filtered and cleared. The sub-series data is detected through multiple mutation points to calculate the sub-series fluctuation amount and eliminate the order of magnitude difference between the fluctuation values; 2-4-2: Analyze the periodic characteristics of time series data. When a ship arrives at the lock chamber, the next ship arrives at the lock chamber. The time series of the two adjacent lock-passing units are consistent with the lock operation rhythm, and there are similarities and periodicity. 2-4-3: Obtain the real-time data V(i) collected by the ship supervision system during the process of the same number of ships passing through the lock. After normalization and clock synchronization preprocessing, the data is unified into a standard format. The distance between the ship passing time series and the lock operation process time series is calculated based on the Euclidean distance algorithm. , where m is the length of the time series, oi and vi are the time series samples of ships and locks, respectively.

[0009] The present invention provides a large-scale ship lock ship passage warning system based on historical database analysis, which has the following beneficial effects: 1. Data statistics and vector analysis are performed on the ship passage situation to establish an application basis for the subsequent ship database; 2. The operation of ships passing through the lock is essentially a scene in which the traffic organization process of a hub navigation is continuously repeated, with the characteristics of fixed spatial range and high process similarity. Through the collection, vector processing and comparative analysis of the data of this patent, there is a passive and active full-process monitoring of the safety of ships passing through the lock, providing certain theoretical and scientific support for the efficient and automated operation of the lock. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0011] like Figure 1 As shown, a large ship lock ship passing warning system based on historical database analysis includes the following steps: S1: By collecting the historical data of ship operation and fault history through the lock, the historical data are processed for distribution similarity and common induction to obtain the processed historical database; S2: Compare and analyze the real-time ship passing through the lock data set at each stage with the database, and issue real-time warning for abnormal vectors; S3: Prompt the ship or lock personnel to conduct inspection and analysis to ensure the safe and efficient passage of the ship through the lock and minimize the probability and impact of accidents.

[0012] Preferably, step S1 comprises the following steps: S1-1: Collect historical data of ships passing through the lock. When a ship passes through the lock, the logical process of waiting for the lock, entering the lock, lock servo, and exiting the lock can be decomposed into the ship's navigation process, dispatching and commanding process, lock equipment and facility state conversion process, and lock water delivery process according to the time and space sequence. Each process contains lock equipment operation status data, lock control system data, dispatching system business data, gear command process data, and ship automatic identification system data; S1-2: Collect historical data on ship lock failures, record the type, time, location, frequency and level of failures; and collect data on the ship that experienced the failure, including ship type, load and cargo; S1-3: Based on step 1-1, the historical data is processed to calibrate the tasks that should be sampled in the current iteration step. For tasks that require further sequential sampling, , , through the maximization formula: ; Find the current target task ; where C represents the number of samples of the sampling task data, C' represents the number of data points of a certain sampling task, and Q(c) represents the prediction quality of the multi-task Gaussian process model for each task. To describe, x is the sampling data point (data of ships passing through a certain lock), D is the data description, and the prediction variance is The task association information is implicitly included, and the coefficient It can help to explicitly consider the impact of related information and task computational efficiency in asymmetric task scenarios; S1-4: Based on steps 1-3, for the selected tasks , construct an adaptive sampling criterion from the perspective of model bias-variance decomposition by maximizing: ; Finding new points , the variance term in the formula It represents the uncertainty of the model prediction at point x, tends to be a global search, and can be based on the sample Euclidean distance metric; the deviation term in the formula It represents the deviation between the model prediction and the actual observation value at point x, and describes the local characteristics of the task output. This item can be measured by the leave-one-out cross-validation error. The coefficient It is a trade-off between local and global search. Reasonable dynamic trade-off can effectively improve sampling quality. S1-5: Based on steps 1-4, establish an automatic matching and fusion relationship between the planning information in the ship dispatching system, the historical fault information of the ship and the lock control system; establish a connection between the ship's lock number information and the lock operation process control system data in the order of time and space position, and associate and fuse the ship's lock number with the lock control system operation data and fault data according to the time tag.

[0013] Preferably, step S2 comprises the following steps: S2-1: Each sample collected and processed during the basic time sequence of ship navigation, dispatching and commanding (human-computer interaction), lock equipment and facility state conversion, and lock water delivery, as well as each information sample of the faulty ship, is represented as a vector. The vectors generated by different process models include the embedding of spatial position information and time information. The data processing model is: ; The goal of the data and model after the fusion of each information is to obtain the best model f, so that the actual time of the ship passing through the hub lock and the model prediction calculation time are minimized: ; In the formula Indicates the actual time taken to pass the gate. Indicates the time required for model prediction calculation; the maximum value of the actual time of a ship passing through the hub lock and the model prediction calculation time is: ; S2-2: Decompose the process of a ship passing through a lock into subtask sets according to the direction of the ship passing through the lock; S2-3: Aggregate the manipulation data corresponding to each local model according to a preset time period, and perform standardization on the aggregated manipulation data to obtain a historical manipulation behavior record and a fault history record; S2-4: Mark the historical operation behavior records with identification information, using the identification information, fault information and historical operation behavior records as input parameters. When the time a ship passes through the lock or the current equipment operation time exceeds the maximum value in the database, or is less than the minimum value in the archived database, the system will list the current lock vector as an abnormal variable and issue an alarm.

[0014] Preferably, the step S2-2 comprises the following steps: S2-2-1: Ship navigation process through the lock. The time series data of ship navigation (including entry and exit) is mainly derived from the ship supervision system data. The coordinate curve of the cumulative distance and time of ship navigation is established: ; In the formula, the ship's navigation track point time The track position status Convert to Accumulated distance over time The relationship between the trajectory point moment Convert to curve time , the trajectory position state The accumulated distance between the initial trajectory points is associated with the average speed of the ship passing through the lock, and then the average entry and exit time of the lock passing unit of a group of ships bundled through the lock is obtained; S2-2-2: The ship lock servo process, i.e. the operation of the ship lock equipment, mainly includes the opening and closing time of the upstream and downstream gate valve equipment, and the key technical parameters of the operation process. Since the equipment and control system of each gate are not completely consistent, the operation time sequence of each gate equipment is embedded with the data number. The operation time sequence of the ship lock equipment is Through multi-task learning sampling and model training, the training data set of each gate task is divided into independent sub-datasets according to the gate. , train the multi-task Gaussian process expert model of the corresponding lock head, extract the equipment operation data characteristics and fault characteristics; perform feature fusion to obtain the expert model prediction value, and then derive the operation time results and fault characteristic results of the lock equipment; S2-2-3: Human-computer interaction process of ship passing through the lock. The process information mainly comes from the business data of the navigation dispatching system of the ship lock, the ship supervision information and the operation information of the lock monitoring and management facility system; the business data of the navigation dispatching system mainly includes the ship passing through the lock declaration information and the dispatching plan; among them, the lock declaration information includes the declaration method, declaration time, ship information, ship passing through the lock declaration items, etc.; the dispatching plan information includes the lock number, lock number, course, number of barges, rated passenger and cargo capacity, cargo type, actual passenger and cargo capacity, formation, lock passing method, lock passing start and end time, according to the lock operation history data collected by multi-task learning, and at the same time record the historical data of the same type of ships with faults, to establish a data model , the data points are assigned to multiple local models corresponding to the locks and trained separately.

[0015] Preferably, step S2-4 comprises the following steps: 2-4-1: The time series data of the ship passing through the lock is normalized and the clock is synchronized. The adaptive sampling data Oi is normalized and standardized into a unified format, and the drift abnormal data and error records are filtered and cleared. The sub-series data is detected through multiple mutation points to calculate the sub-series fluctuation amount and eliminate the order of magnitude difference between the fluctuation values; 2-4-2: Analyze the periodic characteristics of time series data. When a ship arrives at the lock chamber, the next ship arrives at the lock chamber. The time series of the two adjacent lock-passing units are consistent with the lock operation rhythm, and there are similarities and periodicity. 2-4-3: Obtain the real-time data V(i) collected by the ship supervision system during the process of the same number of ships passing through the lock. After normalization and clock synchronization preprocessing, the data is unified into a standard format. The distance between the ship passing time series and the lock operation process time series is calculated based on the Euclidean distance algorithm. , where m is the length of the time series, oi and vi are the time series samples of ships and locks, respectively.

[0016] Implementation case 1: Step 1: After the first miter gate of the multi-stage ship lock is opened, the corresponding four ship locks S1 start to enter the lock, and the start time is marked as TS1; Step 2: By establishing a ship lock database, statistically analyzing the entry time of four ships, the normal entry time range Mentry of the four ships is obtained. According to the actual situation of the ship lock, a reasonable warning range is set to obtain the warning value Mentry'; Step 3: By comparing the real-time ship lock entry situation, the corresponding lock entry time vector TS1 in the lock vector set is compared with the warning value Mentry' of the same number of lock entry ranges in the database; Step 4: When the real-time comparative analysis shows that TS1 is greater than M', it means that the S1 entry status of the lock is abnormal, and the centralized control personnel are prompted to check whether the abnormality is caused by ship factors, human factors or other conditions such as weather, so as to deal with it in time and reduce the impact of abnormal conditions on the efficiency of ship passing through the lock.

[0017] Implementation case 2: Step 1: Ship S1 starts to pass through the multi-stage ship lock, the first-stage miter gate starts to close the gate, and the time is marked as Tfirst stage; Step 2: By establishing a ship lock passing database, statistics are collected on the closing conditions of the first miter gate in different seasons to obtain the normal operating time range of the first miter gate M closing gate, and according to the actual situation of the ship lock, a reasonable warning range is set to obtain the warning value M entering the lock'; Step 3: by correlating and comparing the real-time ship lock entry situation, the first-stage miter gate closing time Tfirst-stage corresponding to the lock vector concentration is compared with the first-stage miter gate normal operation time range warning value Mentry' in the database; Step 4: When the real-time comparison analysis shows that TS1 is not within the value M'entry' range, it means that the gate closing state of S1 at this lock is abnormal, and the control personnel are prompted to check whether the equipment factors cause the fault, so as to remind the control personnel to deal with it in time and reduce the impact of abnormal conditions on the efficiency of ship passing through the lock.

[0018] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limiting the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A large ship lock ship passing warning system based on historical database analysis, characterized in that: The following steps are involved: S1: By collecting the historical data of ship operation and fault history through the lock, the historical data are processed for distribution similarity and common induction to obtain the processed historical database; S2: Compare and analyze the real-time ship passing through the lock data set at each stage with the database, and issue real-time warning for abnormal vectors; S3: Prompt the ship or lock personnel to conduct inspection and analysis to ensure the safe and efficient passage of the ship through the lock and minimize the probability and impact of accidents.

2. According to claim 1, a large ship lock ship passing warning system based on historical database analysis is characterized in that: The step S1 comprises the following steps: S1-1: Collect historical data of ships passing through the lock. When a ship passes through the lock, the logical process of waiting for the lock, entering the lock, lock servo, and exiting the lock can be decomposed into the ship's navigation process, dispatching and commanding process, lock equipment and facility state conversion process, and lock water delivery process according to the time and space sequence. Each process contains lock equipment operation status data, lock control system data, dispatching system business data, gear command process data, and ship automatic identification system data; S1-2: Collect historical data on ship lock failures, record the type, time, location, frequency and level of failures; and collect data on the ship that experienced the failure, including ship type, load and cargo; S1-3: Based on step 1-1, the historical data is processed to calibrate the tasks that should be sampled in the current iteration step. For tasks that require further sequential sampling, , , through the maximization formula: ； Find the current target task ; where C represents the number of samples of the sampling task data, C' represents the number of data points of a certain sampling task, and Q(c) represents the prediction quality of the multi-task Gaussian process model for each task. To characterize, x is the sampling data point, D is the data description, and the prediction variance The task association information is implicitly included, and the coefficient It can help to explicitly consider the impact of related information and task computational efficiency in asymmetric task scenarios; S1-4: Based on steps 1-3, for the selected tasks , construct an adaptive sampling criterion from the perspective of model bias-variance decomposition by maximizing: ； Finding new points , the variance term in the formula It represents the uncertainty of the model prediction at point x, tends to be a global search, and can be based on the sample Euclidean distance metric; the deviation term in the formula It represents the deviation between the model prediction and the actual observation value at point x, and describes the local characteristics of the task output. This item can be measured by the leave-one-out cross-validation error. The coefficient It is a trade-off between local and global search. Reasonable dynamic trade-off can effectively improve sampling quality. S1-5: Based on steps 1-4, establish an automatic matching and fusion relationship between the planning information in the ship dispatching system, the historical fault information of the ship and the lock control system; establish a connection between the ship's lock number information and the lock operation process control system data in the order of time and space position, and associate and fuse the ship's lock number with the lock control system operation data and fault data according to the time tag.

3. According to claim 1, a large ship lock ship passing warning system based on historical database analysis is characterized in that: The step S2 comprises the following steps: S2-1: Each sample collected and processed during the basic time sequence of ship navigation, dispatching and commanding, lock equipment and facility status conversion, and lock water delivery when the ship passes through the lock, as well as each information sample of the faulty ship, is represented as a vector. The vectors generated by different process models include the embedding of spatial position information and time information. The data processing model is: ； The goal of the data and model after the fusion of each information is to obtain the best model f, so that the actual time of the ship passing through the hub lock and the model prediction calculation time are minimized: ； In the formula Indicates the actual time taken to pass the gate. Indicates the time required for model prediction calculation; the maximum value of the actual time of a ship passing through the hub lock and the model prediction calculation time is: ； S2-2: Decompose the process of ships passing through the lock into subtask sets according to the direction of the ships passing through the lock; S2-3: Aggregate the manipulation data corresponding to each local model according to a preset time period, and perform standardization on the aggregated manipulation data to obtain a historical manipulation behavior record and a fault history record; S2-4: Mark the historical operation behavior records with identification information, using the identification information, fault information and historical operation behavior records as input parameters. When the time a ship passes through the lock or the current equipment operation time exceeds the maximum value in the database, or is less than the minimum value in the archived database, the system will list the current lock vector as an abnormal variable and issue an alarm.

4. According to claim 2, a large ship lock ship passing through the lock early warning system based on historical database analysis is characterized in that: The step S2-2 comprises the following steps: S2-2-1: During the ship's navigation process, the ship's navigation time series data mainly comes from the ship supervision system data, and a coordinate curve of the cumulative distance and time of the ship's navigation is established: ； In the formula, the ship's navigation track point time Track position status Convert to Accumulated distance over time The relationship between the trajectory point moment Convert to curve time , the trajectory position state The accumulated distance between the initial trajectory points is associated with the average speed of the ship passing through the lock by taking the derivative of the trajectory curve, and then the average entry and exit time of the lock passing unit of a group of ships bundled through the lock is obtained; S2-2-2: The ship lock servo process, i.e. the operation of the ship lock equipment, includes the opening and closing time of the upstream and downstream gate valve equipment, and the key technical parameters of the operation process. Since the equipment and control system of each gate are not completely consistent, the operation time sequence of each gate equipment is embedded with the data number. The operation time sequence of the ship lock equipment is Through multi-task learning sampling and model training, the training data set of each gate task is divided into independent sub-datasets according to the gate. , train the multi-task Gaussian process expert model of the corresponding lock head, extract the equipment operation data characteristics and fault characteristics; perform feature fusion to obtain the expert model prediction value, and then derive the operation time results and fault characteristic results of the lock equipment; S2-2-3: Human-computer interaction process of ship passing through the lock. The process information comes from the business data of the navigation dispatching system of the ship lock, the ship supervision information and the operation information of the lock monitoring and management facility system; the business data of the navigation dispatching system includes the ship passing through the lock declaration information and the dispatching plan; among them, the lock declaration information includes the declaration method, declaration time, ship information, and ship passing through the lock declaration items; the dispatching plan information includes the lock number, lock number, course, number of barges, rated passenger and cargo capacity, cargo type, actual passenger and cargo capacity, formation, lock passing method, and lock passing start and end time. According to the lock operation history data collected by multi-task learning, the historical data of the same type of ships with faults are recorded at the same time to establish a data model , the data points are assigned to multiple local models corresponding to the locks and trained separately.

5. According to claim 2, a large ship lock ship passing warning system based on historical database analysis is characterized in that: The step S2-4 comprises the following steps: 2-4-1: The time series data of the ship passing through the lock is normalized and the clock is synchronized. The adaptive sampling data Oi is normalized and standardized into a unified format, and the drift abnormal data and error records are filtered and cleared. The sub-series data is detected through multiple mutation points to calculate the sub-series fluctuation amount and eliminate the order of magnitude difference between the fluctuation values; 2-4-2: Analyze the periodic characteristics of time series data. When a ship arrives at the lock chamber, the next ship arrives at the lock chamber. The time series of the two adjacent lock-passing units are consistent with the lock operation rhythm, and there are similarities and periodicity. 2-4-3: Obtain the real-time data V(i) collected by the ship supervision system during the process of the same number of ships passing through the lock. After normalization and clock synchronization preprocessing, the data is unified into a standard format. The distance between the ship passing time series and the lock operation process time series is calculated based on the Euclidean distance algorithm. , where m is the length of the time series, oi and vi are the time series samples of ships and locks, respectively.

Images