A method for estimating the cruising range and energy consumption of a ship
Through intelligent data analysis module and dynamic environment adaptation algorithm, the ship's navigation parameters are monitored and adjusted in real time, and the simplified model of ship mileage and energy consumption estimate calculation in the existing technology, insufficient data and insufficient real-time adaptability are solved, and the problem of insufficient data is insufficient and real-time adaptability is insufficient. Efficient and economical energy consumption optimization and battery life improvement, enhancing navigation safety and sustainability.
Patent Information
- Application Number
- CN202410399911.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-03
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-04-03
AI Technical Summary
The existing ship mileage and energy consumption estimate calculation methods have simplified models, insufficient data or low quality, lack of real-time adaptability, inability to adapt to different types of ships and navigation conditions, high computing resources, high initial investment, safety and privacy Problems and other issues affect the accuracy and reliability of predictions.
Using intelligent data analysis module to collect and analyze ship historical data, weather conditions and navigation speed through machine learning and big data technology, combined with IoT devices and edge computing technology, uses deep learning algorithms and dynamic environment adaptation algorithms to monitor and adjust navigation speeds in real time , heading and engine output, optimize energy consumption and range.
Significantly improve the economy, efficiency and environmental sustainability of navigation, reduce energy consumption and operational costs, enhance the safety and adaptability of navigation, provide strong decision-making support, and promote the intelligent and digital transformation of the shipping industry. .
Smart Images

Figure CN118313247B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for estimating and calculating mileage and energy consumption, and in particular to a method for estimating and calculating the cruising range and energy consumption of a ship. Background Art
[0002] Although the current ship range and energy consumption estimation calculation methods have made some progress in practical applications, there are still some drawbacks and shortcomings, which may affect the accuracy, practicality and reliability of the prediction. The following are some of the main drawbacks and shortcomings of these methods: First, many existing estimation methods rely on simplified models and assumptions, which may not fully capture the complex dynamic changes during navigation, such as the changing ocean environment, different ship operation modes and changing weather conditions. This simplification may cause the prediction results to deviate from the actual situation, affecting the effectiveness of navigation planning. Secondly, accurate prediction depends on a large amount of high-quality data. However, existing methods may face problems such as insufficient data or low data quality, such as sensor failure, incomplete data collection or data processing errors, which will affect the accuracy of the prediction. Thirdly, traditional prediction methods often lack real-time adaptability and cannot respond quickly to environmental changes and changes in ship status. This means that in the face of emergencies or unexpected ocean conditions, these methods may not provide accurate predictions, thus affecting the safety and efficiency of navigation.
[0003] In addition, different ships and routes may have different characteristics and needs. However, existing prediction methods often lack sufficient generalization ability and cannot adapt to different types of ships and diverse navigation conditions, which limits their widespread application. There are also some high-precision prediction methods, especially those involving complex models and large amounts of data analysis, which may require a lot of computing resources. This may be a problem for ships with limited resources, especially for those small or old ships. Therefore, even if the prediction model itself is powerful, if there is a lack of easy-to-use interfaces and tools, ship operators and managers may find it difficult to effectively use these predictions to guide actual operations. This limits the practical application value of advanced prediction technology. In addition, marine environment and weather conditions have an important impact on ship energy consumption and cruising range. However, many existing methods may not fully consider these factors, or the climate models used are overly simplified, which may lead to inaccurate prediction results. For small ships or small-scale shipping companies, deploying advanced prediction systems may require large initial investment and operating costs. If the cost is too high, it may affect the promotion and application of these methods. In addition, collecting and analyzing a large amount of navigation data may raise security and privacy issues, especially during data transmission and processing. If these issues are not properly handled, they may pose risks to shipping companies and passengers. In addition, the shipping industry and the marine environment are constantly changing, which requires forecasting methods to be able to quickly adapt to new situations and needs. However, many existing methods may lack sufficient adaptability and flexibility to keep up with the rapidly changing external environment. Summary of the invention
[0004] The purpose of the present invention is to provide a method for estimating and calculating the cruising range and energy consumption of a ship, thereby solving some of the drawbacks and shortcomings pointed out in the background technology.
[0005] The present invention solves the above-mentioned technical problems by adopting the following technical solutions: including an intelligent data analysis module for collecting and analyzing historical ship endurance data, weather conditions, and navigation speed information through machine learning and big data technology to establish an energy consumption prediction model; a dynamic environmental adaptation algorithm for adjusting the prediction model in real time according to current marine environmental changes to dynamically estimate the cruising range and energy consumption; and an energy optimization management system, which monitors the ship's energy usage in real time and automatically adjusts the ship's driving speed, heading, and engine output according to the prediction results of the intelligent data analysis module and the dynamic environmental adaptation algorithm to optimize energy consumption and cruising range.
[0006] Furthermore, the intelligent data analysis module includes the following three parts:
[0007] P1. Use IoT devices and sensor networks deployed on ships to collect comprehensive data in real time, including sailing speed, engine status, fuel consumption, and ocean weather conditions; and use edge computing technology to pre-process data to reduce the amount of data transmitted to the central server and improve processing speed and efficiency;
[0008] P2. Use deep learning algorithms to analyze time series data to capture complex patterns and relationships; thereby implementing model optimization strategies to adapt to subtle changes in specific vessels and routes;
[0009] P3. Combine multiple data sources, including ship data, open weather data, ocean current information, and satellite monitoring data, to achieve a comprehensive understanding of the environment; and use data fusion technology to integrate data from different sources and types to provide more comprehensive and accurate energy consumption forecasts.
[0010] Furthermore, the IoT devices and sensor network collect comprehensive data in real time on the ship, including the navigation speed v(t) = a·e bt +c·sin(dt), to simulate the complex conditions faced by actual sailing speed, where a, b, c, and d are adjustment parameters specific to the ship and the environment;
[0011] Engine status:
[0012]
[0013] Used to simulate the nonlinear changes of engine performance under certain conditions, where k and m adjust the steepness and center point of the curve;
[0014] Fuel consumption:
[0015]
[0016] The integral form represents the nonlinear relationship between fuel consumption and operating conditions, where γ and n are adjustment coefficients;
[0017] Ocean weather conditions W(w) = αw 2 +βw+χ, used to simulate the impact of weather conditions on ship operations, where α, β, and χ are adjustment parameters;
[0018] The edge computing technology described pre-processes the collected data, including the preliminary analysis function of the data:
[0019]
[0020] The effects of speed, engine status, fuel consumption and weather conditions are combined; logarithmic and square root functions are used to enhance nonlinear relationships and interactions between data;
[0021] And the data compression conversion function:
[0022]
[0023] A method that uses Taylor series approximation to reduce the amount of data transmitted while maintaining information integrity.
[0024] Furthermore, the deep learning algorithm analyzes time series data:
[0025] T s (t) = A(t)·e B(t) +C(t)·sin(D(t))
[0026] Where A(t), B(t), C(t), and D(t) are time-varying coefficients that can capture the complex patterns and relationships of navigation speed, engine status, fuel consumption, and ocean weather conditions;
[0027] Through the feature engineering and data preprocessing modules, the system automatically extracts features from raw data and standardizes them into a format optimized for deep learning models:
[0028]
[0029] Where μ(T s ) and σ(T s ) are the mean and standard deviation of the time series, respectively; the system further implements model optimization strategies, including transfer learning, which uses model parameters Θ pre-trained on similar tasks pre , by minimizing the loss function:
[0030]
[0031] Where P m The system is a prediction model that can quickly adapt to the specific data of new ships and routes. Combined with reinforcement learning, the system learns the optimal strategy by interacting with the environment and optimizes the objective function:
[0032]
[0033] Where γ is the discount factor, R is the reward function, and α(t) and β(t) are the action and state variables.
[0034] Furthermore, the combination of multiple data sources, including ship data S b (t), open weather data W d (t), ocean current information H i (t), and satellite monitoring data W d (t), where t represents time; using a custom function:
[0035]
[0036] Where α, β, δ, and λ are model parameters to achieve comprehensive environmental understanding;
[0037] Further adopting data fusion techniques, including feature-level fusion:
[0038]
[0039] where F k represents the feature function extracted from each data source, ω k is the weight coefficient, φ is the nonlinear conversion function, and decision-level fusion:
[0040]
[0041] Where ψ is the decision function used to integrate various prediction results, and T is the prediction time range.
[0042] Furthermore, the dynamic environment adaptation algorithm includes an environment perception model:
[0043]
[0044] where α i and β i Adjust the influence weights and nonlinearity of each factor to interpret the current marine environmental data and build an adaptive prediction model that can dynamically adjust its parameters based on the output of the environmental perception model, using reinforcement learning or online machine learning methods to ensure that the model can be updated and adapted to new environmental data in real time;
[0045] The predictions are then compared to the actual observations through a real-time feedback loop, using the differences:
[0046]
[0047] Where P m (t) is the prediction model result, O m (t) is the observed data and the optimized model.
[0048] Furthermore, the energy optimization management system includes an intelligent data analysis module, which uses a machine learning algorithm, especially a deep learning model in is the depth of the hidden layer, σ is the activation function, and is the network parameter;
[0049] For analyzing historical energy usage data Where D(s) represents historical energy data, γ is the attenuation parameter; combined with the real-time ocean Mo (t) = α m cos(β m t) and meteorological W e (t) = α w sin(β w t+φ w ) data, and the current status of the ship S c (t) = k1v(t) 2 +k2h(t) and the planned route, where v(t) and h(t) represent the navigation speed and heading respectively, for comprehensive analysis;
[0050] Provide accurate energy consumption trend forecast where λ p and μ p To adjust the parameters of the prediction results.
[0051] Furthermore, the energy optimization management system also includes a dynamic environment adaptation algorithm, which monitors and analyzes changes in the ocean environment in real time, including the ocean current H(t)=a h sin(b h t+c h ), wind speed W(t) = a w cos(b w t+c w ), and temperature T(t) = a t exp(-b t t 2 )+c t ;
[0052] where a h , b h , c h , a w , b w , c w , a t , b t , and c t is the specific coefficient of the environmental variable;
[0053] The system builds an environmental state model To synthesize the changes, among which θ1, θ2, θ3, δ h , δ w , and δ t is the model parameter; the algorithm further uses a mechanism P to adjust the prediction model in real time adj (t) = P(t)·exp(κ·E s (t)), where P(t) is the initial prediction, P adj (t) is the adjusted forecast, and κ is the environmental impact adjustment coefficient.
[0054] Furthermore, the energy optimization management system also includes an automatic adjustment control system, which automatically calculates and adjusts the optimal driving speed through an optimization algorithm based on the prediction results provided by the intelligent data analysis module and the dynamic environment adaptation algorithm. course and engine output
[0055] in,
[0056]
[0057] are the energy consumption influence functions of driving speed, heading and engine output, α v (t), α h (t), and α e (t) is the time-related adjustment coefficient, and T is the forecast period.
[0058] Beneficial effects of the present invention:
[0059] The present invention significantly improves the economy, efficiency and environmental sustainability of navigation. The following are the main benefits of the method:
[0060] 1. Energy consumption optimization: This method can significantly reduce the energy consumption of ships by accurately predicting and adjusting the sailing speed, heading and engine output. The optimization algorithm ensures that the ship sails in the most energy-efficient way in all sea and weather conditions, thereby reducing fuel consumption and operating costs.
[0061] 2. Improved cruising range: The precise predictions provided by the intelligent data analysis module help ships plan routes and speeds more effectively, allowing them to travel farther with a given amount of fuel. This not only improves navigation efficiency, but also helps ensure the reliability and time accuracy of navigation plans.
[0062] 3. Reduced environmental impact: By reducing fuel consumption, this method helps reduce greenhouse gas emissions and other pollutants, thereby reducing the impact of the shipping industry on the environment. This is essential for addressing global climate change and protecting marine ecosystems.
[0063] 4. Real-time adaptability: The dynamic environmental adaptation algorithm can monitor and analyze changes in the marine environment in real time, and adjust the prediction model in time, so that the ship can quickly adapt to sudden weather changes and marine conditions. This enhances the ship's ability to cope with complex environments and improves the safety and reliability of navigation.
[0064] 5. Decision support: The method provides detailed data analysis and forecasts, providing strong decision support for ship operators and managers. With this information, they can better plan voyages, allocate resources and manage operations, thereby improving the overall efficiency and profitability of shipping operations.
[0065] 6. Technological innovation: This approach combines advanced machine learning algorithms, big data analysis and optimization techniques, demonstrating the great potential of applying innovative technologies in the shipping industry. It not only provides an effective way to improve the energy efficiency and environmental sustainability of ship operations, but also promotes the entire industry's transformation towards intelligence and digitalization. BRIEF DESCRIPTION OF THE DRAWINGS
[0066] Figure 1 The present invention is a flow chart of a method for estimating and calculating the cruising range and energy consumption of a ship.
[0067] Figure 2 This is a system diagram of the intelligent data analysis module of the present invention. DETAILED DESCRIPTION
[0068] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.
[0069] This case is mainly implemented through the intelligent data analysis module, which uses machine learning and big data technology to collect and analyze information such as the ship's historical cruising data, weather conditions, and sailing speed. First, the intelligent data analysis module collects data through sensor networks, log records, and external data sources. This data covers the ship's operating status, energy consumption, historical cruising range, and environmental factors such as weather conditions and ocean conditions. The collected data is pre-processed, including cleaning, normalization, denoising, and feature extraction to make it suitable for the needs of machine learning models. Next, the module uses machine learning techniques such as deep learning, support vector machines, or random forest algorithms to build prediction models based on historical data. These models are able to identify complex patterns and relationships in the data to accurately predict future energy consumption and cruising range. After the model is trained, it can provide instant predictions of energy consumption and cruising range based on real-time input variables such as sailing speed and current weather conditions. In addition, the intelligent data analysis module can continue to learn and adapt to new data, and the prediction accuracy will be further improved over time.
[0070] This case involves a dynamic environmental adaptation algorithm, which is specially designed to monitor and respond to changes in the current marine environment in real time in order to dynamically adjust the prediction model to achieve real-time prediction of the ship's range and energy consumption. The algorithm first uses sensors and external data sources to collect information about the current marine environment, including key parameters such as ocean currents, wind speed, temperature, humidity, etc. Then, it uses advanced data processing and pattern recognition techniques, such as machine learning or time series analysis, to analyze this environmental data to identify current and expected environmental conditions. Based on these analyses, the dynamic environmental adaptation algorithm adjusts the parameters of the energy consumption and range prediction model in real time by changing the weights of factors that affect energy consumption or adjusting the structure of the prediction model to reflect the latest ocean conditions. In this way, the prediction model can dynamically adapt to environmental changes and provide accurate range and energy consumption predictions. The algorithm can also identify and respond to sudden marine events or abnormal environmental conditions to ensure that the predictions remain accurate and reliable in all situations.
[0071] The case also covers the Energy Optimization Management System, which collects key data by monitoring the ship's energy usage in real time, such as fuel consumption rate, power usage and engine efficiency. It not only focuses on current energy consumption, but also uses intelligent data analysis modules to predict future energy consumption trends. This module combines machine learning and big data technologies to analyze historical energy consumption data, navigation conditions such as navigation speed, and external environmental factors, such as weather conditions and ocean current information, to provide accurate energy consumption forecasts. At the same time, the system includes a dynamic environmental adaptation algorithm that can adjust the prediction model in real time to reflect the latest environmental conditions and navigation data to ensure the real-time and accuracy of the prediction. Based on the prediction, the Energy Optimization Management System automatically calculates the optimal driving speed, heading and engine output, which are determined by a complex optimization algorithm that takes into account the predicted energy consumption, the performance limitations of the ship, and safety and regulatory requirements. The system then automatically adjusts the actual operating status of the ship by integrating with the ship's automatic navigation system and power control system to optimize energy consumption and cruising range. This optimization is dynamic, and the system continuously monitors energy usage and environmental conditions and adjusts navigation parameters based on real-time data and prediction results. In this way, the energy optimization management system can ensure that the ship operates in the most energy-efficient way while meeting the needs of the navigation mission, improving energy efficiency and reducing operating costs, which is of great significance to promoting sustainable shipping.
[0072] In addition, in this case: the steps of the intelligent data analysis module include several key components to ensure the effective collection, processing, and analysis of the ship's operating data. First, it is crucial to deploy the Internet of Things (IoT) devices and sensor networks on the ship. In the implementation, sensors may be installed on the ship's engine to monitor its performance and fuel consumption, or meteorological instruments may be installed on the outside of the ship to collect data on marine weather conditions. These devices can collect comprehensive data about the ship's sailing speed, engine status, fuel consumption, etc. in real time. For the sake of instantiation, the sensor is set to collect engine speed and temperature data every second, as well as the ship's GPS location and the temperature of the surrounding waters. Next, pre-processing the collected data using edge computing technology is a key step. Edge computing is a distributed computing architecture in which data is processed close to the data source (i.e., the "edge") rather than on a remote central server. In this case, the edge device (which may be a small computer or a specialized processor installed on the ship) processes the sensor data in real time. It can perform tasks such as data cleaning (in the implementation, removing outliers or filling missing values), data compression (reducing the amount of data that needs to be transmitted), and preliminary analysis (in the implementation, calculating the average performance indicators of the engine or identifying specific patterns). In this way, only necessary and relevant information needs to be sent to the central server, greatly reducing the amount of data transmission and improving processing speed and efficiency. In the implementation method, instead of sending all the raw data every second, the edge device may only send summary information of engine performance and any important abnormalities detected every minute. Therefore, the implementation method of this case not only enables the ship to monitor its operating status in real time, but also greatly improves the speed and efficiency of data processing by performing intelligent processing near the source of data, providing high-quality and relevant data for subsequent in-depth analysis.
[0073] In addition, the intelligent data analysis module also involves analyzing time series data using deep learning algorithms to capture complex patterns and relationships in ship operations and environmental conditions, and then implementing model optimization strategies to adapt to subtle changes in specific ships and routes. First, deep learning algorithms, such as recurrent neural networks (RNN) or long short-term memory networks (LSTM), are used to analyze time series data. These data may include sailing speed, engine status, fuel consumption, and ocean weather conditions collected by step P1. In implementation, an LSTM model can be trained to identify how specific engine performance patterns or weather conditions affect fuel efficiency.
[0074] Once a model is trained, it is able to identify complex patterns and relationships in the data, for example, it might discover that a vessel’s fuel efficiency decreases under certain current and wind speed conditions. These insights are critical to implementing model optimization strategies. By continuing to learn from new data and feedback, the model can continually optimize to fit the specific behavior of the vessel and the specific route it is sailing. For example, if the model discovers that a specific course adjustment can improve fuel efficiency under certain sea conditions, it can automatically recommend that adjustment. Model optimization is not just about improving the accuracy of predictions; it is also about making the model more flexible and adaptable, able to handle unknown or changing conditions.
[0075] Furthermore, the intelligent data analysis module also involves combining multiple data sources and using data fusion technology to achieve a comprehensive understanding of the ship's environment and provide accurate energy consumption predictions. This process begins with collecting and combining data from different sources. Ship data includes speed, direction, fuel consumption, and engine efficiency recorded by onboard sensors; open weather data may come from national meteorological services or commercial weather forecasts, providing information such as temperature, wind speed, wind direction, and air pressure; ocean current information includes the direction and speed of seawater, which is usually obtained through oceanographic research institutions or satellite technology; satellite monitoring data provides a wide range of information from marine biological activities to atmospheric conditions. In the implementation method, when a cargo ship crosses the ocean, its intelligent data analysis module collects real-time data from onboard sensors, obtains instant weather forecasts and ocean current conditions from the Internet, and satellite data provides a broad view of the sea area.
[0076] The subsequent data fusion process involves technologies such as Kalman filtering, neural networks or fusion algorithms, which can process and integrate data from different sources and in different formats to extract useful information and reduce uncertainty. In practice, data fusion technology can combine real-time ship position and speed data from sensors with wind speed and direction data from weather forecasts to predict the actual path of the ship and possible energy consumption. By integrating multiple data sources, the system can obtain a comprehensive view of the ship's operating environment, taking into account everything from atmospheric conditions to seawater dynamics, from the operating status of the ship itself to changes in the surrounding environment.
[0077] Ultimately, this comprehensive and highly information-based approach makes energy consumption forecasting more comprehensive and accurate. The system can not only provide forecasts based on current data, but also understand complex patterns and trends through a large amount of historical and real-time data, and predict the energy consumption and endurance performance of ships under different conditions. In implementation, by analyzing historical data, the system may find that in specific sea areas and specific seasons, specific speed and heading combinations will lead to higher fuel efficiency. In this way, not only the economy of navigation is improved, but also strong data support is provided for ship management.
[0078] Example 1: The ship "PACIFIC VISION" is sailing across the ocean. In order to optimize the sailing process and reduce energy consumption, IoT devices and sensor networks are installed on the ship to collect comprehensive data in real time, and edge computing technology is used to pre-process this data:
[0079] Data collection phase:
[0080] Speed monitoring: GPS and speed sensors installed on board monitor the speed in real time. Set the parameters a = 10 knots, b = 0.02, c = 5 knots, d = 0.1 to simulate the actual speed v(t) = 10·e 0.02t +5·sin(0.1t). This model takes into account acceleration and speed fluctuations caused by factors such as sea waves.
[0081] Engine status monitoring: Engine performance sensors measure engine status. Set k=3 and m=50 to simulate engine status This model accounts for changes in engine performance as engine age or other conditions change.
[0082] Fuel consumption monitoring: The fuel flow meter measures the real-time fuel consumption. Set γ = 0.5 and n = 2 to represent the fuel consumption:
[0083]
[0084] This reflects the non-linear relationship between fuel consumption and operating conditions such as speed and engine load.
[0085] Ocean weather monitoring: Integrate data from satellites and weather stations to monitor ocean weather. At a certain moment, the ocean weather conditions are simulated as W(w) = 2w 2 +0.5w+10, where w represents wind speed or other relevant weather factors.
[0086] Data preprocessing stage:
[0087] Preliminary analysis function: Use edge computing technology to perform preliminary analysis on the collected data. For example, at t = 10 hours, s = 60 hours, c = 100 hours, and w = 5 units of wind speed, calculate the preliminary analysis function:
[0088]
[0089] This function takes into account the impact of speed, engine status, fuel consumption and weather conditions on energy consumption.
[0090] Data compression transformation: In order to reduce the amount of data transmitted to the central server, a data compression transformation function is applied. For example, C t (Da ) are approximated as Taylor series.
[0091] Through these steps, ships can monitor their energy usage in real time and automatically adjust their speed, heading and engine output according to the pre-processing results of the intelligent data analysis module to optimize energy consumption and cruising range. By applying these technologies in practice, navigation efficiency and safety can be significantly improved while reducing energy consumption and operating costs.
[0092] Example 2: The navigation of the ship "Xin Mingzhou 98" is further illustrated by substituting actual data into the calculation results. Let this example expand each step in detail:
[0093] The "Xin Mingzhou 98" ship starts a new voyage. In this embodiment, the initial parameters are set to simulate the function of the sailing speed v(t) = 15e 0.01t +5sin(0.02t), where t is hours. At the end of the first day (set to 1.2 hours later), the actual value of the sailing speed is v(1.2) = 15e 0.012 +5sin(0.024), approximately 15.3 knots.
[0094] The engine status can be Indicated by, where s represents the voyage time. After 24 hours, the engine status is It's about 0.98, which indicates the engine is running well.
[0095] Fuel consumption can be calculated by integrating Calculated as follows, where c is the voyage time. Total fuel consumption in 24 hours: Approximately 4608 liters.
[0096] Ocean weather conditions, assuming the wind speed w is 10km / h, can be calculated by W(w) = 2w 2 +0.5w+10 to simulate, we get W(10)=2*10 2 +0.5*10+10, which is approximately 225.
[0097] The data of speed, engine status, fuel consumption and weather conditions are standardized through preliminary analysis performed by edge computing devices. The mean and standard deviation so far are set to μ(T s )=18 nodes and σ(T s )=4 knots, then the standardized sailing speed is Approximately 0.575.
[0098] In the transfer learning phase, the pre-trained model parameters Θ are used pre, the model adjusts parameters by minimizing the loss function to quickly adapt to the specific data of new ships and routes. Set the loss function L(Θ pre ,v(24)) is minimized after a series of iterations, which shows that the model has adapted to the current navigation conditions.
[0099] In the reinforcement learning phase, the system learns the optimal strategy by interacting with the environment. Set the reward function:
[0100] R(v(24),α(24),β(24)) takes into account the sailing speed, fuel efficiency and ocean weather conditions. The system optimizes the objective function G(T s ,α,β), and finally found a strategy to reduce energy consumption while maintaining good navigation performance.
[0101] During the real-time monitoring phase, the intelligent system of the Xinmingzhou 98 continuously receives new data and automatically adjusts the navigation strategy based on the model prediction results. For example, if the model predicts that fuel consumption will increase due to weather changes in the next few hours, the system may automatically adjust the course or slow down to improve fuel efficiency.
[0102] Example 3: The cargo ship “CUL SHEKOU” is sailing a long distance from the port of “Qingdao” to the port of “Guangzhou”. The ship is equipped with advanced sensors and data analysis systems to optimize the energy efficiency and safety of the voyage.
[0103] Step 1: Data Collection
[0104] At a specific time point of t=10 hours, the following data were collected in this embodiment:
[0105] Ship Data b (t): The cruising speed is 13 knots, the fuel consumption is 500 liters / hour, and the engine efficiency is 90%.
[0106] Weather Data d (t): Wind speed is 15km / h.
[0107] Ocean current information i (t): The ocean current speed is 5km / h.
[0108] Satellite monitoring data s (t): The sea surface temperature is 18℃.
[0109] Step 2: Customize the function application
[0110] Environmental understanding function F s :
[0111] Calculate the ship data integral ∫S b(t)dt. Assuming that the sailing speed remains basically stable within 10 hours, the integral result is approximately 13*10=130 knots.
[0112] Calculate the square of weather data δW d (t) 2 Given δ=0.1, the result is 0.1*15 2 =22.5.
[0113] Calculate the square root of ocean current information get
[0114] Calculate the cube λM of satellite monitoring data s (t) 3 Given λ=0.05, the result is 0.05*18 3 ≈291.6.
[0115] Finally, substitute these values into F s In the equation, given α = 2 and β = 1.5, we get:
[0116] F s =2[130+22.5]+1.5[2.24+291.6]≈745.76
[0117] Step 3: Application of data fusion technology
[0118] Feature-level fusion f :
[0119] Assume that this embodiment extracts the corresponding feature function F from each data source k , and assign a weight ω to each feature k For example, ω1=0.3, ω2=0.2, ω3=0.25, ω4=0.25.
[0120] Use a nonlinear transformation function φ, such as the sigmoid function. Calculate The result is then transformed through the sigmoid function. The calculation result is set to 0.85, and sigmoid0.85 is approximately 0.7, which is the result of feature-level fusion.
[0121] Decision-level fusion c :
[0122] Use a decision function ψ, such as linear regression or another neural network, to integrate T f (t).
[0123] Given that the prediction time range T is 10 hours, this embodiment calculates ψ(T f(t)) is integrated within this range and the result is set to 560, which represents the comprehensive prediction result.
[0124] Step 4: Implementation and Monitoring
[0125] Real-time monitoring and adjustment: The system monitors the energy usage and environmental conditions of the "D" wheel in real time and adjusts the c The ship’s speed, heading and engine output are automatically adjusted based on the results.
[0126] Example 4: The vessel "Xin Mingzhou 106" is equipped with advanced sensors, data analysis systems and a dynamic environmental adaptation algorithm designed to adjust its prediction model in real time to optimize cruising range and energy consumption:
[0127] Establishing the environment perception model E s (t): This embodiment first defines an environmental perception model, which will comprehensively consider the sailing speed V(t), engine status E(t), fuel consumption F(t), and ocean weather conditions M(t). Set at a specific time point of t=10 hours, this embodiment collects the following data:
[0128] The navigation speed V(10) is 12.5 knots;
[0129] Engine state E (10) indicates that the engine efficiency is 90%;
[0130] Fuel consumption F(10) is 500 liters / hour;
[0131] Marine weather M(10) means wind speed is 15km / h;
[0132] Model parameter setting: Setting parameter α i and β i To reflect the influence weight and nonlinearity of these factors. For example:
[0133] α1=0.5, β1=2, α2=0.3, β2=2, α3=0.1, β3=1, α4=0.1, β4=1
[0134] Calculate the output of the environmental perception model: Using the above parameters and the collected data, this embodiment can calculate E s (10), which will provide a comprehensive environmental perception value for this embodiment, which reflects the current navigation and environmental conditions.
[0135] Then, construct the prediction model P m (t): The model is designed to dynamically adjust its parameters based on the output of the environment perception model, using reinforcement learning or online machine learning methods to update and adapt to new environmental data in real time.
[0136] Real-time feedback loop: By comparing the results of the prediction model m (t) and the actual observed data O m (t), this embodiment can calculate the difference ΔR(t). Assuming that at t=10 hours, the predicted fuel consumption is 480 liters / hour, while the actually observed fuel consumption is 500 liters / hour.
[0137] Optimizing the model: By analyzing the value of ΔR(t), the present embodiment can optimize the prediction model to reduce the difference between the prediction and the actual. For example, the present embodiment may adjust the parameters of the model or change certain aspects of the learning algorithm to improve the accuracy of the prediction.
[0138] Continuous data collection and analysis: A ship will continue to collect data on sailing speed, engine status, fuel consumption and ocean weather conditions, and analyze this data in real time. Model iteration optimization: The dynamic environmental adaptation algorithm will continuously adjust the prediction model based on new data to ensure that the prediction is always the latest and most accurate. Real-time adjustment of the voyage plan: Based on the optimized prediction model, the ship's navigation system can adjust the sailing speed, heading and engine output in real time to optimize energy consumption and cruising range.
[0139] Example 5: The vessel “NEW JOURNEY” is equipped with advanced sensors and data analysis systems, including an intelligent data analysis module using deep learning models to predict and optimize energy consumption in real time.
[0140] This example initializes the deep learning model in is the depth of the hidden layer, and σ is the activation function, such as ReLU or Sigmoid. This embodiment randomly initializes the network parameters and
[0141] Historical energy usage data analysis: Assume that this embodiment collects historical energy usage data D(s) from the "E" wheel. In order to calculate the historical energy consumption E(t), this embodiment sets γ=0.1, and then integrates the historical data to obtain the energy consumption decay model over time.
[0142] Real-time data collection: At t = 24 hours, the real-time data of the ship is collected. Set the real-time ocean data M o (t) = 10cos(0.05t) and meteorological data W e (t) = 5sin(0.1t+π / 4), the current speed of the ship v(t) = 12 knots, and the heading h(t) = 30 degrees.
[0143] Comprehensive analysis: This embodiment uses a deep learning model to analyze historical energy data, real-time ocean and meteorological data, and the current state of the ship to obtain a forecast of the energy consumption trend at the current moment. For example, this embodiment normalizes these data and inputs them into the model, setting λ p =1.5 and μ p =10 to adjust the prediction results.
[0144] Model training and prediction: This example uses historical and real-time data to train a deep learning model, and then uses the model to predict current and future energy consumption trends. The model is set to predict that the ship's energy consumption will increase by 10% in the next 12 hours.
[0145] Adjustment of forecast results: Based on the forecast of the model, this embodiment calculates the adjusted energy consumption forecast This will provide guidance for the rest of the voyage.
[0146] Real-time monitoring and adjustment: The system monitors the energy usage and environmental conditions of a cargo ship in real time, and automatically adjusts the ship's speed, heading and engine output according to the prediction results to optimize energy consumption. Continuous optimization: As more data is accumulated and more voyages are completed, the prediction model of the intelligent data analysis module will continuously optimize and adjust itself to ensure the accuracy of the prediction and the energy efficiency of the voyage.
[0147] Example 6: The vessel "Xin Minh 102" is planning to travel from the port of "Ho Chi Minh City" to the port of "Chin Chow" and is equipped with an advanced monitoring system for real-time monitoring and analysis of changes in the marine environment, as well as a dynamic environmental adaptation algorithm designed to adjust its prediction model in real time:
[0148] Marine environment data collection: At a specific time point of t = 12 hours, the sensors of a certain ship collected the following environmental data:
[0149] The ocean current H(t)=2sin(0.1t+π / 6) means that the ocean current speed at a specific time point is 2 knots, the frequency of change is once every 10 hours, and there is a phase offset.
[0150] Wind speed W(t) = 15cos(0.05t+π / 4), which means the wind speed at a specific time point is 15km / h, the frequency of change is once every 20 hours, and there is a phase shift. Temperature T(t) = 20exp(0.001t 2 )+10, which means the temperature gradually decreases from 20℃, and the base temperature is 10℃.
[0151] Environment state model setting: This embodiment defines the environment state model E s(t) to synthesize the above changes. Set the model parameters θ1 = 1.5, θ2 = 1, θ3 = 0.5, δ h =2,δ w =1,δ t =2.
[0152] Calculate the output of the environmental state model: At t = 12 hours, this embodiment calculates E s (12) = 1.5H(12) 2 +1W(12)+0.5T(12) 2 , which will provide a comprehensive environmental perception value for this embodiment, which reflects the current navigation and environmental conditions.
[0153] Build and adjust the prediction model: Use the deep learning model P(t) to make an initial prediction of energy consumption. Set it to predict that the ship's energy consumption will increase by 5% in the next 24 hours without environmental adjustments. Real-time adjustment of the prediction model: In order to reflect the impact of real-time environmental changes on energy consumption, this embodiment uses the output of the environmental state model to adjust the energy consumption prediction. Set the environmental impact adjustment coefficient κ = 0.1, and then calculate the adjusted prediction P adj (12) = P(12) exp(0.1 E s (12)).
[0154] Real-time monitoring and adjustment: The system monitors the energy usage and environmental conditions of the "F" wheel in real time and adjusts the predicted P adj (t) Automatically adjust the ship's speed, heading and engine output to optimize energy consumption. Continuous optimization: As more data is accumulated and more voyages are completed, the dynamic environmental adaptation algorithm will continuously optimize and adjust itself to ensure the accuracy of predictions and the energy efficiency of voyages.
[0155] Example 7: The ship “New Valor” is equipped with an advanced monitoring system and a dynamic environmental adaptation algorithm designed to adjust its prediction model in real time:
[0156] Initialization control system: The automatic regulation control system of a certain ship adopts an optimization algorithm to automatically calculate and adjust the optimal driving speed, heading and engine output according to the prediction results provided by the intelligent data analysis module and the dynamic environment adaptation algorithm.
[0157] Real-time data collection and prediction:
[0158] Real-time environmental monitoring: At t = 24 hours, the sensor collected current, wind speed and temperature data. Set the current speed H(24) = 2sin(0.1*24+π / 6) knots, wind speed W(24) = 15cos(0.05*24+π / 4)km / h, temperature T(24) = 20exp(0.001*242 )+10℃.
[0159] Predicted energy consumption trends: The intelligent data analysis module and dynamic environmental adaptation algorithm combine real-time environmental data and historical ship operation data to provide predicted energy consumption trends.
[0160] Calculate optimal driving parameters: Based on the provided prediction results, the automatic regulation control system calculates the optimal driving speed, heading and engine output through the optimization algorithm. For example, set the current driving speed to 15 knots, heading to north, and engine output to 70%.
[0161] Driving speed optimization Ω(t,v,h,e): Calculate the energy consumption impact of a series of speed values and find the speed v that minimizes energy consumption * (24). Set α v (24) = 0.05, and the speed with the lowest energy consumption is found through the optimization algorithm. Course optimization Ψ(t,v,h,e): Similarly, considering the influence of wind direction and ocean current, find the optimal course h * (24). Set α h (24) = 1, and the energy consumption impact of a series of possible headings is calculated.
[0162] Engine output optimization Λ(t,v,h,e): Analyze the energy consumption under different engine outputs and find the optimal output e * (24). Set α e (24) = 0.1, calculate the energy consumption impact of a series of engine outputs.
[0163] Real-time adjustment of navigation plan: Based on the calculated optimal speed, heading and engine output, a ship's navigation system and power system automatically adjust to optimize energy consumption. Continuous monitoring and optimization: The system continuously monitors navigation status and environmental conditions, and continuously adjusts navigation strategies based on new data and forecast results to ensure continuous optimization of energy consumption.
[0164] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. A method for estimating and calculating the cruising range and energy consumption of a ship, characterized by: It includes an intelligent data analysis module for collecting and analyzing historical ship endurance data, weather conditions, and sailing speed information through machine learning and big data technology to establish an energy consumption prediction model; a dynamic environmental adaptation algorithm that adjusts the prediction model in real time according to current marine environmental changes to dynamically estimate the cruising range and energy consumption; and an energy optimization management system that monitors the ship's energy usage in real time and automatically adjusts the ship's speed, heading, and engine output according to the prediction results of the intelligent data analysis module and the dynamic environmental adaptation algorithm to optimize energy consumption and cruising range; The intelligent data analysis module includes the following three parts: P1. Use IoT devices and sensor networks deployed on ships to collect comprehensive data in real time, including sailing speed, engine status, fuel consumption, and ocean weather conditions; and use edge computing technology to pre-process data to reduce the amount of data transmitted to the central server and improve processing speed and efficiency; P2. Use deep learning algorithms to analyze time series data to capture complex patterns and relationships; thereby implementing model optimization strategies to adapt to subtle changes in specific vessels and routes; P3. Combine multiple data sources, including ship data, open weather data, ocean current information, and satellite monitoring data, to achieve a comprehensive understanding of the environment; and use data fusion technology to integrate data from different sources and types to provide more comprehensive and accurate energy consumption forecasts; The dynamic environment adaptation algorithm includes an environment perception model: where α i and β i Adjust the influence weights and nonlinearity of each factor to interpret the current marine environmental data, and build an adaptive prediction model to dynamically adjust its parameters according to the output of the environmental perception model, using reinforcement learning or online machine learning methods to ensure that the model can be updated and adapted to new environmental data in real time; The predictions are then compared to the actual observations through a real-time feedback loop, using the differences: Where P m (t) is the prediction model result, O m (t) is the observed data and the optimized model; The energy optimization management system includes an intelligent data analysis module, using a deep learning model in is the depth of the hidden layer, σ is the activation function, and is the network parameter; For analyzing historical energy usage data Where D(s) represents historical energy data, γ is the attenuation parameter; combined with the real-time ocean M o (t) = α m cos(β m t) and meteorological W e (t) = α w sin(β w t+φ w ) data, and the current status of the ship S c (t) = k1v(t) 2 +k2h(t) and the planned route, where v(t) and h(t) represent the navigation speed and heading respectively, for comprehensive analysis; Provide accurate energy consumption trend forecast where λ p and μ p To adjust the parameters of the prediction results; The energy optimization management system also includes a dynamic environmental adaptation algorithm to monitor and analyze changes in the ocean environment in real time, including the ocean current H(t)=a h sin(b h t+c h ), wind speed W(t) = a w cos(b w t+c w ), and temperature T(t) = a t exp(-b t t 2 )+c t ; where a h , b h , c h , a w , b w , c w , a t , b t , and c t is the specific coefficient of the environmental variable; The system builds an environmental state model To synthesize the changes, among which θ1, θ2, θ3, δ h , δ w , and δ t is the model parameter; the mechanism P used to adjust the prediction model in real time adj (t) = P(t)·exp(κ·E s (t)), where P(t) is the initial prediction, P adj (t) is the adjusted forecast, and κ is the environmental impact adjustment coefficient.
2. A method for estimating and calculating the cruising range and energy consumption of a ship according to claim 1, characterized in that The IoT devices and sensor network collect comprehensive data in real time on the ship, including the navigation speed v(t) = a·e bt +c·sin(dt), to simulate the complex conditions faced by actual sailing speed, where a, b, c, and d are adjustment parameters specific to the ship and the environment; Engine status: Used to simulate the nonlinear changes of engine performance under certain conditions, where k and m adjust the steepness and center point of the curve; Fuel consumption: The integral form represents the nonlinear relationship between fuel consumption and operating conditions, where γ′ and n are adjustment coefficients; Ocean weather conditions W(w) = α′w 2 +β′w+χ, to simulate the impact of weather conditions on ship operations, where α′, β′, and χ are adjustment parameters; The edge computing technology described pre-processes the collected data, including the preliminary analysis function of the data: The effects of speed, engine status, fuel consumption and weather conditions are combined; logarithmic and square root functions are used to enhance nonlinear relationships and interactions between data; And the data compression conversion function: A method that uses Taylor series approximation to reduce the amount of data transmitted while maintaining information integrity.
3. A method for estimating and calculating the cruising range and energy consumption of a ship according to claim 1, characterized in that The deep learning algorithm described analyzes time series data: T s (t)=A(t)·e B(t) +C(t)·sin(D(t)) Where A(t), B(t), C(t), and D(t) are time-varying coefficients that can capture the complex patterns and relationships of navigation speed, engine status, fuel consumption, and ocean weather conditions; Through the feature engineering and data preprocessing modules, the system automatically extracts features from raw data and standardizes them into a format optimized for deep learning models: Where μ(T s ) and σ(T s ) are the mean and standard deviation of the time series, respectively; the system further implements model optimization strategies, including transfer learning, which uses model parameters Θ pre-trained on similar tasks pre , by minimizing the loss function: Where P m The system is a prediction model that can quickly adapt to the specific data of new ships and routes. Combined with reinforcement learning, the system learns the optimal strategy by interacting with the environment and optimizes the objective function: Where γ″ is the discount factor, R is the reward function, and α(t) and β(t) are the action and state variables.
4. A method for estimating and calculating the cruising range and energy consumption of a ship according to claim 1, characterized in that The combination of multiple data sources, including ship data S b (t), open weather data W d (t), ocean current information H i (t), and satellite monitoring data M s (t), where t represents time; Using custom functions: Where α, β, δ, and λ are model parameters to achieve comprehensive environmental understanding; Further adopting data fusion techniques, including feature-level fusion: where F k represents the feature function extracted from each data source, ω k is the weight coefficient, φ is the nonlinear conversion function, and decision-level fusion: Where ψ is the decision function used to integrate various prediction results, and T is the prediction time range.
5. A method for estimating and calculating the cruising range and energy consumption of a ship according to claim 4, characterized in that The energy optimization management system also includes an automatic adjustment control system, which automatically calculates and adjusts the optimal driving speed through an optimization algorithm based on the prediction results provided by the intelligent data analysis module and the dynamic environment adaptation algorithm. course and engine output in, are the energy consumption influence functions of driving speed, heading and engine output, α v (t), α h (t), and α e (t) is the time-related adjustment coefficient, and T is the forecast period.
Citation Information
Patent Citations
Method for predicting ship navigation meteorological factors according to marine meteorological information
CN106599995A
Ship energy efficiency information collection management system and energy consumption state analysis method based on same
CN116362296A
Machine Learning to Accelerate Design of Energetic Materials
US20220067249A1
