SARIMA prediction-based inland river navigation draught optimization management and control method for ship with overrun draught

Abstract

According to the SARIMA prediction-based inland river navigation draught optimization management and control method for the ship with the overrun draught, a multi-source time sequence data set is obtained and preprocessed to obtain an input time sequence set, and the input time sequence set is processed through a constructed SARIMA model to obtain a water level prediction sequence; a ship dynamic draught prediction sequence is obtained through the constructed ship dynamic draught model based on water level prediction sequence processing, a draught safety margin judgment operator is constructed to carry out draught safety margin evaluation and risk grading, and finally a draught optimization management strategy is solved through the draught optimization management model and issued. A ship draught control closed-loop chain of prediction-calculation-evaluation-optimization-feedback is formed, through deep coupling of SARIMA time sequence prediction, ship dynamic draught calculation and draught optimization management, prospective recognition and dynamic regulation and control of draught safety margin of a draught overrun ship are achieved, and under the premise of guaranteeing navigation safety, the ship draught safety margin is effectively controlled. And the water depth utilization efficiency of the inland waterway and the ship loading efficiency are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of water traffic safety technology, and in particular to a method for optimizing and controlling the draft of vessels exceeding draft limits for inland waterway navigation based on SARIMA prediction. Background Technology

[0002] "Inland waterway navigation" refers to navigation activities conducted by vessels on rivers, waterways, lakes, and other inland waterways within a country or region. Compared to maritime navigation, inland waterway navigation in my country mainly occurs in the Yangtze River, Pearl River, Xijiang River, Beijing-Hangzhou Grand Canal, port basins, and upstream and downstream waterways of locks. "Vessels with excessive draft" refers to vessels whose actual draft exceeds the corresponding safety control limit or the maximum permissible draft stipulated by regulations under the conditions of a given waterway or section. Excessive draft generally includes, but is not limited to, the following situations: 1) After loading cargo and / or ballast water, the actual draft of the vessel exceeds the maximum permissible draft stipulated by the vessel itself or the waterway, bridge, lock, dam, or other facilities where it is located; 2) Although the actual draft of the vessel does not reach the mandatory maximum permissible draft limit, it exceeds the safety control draft set for inland waterways or specific sections; 3) The vessel's draft distribution is abnormal, including at least the bow draft, stern draft, or midship draft exceeding the limit, resulting in insufficient overall or partial safety margin for the vessel.

[0003] With the continuous growth of domestic shipping volume and the increasing number of large and heavy-duty vessels, the frequency of over-draft vessels appearing in inland waterways is showing an upward trend year by year. On the one hand, some shipping companies, in order to reduce transportation costs and increase revenue per voyage, tend to operate with heavy loads in their loading management, making it easy for vessels to approach or even exceed the safety control limits in specific water levels or shallow sections of the river. On the other hand, affected by factors such as seasonal changes in water levels, channel siltation, bridge-controlled water depth, and the dimensions of navigation structures, the safe draft margin of vessels under different river sections and operating conditions is further compressed, leading to the continuous accumulation and exposure of over-draft risks. In addition, the insufficient management level of some vessels in terms of ballast water allocation, trim and list control, and cargo stowage also objectively increases the probability of over-draft and abnormal draft distribution. The combined effect of these factors has resulted in an increasing number of over-draft vessels in my country's inland waterway shipping system, a wider distribution range, greater regulatory difficulty, and a surge in safety hazards.

[0004] Currently, there has been some research on the perception and analysis of inland waterway water levels and navigation environments. For example, the multi-station daily water level prediction method for inland waterways improves the accuracy of multi-station water level prediction through a "clustering-then-modeling" approach; the inland waterway navigation water area classification method achieves the classification of water areas with different navigation environments through cluster analysis; and bridge monitoring and early warning methods and ship navigation safety management methods improve water traffic safety from the perspectives of bridge area safety and multi-source data fusion early warning, respectively. However, the above technologies mostly focus on the identification, classification, and early warning of water levels or navigation environments themselves, and have not yet been deeply coupled with the specific draft control of ships, making it difficult to directly apply to the refined dynamic management of ships with excessive draft.

[0005] Traditional ship draft management methods are usually based on "water depth" The simplistic relationship of "fixed margin depth" involves taking the geometric depth from the water surface to the riverbed as the usable depth, assigning a "margin depth" based on empirical fixed values ​​or according to the draft ratio, and then deducing the allowable draft value from this. In this method, the margin depth is generally not broken down into components such as ship sinking, attitude changes, and the uncertainty of riverbed undulations, and it basically does not change in real time with current velocity, draft, loading conditions, and time. Consequently, draft checks are mostly completed statically during loading or berthing and unberthing phases, which makes it difficult to reflect the impact of future water level changes on draft safety margins in a timely manner, and also fails to reflect the time-varying characteristics of dynamic draft of ships under different navigation sections and operating conditions.

[0006] On the other hand, current methods for predicting and managing ship draft largely rely on the experience of navigators and shipping companies, or on simple estimations based on rather crude water level statistics, lacking a systematic characterization of the seasonality, periodicity, and randomness of water levels. Under the complex hydrological conditions of inland waterways, water levels exhibit significant seasonal variations throughout the year, and are also influenced by factors such as flow rate, scheduling, reservoir operation, and local rainfall, exhibiting multi-scale fluctuation characteristics. If static water levels or empirical safety margins are still used for draft control, it is easy to encounter two extreme situations: either "overly conservative approaches leading to insufficient loading" or "insufficient safety margins leading to excessive risk."

[0007] The existing draft management model still has obvious phased and lagging characteristics: management measures are mainly focused on post-event inspection and passive handling, such as taking measures such as detention, load reduction or passage restriction after on-site measurement finds that the draft exceeds the limit. There is a lack of a joint prediction and pre-control mechanism for water level and operating conditions in the future. There is also a lack of an integrated model that organically links the water level prediction results with the ship's dynamic draft calculation, draft safety margin assessment and control decisions such as loading / speed / attitude, making it difficult to form a closed-loop management chain of "prediction-calculation-evaluation-optimization-feedback".

[0008] In summary, existing technologies have at least several shortcomings: 1) Water level and draft are mostly based on static or empirical values, failing to fully consider the seasonal evolution of water level and the dynamic changes in ship draft, resulting in insufficient draft prediction capabilities; 2) The margin of water depth is not broken down into measurable dynamic sub-items such as fuel consumption, attitude changes, shallow water subsidence, and riverbed siltation, making it difficult to calculate and update the draft safety margin in real time; 3) There is a lack of quantitative assessment methods that unify multiple types of draft control limits under the same criterion, making it difficult to classify and finely supervise ships exceeding draft limits in the time-segment dimension; 4) Existing management methods are mostly based on post-event handling and static limit control, and have not yet formed a draft optimization management model driven by prediction results, making it impossible to fully explore the potential of water level windows and channel depth while ensuring safety. Summary of the Invention

[0009] In view of this, the present invention proposes a method for optimizing and controlling the draft of vessels with excessive draft based on SARIMA prediction for inland waterway navigation. This method enables the forward-looking identification and dynamic control of the draft safety margin of vessels with excessive draft, thereby significantly improving the efficiency of channel depth utilization and vessel loading efficiency while ensuring the safety of inland waterway navigation.

[0010] The technical solution of this invention is implemented as follows: The first aspect of this invention provides a method for optimizing and controlling the draft of vessels exceeding draft limits for inland waterway navigation based on SARIMA prediction, comprising the following steps: Step S1: Obtain hydrological and meteorological data, channel condition data, and navigation data of vessels with excessive draft for the target inland waterway section, and construct a multi-source time series dataset of inland waterway navigation for vessels with excessive draft. Step S2: Preprocess the multi-source time series dataset to obtain the input time series set; Step S3: Construct a SARIMA model for predicting water levels in the target inland waterway section, and obtain the water level prediction sequence in the future time domain based on the input time series set; Step S4: Decompose the dynamic surplus water depth into several measurable sub-items, construct a ship dynamic draft model based on the water level prediction sequence and the dynamic surplus water depth, and predict the ship dynamic draft prediction sequence in the time domain using the ship dynamic draft model. Step S5: Define the comprehensive minimum control limit of the target inland waterway section at a certain predicted time, and construct a draft safety margin determination operator with the predicted dynamic draft of the ship at the corresponding time. Based on the full draft margin determination operator, obtain the ship draft safety margin assessment result and risk classification result. Step S6: Construct a draft optimization management model. The draft optimization management model is based on the ship's dynamic draft prediction sequence and the ship's draft safety margin assessment results to obtain the draft optimization management strategy, and the draft optimization management strategy is distributed to the ship and shore-based scheduling system.

[0011] Preferably, the hydrological and meteorological data includes river level, flow rate, water depth, flow velocity, wind speed, and wind direction; the waterway condition data includes waterway maintenance depth, annual riverbed siltation, bridge area control depth, and dimensions of navigation structures; and the navigation data for vessels exceeding draft limits includes vessel load, ballast water configuration, bow draft, stern draft and / or midship draft, trim and heel parameters, speed, heading, track, and vessel size parameters.

[0012] Preferably, step S2 includes the following specific steps: Denoising is performed on hydrological and meteorological data, waterway condition data, and navigation data of vessels with excessive draft, removing meaningless records, outliers, and duplicate data. Error correction processing is performed on data containing symbol errors, encoding errors, and reversed time sequences; Standardize the format of various types of data by unifying units and decimal places; Add a timestamp and a spatial stamp to each record, and use the timestamp + spatial stamp as a composite index to identify duplicate records and delete redundant records; By interpolating and filling in missing data using adjacent time points or adjacent observations, a time-continuous and data-complete sequence can be obtained. Normalize or standardize the key variables to construct an input time series set with a unified time axis.

[0013] Preferably, the SARIMA model construction step in step S3 is as follows: The monthly minimum water level sequence of water level stations along the ship's navigation route is used as the modeling object. Outliers and duplicate records are removed, and the water level unit and time scale are unified to form a water level time series sorted by month. STL decomposition was used to perform seasonal trend decomposition on the water level time series, yielding seasonal components, trend components, and residual components. The stationarity of water level time series is determined by using the ADF test. For non-stationary series, first-order constant difference and necessary seasonal difference are performed until the stationarity test is passed at the preset significance level. Based on the obtained stationary series, the autocorrelation function (ACF) plot and the partial autocorrelation function (PACF) plot are plotted. The autoregression order p is determined according to the significant truncation position of PACF, and the moving average order q is determined according to the significant truncation position of ACF. Combined with the seasonal period s and the seasonal difference order D, the SARIMA model parameters are determined. In the SARIMA model, flow-related exogenous variables are introduced. The monthly average flow and its changes over the past few years are used as exogenous inputs. The direction and intensity of the influence of flow on water level prediction are characterized by regression coefficients. Based on the AIC / BIC and residual tests, different order combinations were compared, and the SARIMA model with the best fit and generalization performance was selected as the target water level prediction model.

[0014] Preferably, the flow-related exogenous variables include the average flow of this month, the difference between the average flow of this month and the average flow of the previous month, and the difference between the average flow of this month and the average flow of the previous two months. The regression coefficients of the exogenous variables are used to reflect the positive or negative impact and the intensity of the impact of the corresponding flow characteristics on the corrected water level.

[0015] Preferably, the calculation formula for the ship's dynamic draft model is as follows:

[0016] in For the dynamic draft of the ship, Due to the depth of the waterway, This is the predicted water level. The dynamic margin of water depth for a ship is decomposed into several measurable sub-items, and the specific calculation formula is as follows:

[0017] in This refers to the change in draft caused by the consumption of ship fuel oil and fresh water. This refers to the change in draft caused by changes in the ship's attitude. This refers to the change in draft caused by the ship's sinking. Prepare silt for the riverbed.

[0018] Preferably, the change in draft caused by the consumption of ship fuel and fresh water is calculated by using the ship's speed and draft records at the current and historical times to determine the rate of change in draft corresponding to fuel consumption per unit time and per unit speed, and combined with the planned future speed. The change in draft caused by the change in the ship's attitude is calculated by the difference between the heel and pitch angles at the current time and the predicted future time. The change in draft caused by the ship's sinking is calculated based on the current water depth-to-draft ratio, the ship's block coefficient, and the planned future speed, using the empirical formula for dynamic water sinking at shallow water speeds. The riverbed siltation reserve is calculated by converting the annual scouring and silting thickness of the riverbed obtained from the survey of the waterway management department to the predicted time, thus obtaining the riverbed siltation reserve thickness at the future time.

[0019] Preferably, the formula for calculating the comprehensive minimum control limit is:

[0020] in For the target inland waterway segment at the predicted time The comprehensive minimum control limit, For the maximum permissible draft of the ship structure, The maximum permissible draft for navigable waterways, To restrict the draft in the bridge area, To ensure safe control of draft; The expression for the draft safety margin determination operator is:

[0021] in For the predicted time The draft safety margin determination value, For the predicted time Predicted dynamic draft of ships; The minimum draft safety margin determination value is obtained based on the draft safety margin determination operator within a given time period. The minimum draft safety margin determination value is compared with a preset threshold to obtain a risk classification result including safe, concerned, warning and high risk.

[0022] Preferably, the specific steps of step S6 are as follows: The decision variables are cargo hold loading capacity, ballast water volume in each ballast tank, ship attitude control, ship speed in different river sections, and time window for passing through restricted river sections. The constraints include a draft safety constraint that the ship's dynamic draft does not exceed the comprehensive minimum control limit at each predicted time, and a capacity and loading constraint that the capacity of each cargo hold and ballast tank is not exceeded. To maximize the effective cargo capacity while also considering fuel consumption costs under the premise of meeting draft safety constraints, an objective function is constructed. The optimization problem is solved using mixed-integer linear programming or heuristic algorithms, and the draft optimization management strategy is obtained. The draft optimization management strategy is then distributed to the ship and shore-based scheduling system.

[0023] Preferably, the expression for the objective function is:

[0024] in For the ship speed vector of each river segment, For cargo hold loading capacity, For cargo hold number n, The ballast water volume of each ballast tank, The ballast water volume of ballast tank number m. These are the pitch control parameters. As a weighting factor, For fuel consumption estimation functions related to speed configuration, This refers to the total amount of ballast water.

[0025] Secondly, the present invention provides a draft optimization and control device for inland waterway navigation of vessels exceeding draft limits based on SARIMA prediction, comprising: The data acquisition module is used to acquire hydrological and meteorological data, channel condition data, and navigation data of vessels with overdrafts in the target inland waterway section, and to construct a multi-source time series dataset of inland waterway navigation for vessels with overdrafts. The preprocessing module is used to perform noise reduction, error correction, unit and format standardization, time and space alignment, missing value imputation, and normalization or standardization on multi-source time series data to obtain the input time series set for draft prediction. The water level prediction module constructs a SARIMA model for predicting water levels in the target inland waterway section, and obtains the water level prediction sequence in the future time domain based on the input time series set. The dynamic draft calculation module is used to decompose the dynamic surplus water depth into several measurable sub-items, construct a ship dynamic draft model based on the water level prediction sequence and the dynamic surplus water depth, and predict the ship dynamic draft prediction sequence in the time domain using the ship dynamic draft model. The safety margin assessment module is used to define the comprehensive minimum control limit of the target inland waterway section at a certain predicted time by combining multiple types of draft control limits, including the maximum permissible draft of the ship structure, the maximum permissible draft of the navigable channel, the restricted draft of the bridge area, and the safety control draft. It also constructs a draft safety margin determination operator with the predicted dynamic draft of the ship at the corresponding time. Based on the full draft margin determination operator, the ship's draft safety margin assessment results and risk classification results are obtained. The optimization decision module is used to construct a draft optimization management model, which is based on the ship's dynamic draft prediction sequence and the ship's draft safety margin assessment results to obtain the draft optimization management strategy.

[0026] The strategy distribution and feedback module is used to distribute multiple draft optimization management strategies, including load adjustment, ballast water configuration optimization, attitude control, speed and navigation timing, to the ship and shore-based scheduling system through the application layer and presentation layer, and collect feedback data to iteratively correct the SARIMA model, ship dynamic draft model and draft optimization management strategies.

[0027] Thirdly, the present invention provides an electronic device, including a memory and a processor, wherein the memory is used to store a program; The processor is coupled to the memory to execute the program to implement the steps of the method for optimizing the draft control of inland waterway navigation of vessels with excessive draft based on SARIMA prediction, thereby realizing automated and intelligent management of vessels with excessive draft.

[0028] Fourthly, the present invention provides a computer-readable storage medium storing one or more programs that can be executed by one or more processors, for implementing the steps of a method for optimizing and controlling the draft of vessels exceeding draft limits for inland waterway navigation based on SARIMA prediction.

[0029] Compared with the prior art, the beneficial effects of the present invention are: ① It breaks through the traditional model of calculating static draft based on fixed surplus water depth, introduces the SARIMA time series model to predict the water level of the target inland waterway section, and constructs a dynamic draft model of "water depth + predicted water level - dynamic surplus water depth" in combination with the waterway water depth. It decomposes uncertain factors such as fuel consumption, attitude change, ship sinking and riverbed siltation into measurable sub-items, so that draft and surplus water depth are transformed from static empirical values ​​into dynamic quantities that can be calculated in real time.

[0030] ② Combining multiple control limits such as the maximum permissible draft of the ship structure, the maximum permissible draft of the navigable channel, the restricted draft in the bridge area, and the safety control draft, a comprehensive minimum control limit and draft safety margin judgment operator are proposed, and a comprehensive risk index is constructed. This enables a graded and quantitative assessment of ships exceeding the draft limit in the time-segment dimension, providing management departments with a refined risk ranking basis.

[0031] ③ The draft prediction results are coupled with multiple decision variables such as loading, ballast, attitude, speed and navigation timing to construct a draft optimization management model that takes into account both safety and economy. The model is then continuously solved in the rolling time domain through intelligent optimization algorithms, which transforms draft management from a passive adjustment after the fact to an active optimization before the fact, effectively improving the utilization rate of inland waterway depth and the loading efficiency of ships.

[0032] ④ Through the collaboration of the data acquisition and preprocessing layer, algorithm layer, application layer and presentation layer, a closed-loop control chain is constructed, from data perception, water level prediction, dynamic draft calculation, safety margin assessment to optimization strategy generation and distribution, implementation feedback. It can achieve system-level docking with existing inland waterway navigation management platforms, bridge area monitoring and early warning systems, etc., and provide integrated technical support for draft safety supervision under new transportation modes such as "river-sea direct transport" and "sea-inland intermodal transport". Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only preferred embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 A traditional diagram illustrating ship draft; Figure 2 This is a schematic diagram of the dynamic draft of a vessel, provided by the SARIMA-based method for optimizing and controlling the draft of vessels exceeding the draft limit for inland waterway navigation according to the present invention. Figure 3 This is a flowchart of the ship navigation draft optimization management strategy of the SARIMA-based method for optimizing and controlling the draft of over-draft vessels in inland waterways, as described in this invention. Figure 4 The flowchart shows the draft optimization and control method for inland waterway navigation of vessels with excessive draft based on SARIMA prediction according to the present invention. Figure 5 This is a diagram illustrating the overall architecture of the electronic device for optimizing and controlling the draft of vessels exceeding draft limits in inland waterway navigation based on SARIMA prediction, as described in this invention. Detailed Implementation

[0035] To better understand the technical content of this invention, a specific embodiment is provided below, and the invention will be further described in conjunction with the accompanying drawings.

[0036] Reference Figure 1 The traditional ship draft diagram shown only includes three basic quantities: Water depth: The geometric depth of water from the surface to the riverbed.

[0037] Ship draft: The vertical distance of the ship's hull submerged in water, generally referring to still water draft or simple average draft.

[0038] The clearance between the bottom of the ship and the riverbed is usually specified according to a fixed proportion or fixed value, which is not less than 10% of the draft or not less than 1.0m.

[0039] Traditional ship draft calculation: Ship draft = water depth - ship's margin water depth.

[0040] In traditional ship draft management, the ship's margin of water depth is mostly a fixed value based on experience, and it is rarely subdivided into factors such as sinking amount, attitude, and riverbed uncertainty, and it does not change in real time with the current speed, draft, and loading conditions.

[0041] Reference Figure 2 The schematic diagram of the ship's dynamic draft shown in this invention decomposes the various dynamic factors affecting the ship's actual draft and safety margin, based on the traditional ship draft method. These factors mainly include: Water depth: The vertical distance from the water surface to the riverbed, which is the water depth condition for maintaining a waterway.

[0042] Predicted water level: Based on historical water level statistics, the SARIMA model is used to estimate the water level height over a certain period of time in the future.

[0043] Dynamic draft of a ship: The actual draft in real time under the combined effects of its current loading status, speed, attitude, etc.

[0044] Dynamic safety margin for ships: DUKC as defined in this invention refers to the dynamic safety margin that should be maintained between the bottom of the ship and the riverbed after considering various uncertainties, which is further decomposed into several refined items.

[0045] Ship fuel and fresh water consumption: As the ship's weight decreases over time, its draft will also decrease, which will affect the draft.

[0046] Ship attitude changes: changes in attitude such as trim and list cause inconsistencies in draft between the bow and stern.

[0047] Ship sinking: The amount of water a ship sinks under the influence of its speed. The amount of sinking is more significant when the ship is in shallow water or traveling at high speed.

[0048] Riverbed siltation allowance: Due to the undulation of the riverbed and the uncertainty of elevation caused by siltation, an additional margin needs to be reserved.

[0049] The proposed ship draft calculation method in this invention is: Dynamic draft = Water depth + Predicted water level - Dynamic surplus water. Its core feature is that, based on factors such as water level prediction, fuel consumption, attitude changes, sinking, and riverbed siltation, a real-time updatable model of dynamic draft and dynamic surplus water depth is constructed. Draft and safety margin are made into "real-time calculable" dynamic quantities and linked with water level prediction results. This allows for more efficient use of channel water depth and water level windows while ensuring navigation safety, thereby achieving refined draft control and optimized management of oversized vessels.

[0050] See Figures 3 to 4 The present invention provides a method for optimizing and controlling the draft of vessels exceeding the draft limit for inland waterway navigation based on SARIMA prediction, comprising the following steps: Step S1: Obtain hydrological and meteorological data, channel condition data, and navigation data of vessels exceeding draft limits for the target inland waterway section, and construct a multi-source time series dataset for inland waterway navigation of vessels exceeding draft limits. The hydrological and meteorological data includes river level, flow rate, water depth, flow velocity, wind speed, and wind direction. The channel condition data includes channel maintenance depth, annual riverbed siltation, bridge area control depth, and navigation structure dimensions. The navigation data of vessels exceeding draft limits includes vessel load, ballast water configuration, bow draft, stern draft and / or midship draft, trim and heel parameters, speed, heading, track, and vessel size parameters.

[0051] Step S2: Perform denoising, error correction, and standardization preprocessing on the multi-source time series dataset to obtain the input time series set. The specific preprocessing steps include: Denoising is performed on hydrological and meteorological data, waterway condition data, and navigation data of vessels with excessive draft, removing meaningless records, outliers, and duplicate data. Error correction processing is performed on data containing symbol errors, encoding errors, and reversed time sequences; Standardize the format of various types of data by unifying units and decimal places; Add a timestamp and a spatial stamp to each record, and use the timestamp + spatial stamp as a composite index to identify duplicate records and delete redundant records; By interpolating and filling in missing data using adjacent time points or adjacent observations, a time-continuous and data-complete sequence can be obtained. Normalize or standardize the key variables to construct an input time series set with a unified time axis.

[0052] After preprocessing, descriptive information is added to hydrological and meteorological data, channel condition data, and navigation data of vessels exceeding draft limits. Specifically, a timestamp and a spatial stamp are added to each water level record. The timestamp is used to accurately indicate the specific time when the water level data was observed, such as 2026-03-13 08:00:00; the spatial stamp is used to indicate the location of the data, marking the water level station number and latitude and longitude coordinates. By uniformly adding time and spatial labels to the water level data, the information of "when, where, and how much" is completely linked, which facilitates time series analysis, spatial distribution analysis, and multi-site comparison.

[0053] In hydrological and meteorological data, waterway condition data, and navigation datasets of vessels with excessive draft, timestamps and spatial stamps are used as a combined index to identify duplicate records at the same time and place. If multiple data entries have the exact same timestamp and spatial stamp, they are considered duplicates in the same observation record. The redundant records need to be deleted, and only one record needs to be kept, thereby avoiding statistical bias and distortion of analysis results caused by duplicate data.

[0054] For meteorological data, waterway condition data, and navigation data of vessels with excessive draft, missing values ​​are estimated by using existing adjacent time points or adjacent observations. When some times are missing, interpolation between adjacent times is used to approximate the missing values ​​with known observations before and after, thus obtaining a time series set that is continuous in time and complete in data.

[0055] Step S3: Construct a SARIMA model for predicting water levels in the target inland waterway section. Based on the input time series set, obtain the water level prediction sequence in the future time domain. The steps for constructing the SARIMA model are as follows: Taking the water level stations along the route in the ship's navigation plan as the research object, the monthly water level statistics (maximum, minimum and average values) of the past three years were collected. In order to ensure the draft safety margin, the monthly minimum water level sequence was used as the modeling object. Before prediction, obvious outliers and missing values ​​were cross-checked and corrected. Duplicate records were removed by using the water level station name + timestamp as an index, and the water level unit and time scale were unified to form a water level time series sorted by month. Seasonal-Trend decomposition using Loess (STL) was used to decompose the water level time series into seasonal components, trend components, and residual components. The decomposition results show that: 1) the overall trend of inland river water level changes is not significant, and the long-term upward or downward trend is not obvious; 2) the seasonal component of inland river water level fluctuates in an annual cycle, but the periodicity is not strong; 3) the residual of inland river water level is randomly distributed and there is no obvious structural pattern.

[0056] The stationarity of water level time series is determined by the Augmented Dickey-Fuller Test (ADF). Water level station series that pass the stationarity test can be used directly. For non-stationary series, first-order constant difference is performed, followed by the ADF test. If the series is still not stationary, seasonal difference is performed in combination with seasonal characteristics until the unit root hypothesis is rejected at the 1%, 5%, and 10% significance levels, at which point the series is considered stationary. Based on the obtained stationary series, the autocorrelation function (ACF) plot and the partial autocorrelation function (PACF) plot are plotted to identify the model order. The autoregression order p is determined based on the significant truncation position of PACF, and the moving average order q is determined based on the significant truncation position of ACF. The inland river water level approximately shows the characteristic that the water level in the same month of each year is similar. The seasonal cycle is taken as s=12. Combining the seasonal cycle s, the difference order d, and the seasonal difference order D, and in conjunction with the ADF test results, the SARIMA model parameters are determined. In the SARIMA model, flow-related exogenous variables are introduced. The monthly average flow and its changes over the past few years are used as exogenous inputs. The direction and intensity of the influence of flow on water level prediction are characterized by regression coefficients. Based on the AIC / BIC and residual tests, different order combinations were compared, and the SARIMA model with the best fit and generalization performance was selected as the target water level prediction model.

[0057] The exogenous variables related to flow include the average flow of this month, the difference between the average flow of this month and the average flow of the previous month, and the difference between the average flow of this month and the average flow of the previous two months. The regression coefficients of the exogenous variables are used to reflect the positive or negative impact and the intensity of the impact of the corresponding flow characteristics on the corrected water level.

[0058] The seasonal difference autoregressive moving average model SARIMA(p,d,q) (P,D,Q) can be written as:

[0059] Where B is the delay operator, similar to a time pointer. Multiplying the current sequence value by a delay operator is equivalent to moving the time of the current sequence value back one moment, expressed as: Especially, Indicates to Perform the first-order difference, and so on. Indicates to Perform d-order difference, This is a seasonal delay operator, equivalent to turning back the time of the current sequence value by one period, denoted as: Especially, Indicates to A seasonal difference with a period of s was performed once, and so on. Indicates to Perform seasonal differences D times with a period of s.

[0060] For the delayed polynomial operator: ; For the delayed polynomial operator:

[0061] For seasonal delay polynomial operators: ; For seasonal delay polynomial operators:

[0062] The SARIMA model is essentially a method for analyzing a time series with seasonal characteristics. Performing D-fold seasonal differencing (de-periodization) and d-fold constant differencing (de-trendization) yields a new series. , represented as:

[0063] After difference Establish the following model:

[0064] Considering the hydrological characteristics of inland rivers, the flow rate has a direct impact on the water level. Therefore, the flow rate is incorporated into the SARIMA model:

[0065] in To correct the water level, Let be the k-th traffic-related exogenous variable in month t. Specifically, This indicates the average flow rate for this month; This indicates the change in the average flow rate this month compared to the average flow rate of the previous month. This indicates the change in the average flow rate this month compared to the average flow rate of the previous two months. .

[0066] For the k-th exogenous variable The corresponding regression coefficients reflect the influence of the flow variable on the predicted object. The direction and intensity of the influence. When hour, and Positively correlated Enlargement will It is showing an upward trend; when hour, and Negative correlation Enlargement will It shows a downward trend; Size is used to measure right The importance of.

[0067] The water level time series is segmented chronologically. The historical data of the previous two years (24 consecutive months) are used as the boundary with the month corresponding to the current time predicted by the model as the training set for model parameter estimation. The historical data of the third year before the current time is used as the model validation set. The data of the most recent 12 months of the current time are used as the test set to evaluate the generalization prediction performance of the model in the future. The water level data of each water level station along the line at the current time t and the water level data of the previous two years (24 months) and their difference values ​​are input into the pre-trained SARIMA model to obtain the water level prediction values ​​in the future time domain t+1,…,t+h.

[0068] Step S4: Decompose the dynamic surplus water depth into several measurable sub-items. Construct a dynamic draft model for the ship based on the water level prediction sequence and the dynamic surplus water depth. Predict the dynamic draft prediction sequence of the ship in the time domain using the dynamic draft model. The calculation formula for the dynamic draft model is as follows:

[0069] in For the dynamic draft of the ship, This is the channel depth, obtained from electronic charts or channel surveys. This value is greater than or equal to the channel maintenance depth published by the navigation authority. This is a predicted water level value, obtained through the SARIMA model. The dynamic margin of water depth for a ship is decomposed into several measurable sub-items, and the specific calculation formula is as follows:

[0070] in This refers to the change in draft caused by the consumption of ship fuel oil and fresh water. This refers to the change in draft caused by changes in the ship's attitude. This refers to the change in draft caused by the ship's sinking. Prepare silt for the riverbed.

[0071] Preferably, the change in draft caused by the consumption of ship fuel and fresh water is calculated by using the ship's speed and draft records at the current and historical times to determine the rate of change in draft corresponding to fuel consumption per unit time and per unit speed, and combined with the planned future speed. Record the ship's speed at the current time (t). and draft and Before the current moment ( The ship's speed and draft at (time) are respectively and draft and The changes in fuel consumption and draft per unit time and per unit speed of the ship are obtained as follows:

[0072] future( The planned speed at (time) is Then predict the future ( Change in draft caused by ship fuel consumption from time t to the present: .

[0073] The change in draft caused by the change in the ship's attitude is calculated by the difference between the heel and pitch angles at the current time and the predicted future time. Record the ship's heel and trim angles at the current time (t). , , assuming the future ( The ship's heel and trim angles at (time) are respectively , Then the future ( The change in draft caused by the change in ship attitude between time t and the current time (t) is: .

[0074] The change in draft caused by the ship's sinking is calculated based on the current water depth-to-draft ratio, the ship's block coefficient, and the planned future speed, using the empirical formula for dynamic water sinking at shallow water speeds. The current (at time t) water depth to draft ratio is ; For the ship's block coefficient; future ( The planned speed at (time) is Then predict the future ( (Time) Ship sinking amount: .

[0075] The riverbed siltation reserve is calculated by converting the annual scouring and silting thickness of the riverbed obtained from the survey of the waterway management department to the predicted time, thus obtaining the riverbed siltation reserve thickness at the future time.

[0076] The survey of the waterway management department yielded the annual sedimentation and erosion thickness of the riverbed. Then predict the future ( (Time) Riverbed siltation capacity .

[0077] The present invention proposes a dynamic draft for ships, which changes the traditional empirical approach of specifying the maximum ship draft according to a fixed proportion or value. Instead, it uses ship speed, draft, attitude, block coefficient, etc., to dynamically calculate the future draft. Dynamic draft of ships at any time : .

[0078] By utilizing historical and current speed and draft records of the vessel, the rate of change in fuel consumption draft is calculated. The changes in heel and roll angles are used to estimate the changes in draft caused by attitude. The dynamic water subsidence is estimated based on the water depth-to-draft ratio, the vessel's block coefficient, and the planned speed. Combined with the annual sedimentation thickness of the riverbed, the amount of sediment prepared in the riverbed is calculated. This enables real-time updating and rolling prediction of the vessel's dynamic draft and dynamic surplus water depth, resulting in a dynamic draft prediction sequence for the vessel divided by time and segment.

[0079] Step S5: Define the target inland waterway segment at a certain predicted time ( The comprehensive minimum control limit is calculated at (time) and a draft safety margin determination operator is constructed by combining it with the predicted dynamic draft of the ship at the corresponding time. Based on the draft safety margin determination operator, the ship's draft safety margin assessment result and risk classification result are obtained. The formula for calculating the comprehensive minimum control limit is as follows:

[0080] in For the target inland waterway segment at the predicted time The comprehensive minimum control limit, The maximum permissible draft of the ship structure is determined based on the ship design documents and load line certificate. The maximum permissible draft for a navigable channel is determined based on the channel's maintenance depth, the classification of navigable waters, and the restrictions issued by the waterway management department. To restrict the draft in the bridge area, the water depth and clearance height are determined based on the controlled water depth and clearance height in the bridge area. The draft is set by the management department to improve the safety of ship navigation in order to control the safety of the vessel. The expression for the draft safety margin determination operator is:

[0081] in For the predicted time The draft safety margin determination value, For the predicted time The predicted dynamic draft of the ship, when the draft safety margin determination operator satisfies At that time, it was determined that the vessel was at risk of exceeding its draft limit at the predicted time; when At that time, it is determined that the system is in a warning state with insufficient safety margin, among which The preset warning threshold is set to 0.1m in this invention.

[0082] The minimum draft safety margin determination value is obtained based on the draft safety margin determination operator within a given time period. The minimum draft safety margin determination value is compared with a preset threshold to obtain a risk classification result including safe, concerned, warning and high risk.

[0083] To facilitate management and dispatching personnel in prioritizing the risks of different vessels and time periods, this invention further constructs a comprehensive risk index for a given time period. The minimum safety margin determination operator for a ship during this period is defined as follows:

[0084] by Draft size is used as the basis for ranking water risk: the smaller the value, the higher the risk level. Based on this, the risk is divided into "safe" and "neutral". ),focus on( ), early warning ( ), high risk ( The system is classified into three levels, which form the risk assessment results for exceeding the draft limit and are pushed to the application layer and the presentation layer to guide navigation decisions.

[0085] Step S6: Construct a draft optimization management model. This model, based on the ship's dynamic draft prediction sequence and the ship's draft safety margin assessment results, calculates the draft optimization management strategy and distributes it to the ship and shore-based dispatching system. The specific steps are as follows: The decision variables are cargo hold loading capacity, ballast water volume in each ballast tank, ship attitude control, ship speed in different river sections, and the time window for passing through restricted river sections. For cargo hold loading capacity, For cargo hold number n, The ballast water volume of each ballast tank, The ballast water volume of ballast tank number m. These are the pitch control parameters. This represents the speed vector of ships in different river sections.

[0086] The constraints include a draft safety constraint that the ship's dynamic draft does not exceed the comprehensive minimum control limit at each predicted time, and a capacity and loading constraint that prevents the capacity of each cargo hold and ballast tank from being exceeded. The draft safety constraint is as follows: For each prediction time The requirement is that the dynamically predicted draft should not exceed the comprehensive control limit, that is:

[0087] in The combined calculations using the SARIMA prediction model and the dynamic draft model demonstrate the impact of speed, loading, and ballast schemes on future draft.

[0088] Cargo capacity and loading constraints: ,

[0089] in , These are the capacity constraints for each cargo hold and ballast tank.

[0090] To maximize effective cargo capacity while meeting draft safety constraints and considering fuel consumption costs, an objective function is constructed, the expression of which is:

[0091] This is a trade-off factor used to balance loading revenue with operating costs. For fuel consumption estimation functions related to speed configuration, This refers to the total amount of ballast water.

[0092] When the relationship between draft and decision variables can be approximated linearly or after linearization, mixed-integer linear programming can be used to solve it; when the relationship is highly nonlinear, heuristic algorithms such as genetic algorithms and particle swarm optimization algorithms can be used to search for approximate optimal solutions and obtain draft optimization management strategies, which can then be distributed to the ship and shore-based scheduling system.

[0093] Historical data related to the target vessel and its navigation environment is collected. This data serves as the foundational input for subsequent water level prediction and dynamic draft calculation. Constraints and control requirements related to future navigation are obtained. This data serves as the boundary conditions and constraints for prediction calculation and safety assessment, inputting into the system to perform "prediction-calculation-assessment-optimization". Under the combined effect of historical data and future conditions, water level prediction and dynamic draft calculation are performed to obtain the dynamic draft of the target vessel at a predetermined time. The obtained dynamic draft calculation results are compared with the maximum permissible draft and the safe control draft value to calculate the draft safety margin. Based on the draft safety margin assessment results, corresponding draft optimization management strategies are generated.

[0094] Alternatively, the above optimization process can be embedded in a rolling time-domain control framework: within each decision cycle, the draft for several future time steps is predicted using the latest hydrological, channel, and vessel status data, and the optimization problem is solved within a finite prediction time domain to obtain the partial control quantities that should be implemented at the current moment (such as the current ballast, the recommended speed for the next river segment, and the navigation attitude). Then, one time step is advanced and the above process is repeated to achieve dynamic updates of the draft optimization management strategy.

[0095] When the optimization results show that the draft safety margin has significantly increased within a future time window while still meeting the navigation schedule requirements, this embodiment can automatically recommend strategies such as "increasing load," "reducing speed," and "utilizing water level"; when it is expected that the river section ahead will appear at the planned passage time... When the speed approaches zero, priority is given to generating a combination of measures, including reducing ballast, decreasing speed, adjusting trim, or changing the passage time window.

[0096] After obtaining the draft optimization management strategy, the strategy is distributed to the vessel or scheduling / management system to guide actual navigation and loading operations. During the implementation of the strategy, the actual draft data and related operating condition data of the vessel are collected in real time or periodically. The effectiveness of the strategy implementation and the degree of improvement in draft safety are compared and evaluated, and the evaluation results are fed back to the optimization management module. After obtaining the actual feedback data, the SARIMA model, the vessel dynamic draft model, and the draft optimization management strategy are iteratively corrected and the parameters are updated. This invention can adapt to changes in hydrological conditions, evolution of waterway conditions, and adjustments in vessel operation modes, achieving continuous optimization of draft management. This invention can realize a closed-loop process from "quantitative assessment of draft safety margin" to "optimization management strategy", which not only ensures the navigation safety of vessels with excessive draft during inland waterway navigation, but also improves the utilization efficiency of waterway depth, taking into account both safety and economy.

[0097] Reference Figure 5 As shown, the present invention also provides an electronic device for the optimization and control of draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction. The overall structure adopts a five-layer architecture from bottom to top, including: perception layer, data layer, algorithm layer, application layer and display layer. Each layer interacts with the others through ordered data flow and control flow, realizing integrated closed-loop management from multi-source perception, data processing, intelligent computing to business application and visualization display.

[0098] Perception layer: The sensing layer, located at the bottom of the system, is the data source for the entire system. It mainly consists of sensing terminals such as hydrological and meteorological observation equipment, hydrological and meteorological data sources, electronic buoys, radar, industrial cameras, AIS receiving equipment, and Beidou positioning equipment.

[0099] The collected hydrological and meteorological data mainly include: water level, flow rate, flow velocity, wind speed, wind direction, wave height, etc.

[0100] The collected waterway condition data mainly include: waterway maintenance depth, annual riverbed siltation, and controlled water depth in the bridge area.

[0101] The collected ship operation data mainly includes: ship size, trim, list, speed, ballast water configuration, and bow / stern draft.

[0102] The perception layer collects the above multi-source information in real time and uploads the raw observation data to the data layer.

[0103] Data layer: The data layer is responsible for data storage and management, and consists of a data acquisition module, a data cleaning and standardization module, a unified timeline management module, a data normalization and preprocessing module, a database, and a data storage server.

[0104] Collect and access multi-source heterogeneous data from the perception layer, complete format conversion, noise reduction, outlier removal and missing value filling, and achieve data cleaning (noise reduction, error correction and standardization).

[0105] By managing the timeline in a unified manner, data with different sampling frequencies and time stamps are time-aligned to build a unified foundation for time-series data.

[0106] Key variables are normalized and standardized, and data is labeled and features are extracted. The processed data is stored in the database server and historical data warehouse according to the theme and business needs, providing reliable data support for algorithm layer calls.

[0107] Algorithm layer: The algorithm layer, deployed on the algorithm server, is the core computing and decision-making unit of the system, integrating multiple model algorithms such as inland water level prediction, water level window analysis, and sorting and control of vessels with excessive draft.

[0108] The predicted water level sequence is obtained by using an inland river water level forecasting module based on the SARIMA time series model, combined with historical water level, flow rate and meteorological data.

[0109] Based on the predicted water level and the reference water depth of the waterway, an algorithm for utilizing water level windows is constructed to identify water level windows suitable for navigation of vessels with excessive draft in different time periods.

[0110] By combining information such as ship dimensions, actual draft, navigation requirements, and arrival order, an adaptive ship sorting and draft optimization algorithm is constructed, providing a comprehensive optimization scheme for ship berthing order, recommended draft, and navigation timing. The algorithm layer calculates key indicators such as ship safety draft control value, draft correction amount, and excess water depth utilization rate according to business needs, and transmits the calculation results and intermediate state data to the application layer.

[0111] Application layer: The application layer, based on an integrated visualization platform, encapsulates the models and calculation results produced by the algorithm layer into business function modules, supporting refined management of vessels exceeding draft limits. This mainly includes: Inland water level forecasting module: Enables querying, comparing, and configuring early warnings for water level forecasts from multiple stations.

[0112] The water level navigation management module is used to assist in scheduling ships to pass through restricted waters during the optimal time period based on the water level window period identification results, thereby improving the efficiency of water level utilization.

[0113] Ship draft optimization management module: Based on ship characteristics and predicted water level, it provides recommended control draft values, draft over-limit verification and adjustment schemes to achieve dynamic optimization control of ship draft.

[0114] It can also expand business applications such as ship scheduling optimization and traffic efficiency assessment to achieve comprehensive management and control of waterway and port operations.

[0115] The business results and decision information processed by the application layer will be pushed to the presentation layer for display and interaction on multiple terminals.

[0116] Presentation layer: The presentation layer is the user-facing interactive and visual interface, encompassing various platforms such as PC web pages, handheld mobile terminals, and shore-based large-screen display devices.

[0117] On PC web pages and shore-based display devices, the system provides visualized displays of multi-dimensional data, including water level process curves, flow changes, ship operation status diagrams, draft control suggestions, and navigation windows.

[0118] On handheld mobile terminals, the system pushes key prompts such as ship draft information, forecast water level, recommended navigation time windows, and operation instructions to ship owners, dispatchers, and managers, enabling information access and decision support anytime, anywhere.

[0119] The presentation layer also supports interactive functions such as early warning information pop-ups, work instruction issuance and confirmation, forming a closed-loop control chain from prediction calculation to on-site execution.

[0120] Through the coordinated operation of the above five-layer architecture, this invention realizes the perception, intelligent analysis and optimized management of the entire process of inland waterway navigation for vessels with overdrafts, thereby improving water level utilization and waterway throughput capacity while ensuring navigation safety.

[0121] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction, characterized in that, Includes the following steps: Step S1: Obtain hydrological and meteorological data, channel condition data, and navigation data of vessels with excessive draft for the target inland waterway section, and construct a multi-source time series dataset of inland waterway navigation for vessels with excessive draft. Step S2: Preprocess the multi-source time series dataset to obtain the input time series set; Step S3: Construct a SARIMA model for predicting water levels in the target inland waterway section, and obtain the water level prediction sequence in the future time domain based on the input time series set; Step S4: Decompose the dynamic surplus water depth into several measurable sub-items, construct a ship dynamic draft model based on the water level prediction sequence and the dynamic surplus water depth, and predict the ship dynamic draft prediction sequence in the time domain using the ship dynamic draft model. Step S5: Define the comprehensive minimum control limit of the target inland waterway section at a certain predicted time, and construct a draft safety margin determination operator with the predicted dynamic draft of the ship at the corresponding time. Based on the full draft margin determination operator, obtain the ship draft safety margin assessment result and risk classification result. Step S6: Construct a draft optimization management model. The draft optimization management model is based on the ship's dynamic draft prediction sequence and the ship's draft safety margin assessment results to obtain the draft optimization management strategy, and the draft optimization management strategy is distributed to the ship and shore-based scheduling system.

2. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 1, characterized in that, The hydrological and meteorological data include river level, flow rate, water depth, flow velocity, wind speed, and wind direction; the waterway condition data include waterway maintenance depth, annual riverbed siltation, bridge area control depth, and dimensions of navigation structures; the navigation data for vessels with overdrafts include vessel load, ballast water configuration, bow draft, stern draft and / or midship draft, trim and heel parameters, speed, heading, track, and vessel size parameters.

3. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 1, characterized in that, The specific steps of step S2 include: Denoising is performed on hydrological and meteorological data, waterway condition data, and navigation data of vessels with excessive draft, removing meaningless records, outliers, and duplicate data. Error correction processing is performed on data containing symbol errors, encoding errors, and reversed time sequences; Standardize the format of various types of data by unifying units and decimal places; Add a timestamp and a spatial stamp to each record, and use the timestamp + spatial stamp as a composite index to identify duplicate records and delete redundant records; By interpolating and filling in missing data using adjacent time points or adjacent observations, a time-continuous and data-complete sequence can be obtained. Normalize or standardize the key variables to construct an input time series set with a unified time axis.

4. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 1, characterized in that, The steps for constructing the SARIMA model in step S3 are as follows: The monthly minimum water level sequence of water level stations along the ship's navigation route is used as the modeling object. Outliers and duplicate records are removed, and the water level unit and time scale are unified to form a water level time series sorted by month. STL decomposition was used to perform seasonal trend decomposition on the water level time series, yielding seasonal components, trend components, and residual components. The stationarity of water level time series is determined by using the ADF test. For non-stationary series, first-order constant difference and necessary seasonal difference are performed until the stationarity test is passed at the preset significance level. Based on the obtained stationary series, the autocorrelation function (ACF) plot and the partial autocorrelation function (PACF) plot are plotted. The autoregression order p is determined according to the significant truncation position of PACF, and the moving average order q is determined according to the significant truncation position of ACF. Combined with the seasonal period s and the seasonal difference order D, the SARIMA model parameters are determined. In the SARIMA model, flow-related exogenous variables are introduced. The monthly average flow and its changes over the past few years are used as exogenous inputs. The direction and intensity of the influence of flow on water level prediction are characterized by regression coefficients. Based on the AIC / BIC and residual tests, different order combinations were compared, and the SARIMA model with the best fit and generalization performance was selected as the target water level prediction model.

5. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 4, characterized in that, The exogenous variables related to flow include the average flow of this month, the difference between the average flow of this month and the average flow of the previous month, and the difference between the average flow of this month and the average flow of the previous two months. The regression coefficients of the exogenous variables are used to reflect the positive or negative impact and the intensity of the impact of the corresponding flow characteristics on the corrected water level.

6. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 1, characterized in that, The calculation formula for the ship's dynamic draft model is as follows: in For the dynamic draft of the ship, Due to the depth of the waterway, This is the predicted water level. The dynamic margin of water depth for a ship is decomposed into several measurable sub-items, and the specific calculation formula is as follows: in This refers to the change in draft caused by the consumption of ship fuel oil and fresh water. This refers to the change in draft caused by changes in the ship's attitude. This refers to the change in draft caused by the ship's sinking. Prepare silt for the riverbed.

7. The method for optimizing and controlling the draft of vessels exceeding draft limits in inland waterway navigation based on SARIMA prediction according to claim 6, characterized in that, The change in draft caused by the consumption of ship fuel and fresh water is calculated by taking the ship's speed and draft records at the current and historical times as the rate of change in draft corresponding to fuel consumption per unit time and per unit speed, and combined with the planned speed for the future. The change in draft caused by the change in the ship's attitude is calculated by the difference between the heel and pitch angles at the current time and the predicted future time. The change in draft caused by the ship's sinking is calculated based on the current water depth-to-draft ratio, the ship's block coefficient, and the planned future speed, using the empirical formula for dynamic water sinking at shallow water speeds. The riverbed siltation reserve is calculated by converting the annual scouring and silting thickness of the riverbed obtained from the survey of the waterway management department to the predicted time, thus obtaining the riverbed siltation reserve thickness at the future time.

8. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 1, characterized in that, The formula for calculating the comprehensive minimum control limit is as follows: in For the target inland waterway segment at the predicted time The comprehensive minimum control limit, For the maximum permissible draft of the ship structure, The maximum permissible draft for navigable waterways, To restrict the draft in the bridge area, To ensure safe control of draft; The expression for the draft safety margin determination operator is: in For the predicted time The draft safety margin determination value, For the predicted time Predicted dynamic draft of ships; The minimum draft safety margin determination value is obtained based on the draft safety margin determination operator within a given time period. The minimum draft safety margin determination value is compared with a preset threshold to obtain a risk classification result including safe, concerned, warning and high risk.

9. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 1, characterized in that, The specific steps of step S6 are as follows: The decision variables are cargo hold loading capacity, ballast water volume in each ballast tank, ship attitude control, ship speed in different river sections, and time window for passing through restricted river sections. The constraints include a draft safety constraint that the ship's dynamic draft does not exceed the comprehensive minimum control limit at each predicted time, and a capacity and loading constraint that the capacity of each cargo hold and ballast tank is not exceeded. To maximize the effective cargo capacity while also considering fuel consumption costs under the premise of meeting draft safety constraints, an objective function is constructed. The optimization problem is solved using mixed-integer linear programming or heuristic algorithms, and the draft optimization management strategy is obtained. The draft optimization management strategy is then distributed to the ship and shore-based scheduling system.

10. The method for optimizing and controlling the draft of vessels exceeding the draft limit in inland waterway navigation based on SARIMA prediction according to claim 9, characterized in that, The expression for the objective function is: in For the ship speed vector of each river segment, For cargo hold loading capacity, For cargo hold number n, The ballast water volume of each ballast tank, The ballast water volume of ballast tank number m. These are the pitch control parameters. As a weighting factor, For fuel consumption estimation functions related to speed configuration, This refers to the total amount of ballast water.

Images