CN-122022306-A - Ship fuel supply data monitoring method and system based on sea area environment

CN122022306ACN 122022306 ACN122022306 ACN 122022306ACN-122022306-A

Abstract

The application relates to the field of ship fuel, and discloses a ship fuel supply data monitoring method and system based on a sea area environment, wherein the method comprises the steps of acquiring real-time navigation data, weather forecast data and ship log data of a ship; the method comprises the steps of obtaining corresponding energy consumption modes and predicted energy consumption trends of first time duration in the future based on real-time navigation data, weather forecast data, ship log data and a trained multi-task learning model, calculating fuel consumption time and minimum replenishment demand based on current fuel stock, residual voyage and the predicted energy consumption trends when the high energy consumption modes are identified, screening candidate replenishment places according to the replenishment demand, the fuel consumption time and a preset route range, prioritizing the candidate replenishment places, and generating a replenishment scheme based on the replenishment place with the highest priority in the prioritization. Through the technical scheme, navigation risks caused by insufficient supply are effectively avoided.

Inventors

LI LI
YAN CONG
Zhao kaixuan
XIONG JIAJI
XIONG QIANG

Assignees

湖北弘仪智能装备有限责任公司

Dates

Publication Date: 20260512
Application Date: 20260126

Claims (8)

1. A marine vessel fueling data monitoring method based on a marine environment, comprising: Acquiring real-time navigation data, weather forecast data and ship log data of a ship; obtaining a corresponding energy consumption mode and a predicted energy consumption trend of a first time length in the future based on the real-time navigation data, the weather forecast data, the ship log data and the trained multi-task learning model, wherein the energy consumption mode comprises a high energy consumption mode and a normal energy consumption mode; When the high energy consumption mode is identified, calculating a fuel consumption time and a minimum replenishment demand based on the current fuel inventory, the remaining range, and the projected energy consumption trend; Screening candidate replenishment sites according to the replenishment demand, the fuel consumption time and a preset route range; The method comprises the steps of obtaining service availability data, fuel price data, port condition data and predicted stop time of each candidate replenishment place, sequencing the priority of the candidate replenishment places on the premise of meeting fuel constraint and time constraint on the basis of the service availability data, the fuel price data, the port condition data and the predicted stop time, wherein the fuel constraint is that the fuel quantity after replenishment meets the requirement of reaching the next destination, and the time constraint is that the time of reaching the destination cannot be caused by replenishment operation exceeds the maximum allowable delay time, wherein the maximum allowable delay time is obtained according to a ship operation plan; A replenishment plan including a recommended route, an estimated time of arrival, and a required replenishment amount is generated based on the replenishment location of the highest priority in the prioritization.
2. The method of claim 1, wherein the multi-task learning model comprises a feature extraction layer, a dual-task sub-model layer, and a fusion layer; the feature extraction layer is configured to extract general features of ship navigation, and comprises a data preprocessing module, a feature engineering module and a feature extraction network, wherein the data preprocessing module performs normalization processing on input data, the feature engineering module constructs ship navigation feature vectors, and the feature extraction network adopts a convolutional neural network architecture and is used for extracting high-level abstract features from the ship navigation feature vectors; the energy consumption prediction sub-model adopts a mixed structure of a time domain convolution network and a bidirectional LSTM (least squares) for predicting the future energy consumption trend of the ship, and the energy consumption pattern recognition sub-model adopts a cascade structure of an improved clustering algorithm and a multi-head self-attention classifier for recognizing the current energy consumption pattern of the ship; The fusion layer is used for synthesizing the output result of the double-task sub-model to generate a comprehensive energy consumption analysis result.
3. The method of claim 2, wherein the time domain convolution network part of the energy consumption prediction sub-model adopts a multi-layer cascade structure, each layer adopts a causal convolution mode, expansion factors of each layer are arranged in an increasing mode, a first time domain convolution network layer adopts a first preset expansion factor for capturing short-term energy consumption fluctuation characteristics, an intermediate time domain convolution network layer adopts a second preset expansion factor for capturing mid-term energy consumption trend characteristics, a top time domain convolution network layer adopts a third preset expansion factor for capturing long-term energy consumption mode characteristics, wherein the first preset expansion factor is smaller than the second preset expansion factor, the second preset expansion factor is smaller than the third preset expansion factor, and batch normalization layers and regularization layers are arranged behind each time domain convolution network layer; The bidirectional LSTM part of the energy consumption predictor model comprises a plurality of bidirectional LSTM layers, a front layer bidirectional LSTM returns to complete sequence output for retaining time-period change details of energy consumption data, a rear layer bidirectional LSTM returns to a final output vector for analyzing the energy consumption trend of the whole sailing section, when the sea state level is detected to be changed remarkably, the forgetting door weight of the LSTM is adjusted to adapt to the influence of environmental change on energy consumption, and the output layer adopts a full-connection layer structure and outputs an energy consumption trend prediction sequence of a first time length in the future.
4. The method of claim 3, wherein the energy consumption pattern recognition sub-model comprises a clustering layer, a feature transformation layer, and a classification layer; the clustering layer adopts an improved K-means algorithm to determine the optimal clustering quantity; The feature conversion layer adopts a multi-head self-attention mechanism, wherein different numbers of attention heads are set to respectively process numerical value type features, category type features and time sequence features, different feature mapping strategies are adopted for each attention head, linear transformation is carried out on input features to generate a query matrix, a key matrix and a value matrix, contribution degrees of different features to an energy consumption mode are calculated based on the attention mechanism, attention weights are determined based on correlation among the features, weighted summation is carried out on the value matrix according to the attention weights to generate attention output, and the output of the plurality of attention heads is spliced and linear transformation is carried out to generate final feature representation; the classification layer adopts a support vector machine algorithm to output probability distribution of various energy consumption modes.
5. The method of claim 4, wherein screening candidate replenishment places based on the replenishment demand, fuel depletion time and a preset course range comprises: constructing a geographic feasible region in a banded region with the current ship position as a starting point and the furthest navigable distance corresponding to the fuel consumption time as a radius and with width thresholds set on two sides of a preset route; selecting a port or an offshore filling point with ship refueling capability as an initial candidate set in the geographic feasible region; calculating the fuel consumption required for voyage from the current ship position to each place in the initial candidate set, and eliminating places which are more than the current fuel stock and have the time to reach the place which is later than the fuel consumption time minus the safe buffer duration and the lateral offset distance from the preset route which exceeds the allowable maximum yaw distance; the remaining sites are retained as final candidate replenishment sites.
6. The method of claim 5, wherein when the high energy consumption mode is identified, calculating the fuel consumption time and minimum replenishment demand based on the current fuel inventory, the remaining voyage, and the projected energy consumption trend comprises: Extracting energy consumption characteristic parameters under a historical high energy consumption scene from a historical database based on the confirmed high energy consumption mode; decomposing the predicted energy consumption trend into fuel consumption predicted values of a plurality of time slices; accumulating the corrected time-period fuel consumption predicted values in time sequence by taking the current fuel stock as a reference to generate an accumulated fuel consumption time sequence and determining fuel consumption time; Calculating the remaining range from the current position of the ship to the next destination; determining an average fuel consumption level in a high energy consumption mode based on the energy consumption characteristic parameters, and calculating the minimum fuel quantity required by completing the voyage according to the average fuel consumption level and the residual voyage; A minimum fueling demand is calculated based on the minimum fuel amount, the safe reserve fuel amount, and the current actual fuel inventory.
7. The method of claim 6, wherein prioritizing the candidate replenishment places based on the service availability data, fuel price data, port condition data, and predicted dock times on the premise that fuel constraints and time constraints are satisfied, comprises: respectively acquiring service availability data, fuel price data, port condition data and predicted stop time for each candidate replenishment site; Converting the service availability data, the fuel price data, the port condition data and the estimated berthing time into corresponding scoring items respectively; Distributing corresponding weights for each scoring item according to the current operation targets of the ship, wherein the operation targets comprise cost priority, time priority or service reliability priority; calculating a comprehensive score of each candidate replenishment site based on each scoring item and the corresponding weight thereof; And prioritizing the reserved candidate sites from high to low according to the composite score.
8. A marine vessel fueling data monitoring system based on a marine environment, comprising: The first processing module is used for acquiring real-time navigation data, weather forecast data and ship log data of the ship; The second processing module is used for obtaining a corresponding energy consumption mode and a predicted energy consumption trend of a first time length in the future based on the real-time navigation data, the weather forecast data, the ship log data and the trained multi-task learning model, wherein the energy consumption mode comprises a high energy consumption mode and a normal energy consumption mode; A third processing module for calculating a fuel consumption time and a minimum replenishment demand based on the current fuel inventory, the remaining voyage, and the predicted energy consumption trend when the high energy consumption mode is identified; The fourth processing module is used for screening candidate replenishment places according to the replenishment demand, the fuel consumption time and the preset route range; The fifth processing module is used for acquiring service availability data, fuel price data, port condition data and expected berthing time of each candidate replenishment place, sequencing the candidate replenishment places according to priority on the premise of meeting fuel constraint and time constraint on the basis of the service availability data, the fuel price data, the port condition data and the expected berthing time, wherein the fuel constraint is that the fuel quantity after replenishment meets the requirement of reaching the next destination, and the time constraint is that the time of reaching the destination cannot be caused by replenishment operation exceeds the maximum allowable delay time; And a sixth processing module for generating a replenishment plan comprising a recommended route, an estimated time of arrival, and a required replenishment amount based on a replenishment place of a highest priority in the prioritization.

Description

Ship fuel supply data monitoring method and system based on sea area environment Technical Field The application relates to the field of ship fuel, in particular to a ship fuel supply data monitoring method and system based on a sea area environment. Background The energy consumption of the ship is affected by sea conditions, stormy waves, ocean currents, loading, host machine working conditions and other factors in the course of sailing, and the ship has high nonlinearity and time-varying characteristics. Under severe sea areas or high-load working conditions, ships are easy to enter a high-energy consumption state, and if the ships cannot timely identify and plan replenishment, fuel can be consumed in advance, so that the period of the ships is seriously influenced. Currently, mainstream marine energy efficiency management systems or electronic chart display and information systems typically estimate the remaining range and fuel demand based on a fixed fuel consumption model or historical average data. Part of the systems introduce meteorological routing technology and optimize the route by combining the information of the predicted stormy waves so as to reduce the energy consumption, but the method cannot early warn in time at the early stage of sudden increase of the energy consumption. How to solve the technical problem is a technical problem that a person skilled in the art needs to overcome. Disclosure of Invention The embodiment of the application provides a ship fuel supply data monitoring method based on a sea area environment, which aims to at least partially solve the technical problems. In order to achieve the above object, according to a first aspect of the present application, there is provided a marine fuel supply data monitoring method based on a marine environment, comprising: Acquiring real-time navigation data, weather forecast data and ship log data of a ship; obtaining a corresponding energy consumption mode and a predicted energy consumption trend of a first time length in the future based on the real-time navigation data, the weather forecast data, the ship log data and the trained multi-task learning model, wherein the energy consumption mode comprises a high energy consumption mode and a normal energy consumption mode; When the high energy consumption mode is identified, calculating a fuel consumption time and a minimum replenishment demand based on the current fuel inventory, the remaining range, and the projected energy consumption trend; Screening candidate replenishment sites according to the replenishment demand, the fuel consumption time and a preset route range; The method comprises the steps of obtaining service availability data, fuel price data, port condition data and predicted stop time of each candidate replenishment place, sequencing the priority of the candidate replenishment places on the premise of meeting fuel constraint and time constraint on the basis of the service availability data, the fuel price data, the port condition data and the predicted stop time, wherein the fuel constraint is that the fuel quantity after replenishment meets the requirement of reaching the next destination, and the time constraint is that the time of reaching the destination cannot be caused by replenishment operation exceeds the maximum allowable delay time, wherein the maximum allowable delay time is obtained according to a ship operation plan; A replenishment plan including a recommended route, an estimated time of arrival, and a required replenishment amount is generated based on the replenishment location of the highest priority in the prioritization. According to a second aspect of the present application there is provided a marine vessel fueling data monitoring system based on a marine environment, comprising: The first processing module is used for acquiring real-time navigation data, weather forecast data and ship log data of the ship; The second processing module is used for obtaining a corresponding energy consumption mode and a predicted energy consumption trend of a first time length in the future based on the real-time navigation data, the weather forecast data, the ship log data and the trained multi-task learning model, wherein the energy consumption mode comprises a high energy consumption mode and a normal energy consumption mode; A third processing module for calculating a fuel consumption time and a minimum replenishment demand based on the current fuel inventory, the remaining voyage, and the predicted energy consumption trend when the high energy consumption mode is identified; The fourth processing module is used for screening candidate replenishment places according to the replenishment demand, the fuel consumption time and the preset route range; The fifth processing module is used for acquiring service availability data, fuel price data, port condition data and expected berthing time of each candidate replenishment place, sequencing the candidate replenishmen