PMS (Power Management System) power management system of geotextile laying ship

Abstract

The invention discloses a PMS power management system of a geotextile laying ship, and belongs to the technical field of intelligent management of geotextile laying ships. According to the method, the problem that the ship cannot operate normally or adverse effects are generated in the operation process due to occasional errors in the use process of the existing PMS power management system of the geotextile laying ship is solved, various historical data in the power system are trained and tested by adopting the neural network model, and the reliability of the power system is improved. Obtaining a model capable of predicting the running state of the power system; the method comprises the following steps of: embedding the data into a PMS (Power Management System) to monitor and predict various data of the power system of the geotextile laying ship in real time so as to obtain prediction results of various energy consumption and power in the power system, thereby helping a ship operator to better manage the power system of the ship; meanwhile, various energy consumption and power of the power system are adjusted according to the prediction result, intelligent monitoring of the geotextile laying ship is achieved, and the purposes of optimizing energy management and improving operation safety of the geotextile laying ship are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent management of laying vessels, in particular to a PMS power management system for laying vessels. Background Art

[0002] A laying vessel is a vessel used in the construction or repair of large structures. It is usually designed as a flatbed ship and can carry large amounts of concrete or other construction materials and then unload them to a nearby work area. A laying vessel can be a single-body design or a modular design to facilitate the formation of a larger work platform to meet different needs. In addition, laying vessels usually have a highly adjustable deck, which enables them to operate in different types of waters and environmental conditions. PMS is a type of power management system for laying vessels, which mainly monitors and intelligently optimizes the power system of laying vessels to achieve energy conservation and emission reduction and improve the reliability of the power system.

[0003] In the prior art, the PMS power management system of laying vessels is very complex, and errors may occasionally occur during use, causing the vessel to be unable to operate normally or causing adverse effects during operation. Therefore, the existing needs are not met, so we proposed a PMS power management system for laying vessels. Summary of the invention

[0004] The purpose of the present invention is to provide a PMS power management system for a laying ship. By adopting a neural network model to train and test various historical data in the power system of the laying ship, a model that can predict the operating status of the power system is obtained; then the model is embedded in the PMS power management system, so that it can monitor and predict various data of the power system of the laying ship in real time, thereby obtaining prediction results of various energy consumption and power in the power system, thereby helping ship operators to better manage the power system of the ship; at the same time, the energy consumption and power of the power system are adjusted according to the prediction results, thereby realizing intelligent monitoring of the laying ship, achieving the purpose of optimizing energy management, saving energy and reducing emissions, reducing operating costs, and improving the operating safety of the laying ship, solving the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: The PMS power management system for laying vessels includes: The data collection unit detects and tests the power system of the laying ship through the PMS power management system, and collects the engine data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data in the power system as data samples; The target unit determines the target operating conditions of the power system according to the use requirements, specified standards and environmental protection requirements of the laying ship. The target operating conditions include: idle point, economic point and maximum continuous working time, which serve as the basis for subsequent management; A model building unit, which builds a neural network model based on multiple data and target working conditions in the power system, uses the neural network model to extract feature information from multiple data of the power system, and describes the operating characteristics of the power system under different working conditions; The model practice unit embeds the neural network model into the PMS power management system and uses the neural network model to make real-time predictions of various energy consumption and power in the power system during the operation of the laying ship; The power management unit immediately takes power management actions when there is a difference between the predicted energy consumption output by the neural network model and the actual energy consumption; and summarizes and stores various management plans to form a strategy library.

[0006] Furthermore, the data collection unit includes: The data processing module cleans, fuses and removes noise from engine data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data; The data partitioning module divides the processed data into training sets and test sets according to the seven-part rule. The training set is used to train the neural network model, and the test set is used to evaluate the performance of the neural network model.

[0007] Furthermore, the model building unit includes: Model building module, defining input layer, hidden layer, output layer, activation function and loss function. The input layer inputs data samples, the hidden layer analyzes the data samples and extracts feature information from the data samples, and the output layer outputs the prediction results. The activation function and loss function prevent the neural network model from overfitting. The model training module imports the processed data samples into the neural network model for training and testing, simulating the prediction results of the power system of the laying ship under different working conditions; and compares the prediction results with the actual results to verify the performance of the neural network model; The model optimization module organizes and arranges the new data in the ship's power system and incorporates it into the neural network model for optimization and updating.

[0008] Furthermore, the model optimization module includes: A partitioning module determines the amount of new data in the power system of the laying ship, divides the global distribution of data into multiple local distribution groups based on data aggregation, and determines the distribution ratio of the new data in the local distribution groups; The data volume determination module determines the historical data volume of the historical data to be optimized according to the following formula based on the distribution ratio of each new data item in the local distribution group and the distribution ratio of the historical data in the local distribution group and the data volume of each new data item;

[0009] Among them, H represents the amount of historical data involved in optimization, N represents the amount of new data, n represents the number of local distribution groups, K i Indicates the proportion of new data in the i-th local distribution group, L i represents the proportion of historical data in the i-th local distribution group, ΔD represents the preset difference ratio, τ i represents the influence weight of the i-th local distribution group, and C represents a custom constant.

[0010] Furthermore, the model optimization module further includes: The layer number determination module determines the number of layers of the neural network model, and determines the optimized number of layers according to the following formula based on the data difference information between the new data and the historical data;

[0011] Among them, s represents the number of optimized levels, R represents the number of levels of the neural network model, m represents the number of items of data, ΔE j represents the data difference between the i-th new data and historical data, β j represents the impact weight of the i-th new data, represents the mean reference total difference; The optimization module randomly selects historical participation data of historical data volume from historical data, forms a new optimization data group with historical participation data and new data, obtains the optimized number of layers to be optimized in order from the top layer to the bottom layer from the neural network model, optimizes the layers to be optimized based on the optimization data group, and realizes the optimization update of the neural network model.

[0012] Furthermore, the model practice unit includes: Model integration module, which embeds the verified neural network model into the PMS power management system of the laying vessel, and uses the neural network model to predict energy consumption and power in real time during the operation of the laying vessel; The real-time processing module transmits various data of the power system to the PMS power management system of the laying ship in real time, analyzes and predicts the real-time power data through the neural network model, and obtains the prediction results of various energy consumption and power in the power system.

[0013] Furthermore, the model practice unit further includes: The abnormal alarm module enables the PMS power management system to immediately send an alarm to the terminal held by the ship operator and provide corresponding suggestions when the neural network model predicts a certain abnormal situation.

[0014] Furthermore, the power management unit comprises: The strategy storage module records the actual energy consumption, predicted energy consumption and selected strategy information each time to form a strategy library, and regularly updates and optimizes the strategies in the strategy library; The automatic management module selects appropriate strategies from the pre-defined strategy library for execution based on the difference between predicted and actual energy consumption and the current operating status of the ship, and feeds back the adjustment results to the terminal held by the ship operator in real time.

[0015] Furthermore, the strategy storage module regularly updates and optimizes the strategies in the strategy library including: Obtain the historical management data of the strategy library for the PMS power management of the laying ship, and obtain the newly added change data on the PMS power management of the laying ship; Determine the power management effect of the current strategy on the current laying ship situation based on the historical management data, and determine the first current strategy whose power management effect does not meet the preset effect and its corresponding power management situation; Obtaining attribute keywords of the newly added change data, obtaining relevant keywords about the attributes in the policy library, matching the attribute keywords with the relevant keywords, and retrieving a second current policy with a matching degree greater than a preset matching degree from the policy library according to the matching result; Establishing a difference sequence based on the difference between all power management situations of the first current strategy and the target management effect, establishing a power value management sequence corresponding to the difference sequence based on all power management situations, matching the difference sequence with the power value management sequence, and determining the corresponding relationship between the difference and the power value according to the sequence matching result; A first optimization strategy for optimizing a first current strategy under a power value based on the corresponding relationship; Inputting the change amplitude of the newly added change data into the neural network model to obtain the corresponding power prediction result, and based on the difference between the power prediction result and the standard power result, determining the optimization value of the second current strategy to obtain the second optimization strategy; Obtaining a first distribution of a first optimization strategy in the PMS power management of a laying ship, and a second distribution of a second optimization strategy in the PMS power management of a laying ship, and integrating, verifying and adjusting the first optimization strategy and the second optimization strategy based on distribution correlation characteristics between the first distribution and the second distribution to obtain a target optimization strategy; The first current strategy and the second current strategy in the strategy library are updated based on the target optimization strategy.

[0016] Furthermore, the data collection unit further includes: The data storage module stores the processed data in the database and displays them using a directed graph structure; The data update module regularly updates the real-time collected engine data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention uses a neural network model to train and test various historical data in the power system of the laying ship, and obtains a model that can predict the operating status of the power system; then embeds it into the PMS power management system, so that it can monitor and predict various data of the power system of the laying ship in real time, thereby obtaining prediction results of various energy consumption and power in the power system, thereby helping ship operators to better manage the power system of the ship; at the same time, the energy consumption and power of the power system are adjusted according to the prediction results, thereby realizing intelligent monitoring of the laying ship, achieving the purpose of optimizing energy management, saving energy and reducing emissions, reducing operating costs, and improving the safety of laying ship operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a diagram of the module composition of the PMS power management system for laying vessels of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In order to solve the technical problem that the PMS power management system of the laying ship is very complex in the existing technology, errors may occur occasionally during use, resulting in the ship being unable to operate normally or causing adverse effects during operation, please refer to Figure 1 , this embodiment provides the following technical solutions: The PMS power management system for laying vessels includes: The data collection unit detects and tests the power system of the laying ship through the PMS power management system, and collects the engine data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data in the power system as data samples; among them, the engine data includes: the number of engine starts, working time, cumulative working time, idling time, acceleration time and deceleration time, etc., based on which the use and health of the engine can be evaluated; the generator set data includes: the number of generator starts, working time, cumulative working time, load change rate, no-load current, load current, voltage and frequency, etc., based on which the working status and efficiency of the generator set can be evaluated. rate; auxiliary equipment data include: the number of auxiliary equipment starts, working time, cumulative working time, operating status, load change rate and working efficiency, etc., based on which the working status and efficiency of the auxiliary equipment can be evaluated; energy consumption data include: fuel consumption, electricity consumption and seawater consumption, etc., which can be used to evaluate the energy consumption of the laying ship; environmental data include: temperature, humidity, air pressure, wind direction and speed, etc., which can be used to evaluate the environmental conditions of the laying ship; fault alarm data include: information such as the time, location, type and severity of the fault, which can be used to quickly locate and eliminate the fault; maintenance data include: information such as the time, content, method and cost of maintenance, which can be used to grasp the maintenance status of the laying ship.

[0022] Data collection unit, including: The data processing module cleans and processes engine data, generator data, auxiliary engine data, energy consumption data, environmental data, fault alarm data, and maintenance data, such as checking and repairing logical errors, incorrect date / time stamps, handling conflicts, etc., to ensure data quality and consistency; data source fusion, such as fusing features from different data sources, using clustering algorithms to group similar equipment or sensor data, or using rule-based methods to associate and combine data from different data sources; and denoising, such as using filtering, smoothing, removing the mean, removing variance, etc. to reduce the impact of Gaussian white noise, periodic noise, salt and pepper noise, etc.; Specifically, by cleaning, processing, and analyzing the above-mentioned data, more accurate, timely, and valuable data information can be provided to help ship operators make more informed decisions. At the same time, the data processing module can also fuse data from different data sources to form a larger data set to further mine valuable information and insights.

[0023] The data partitioning module divides the processed data into a training set and a test set by the seven-point rule. The training set trains the neural network model, and the test set evaluates the performance of the neural network model. In this embodiment, for example, all the data are organized into a list, the total number is calculated, and the total number is divided by 7 to obtain the average of each category, which is regarded as the dividing point between the training set and the test set. Secondly, for each category, the largest value is found and recorded; all entries less than or equal to the minimum value are deleted from the data list and put into the first set, namely, the training set; and the remaining data entries are put into the second set, namely, the test set; thus, 70% of the training set and 30% of the test set are obtained.

[0024] The data storage module stores the processed data in the database and displays them in a directed graph structure. In the present embodiment, for example, an object-oriented database can be deployed in a cloud server connected to the PMS power management system of the laying vessel, and each device can be used as a node to connect the related data to form a directed graph structure. When performing data analysis, the graph structure can be directly traversed to quickly find the related data, which is convenient for subsequent data analysis and prediction.

[0025] The data update module regularly updates the real-time collected engine data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data; in this embodiment, a regular time interval can be set, such as: daily, weekly or monthly, and then the data sampling time points are performed at certain time points; for example: data collection is performed once at 1 a.m. every day, and the power system data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data of the most recent day are all uploaded to the database for storage.

[0026] The target clarification unit determines the target operating conditions of the power system according to the use requirements, specified standards and environmental protection requirements of the laying ship. The target operating conditions include: idle point, economic point and maximum continuous working time, which serve as the basis for subsequent management; among them, the idle point refers to the engine operating point with the minimum fuel consumption when the ship is moored or waiting; when simulating, the idle point should be set at the slowest speed of the ship to ensure the safety and comfort of the ship; the economic point means that the ship can meet certain economic benefits while minimizing the impact of the exhaust gas emitted by the ship on the environment; therefore, it is necessary to clarify the economy, fuel efficiency, emission standards and other factors of the actual laying ship; the maximum continuous working time refers to the longest time that the ship can continue to work under the premise of meeting safety and economy; therefore, it is necessary to clarify the design parameters, operating conditions, maintenance cycle and other factors of the actual laying ship; by clarifying the above three factors, the accuracy of the neural network model simulation can be further ensured.

[0027] The model building unit establishes a neural network model based on multiple data and target working conditions in the power system, uses the neural network model to extract characteristic information from multiple data of the power system, and describes the operating characteristics of the power system under different working conditions; the model building unit includes: The model building module defines the input layer, hidden layer, output layer, activation function and loss function. The input layer inputs data samples, the hidden layer analyzes the data samples and extracts the feature information in the data samples, and the output layer outputs the prediction results. The activation function and loss function prevent the neural network model from overfitting. Specifically, by taking the operating status, environmental parameters and fault information of the laying ship as the input layer, such as speed, heading, draft, load, meteorological data and wave data, the actual operating status of the laying ship can be reflected. After establishing the input layer, one or more hidden layers are set up, the purpose of which is to extract deeper features and convert the original data into higher-order information, such as the energy consumption pattern and power prediction of the laying ship. The output layer will predict the energy consumption and power of the laying ship, so as to help the ship operator decide how to use the engine, thereby saving fuel and reducing carbon emissions. In order to ensure that the neural network model can fit the above complex nonlinear relationship, activation functions such as ReLU and loss functions such as cross entropy and mean square error can be used for calculation.

[0028] The model training module imports the processed data samples into the neural network model in sequence for training and testing, simulates the prediction results of the power system of the laying ship under different working conditions, and compares the prediction results with the actual results to verify the performance of the neural network model; specifically, after constructing the neural network model framework, the data samples in the training set are input into the neural network model one by one for training, and the prediction results of the model on each test set are recorded; and during the training process, the learning process of the model can be observed through the visualized loss function graph and accuracy curve; when the model training is completed, the data samples of the test set are input into the neural network model for prediction; the predicted results can be compared with the actual test results to calculate the prediction error; a smaller prediction error indicates that the model has a better fitting effect, and a larger prediction error indicates that the model has a poor fitting effect; if the prediction error of the model is within an acceptable range, the trained model can be applied to actual scenarios for real-time prediction; if the prediction error is too large, the structure or parameters of the model can be adjusted, and a new round of training can be carried out.

[0029] The model optimization module organizes the new data in the ship's power system and incorporates it into the neural network model for optimization and update to make it more accurate and effective.

[0030] Beneficial effects achieved by the above: Through the above operations, complex laying ship operations and nonlinear relationships can be learned, which helps ship operators to better understand and optimize the power system, thereby providing more accurate predictions and better control strategies.

[0031] In one embodiment, the model optimization module includes: A partitioning module determines the amount of new data in the power system of the laying ship, divides the global distribution of data into multiple local distribution groups based on data aggregation, and determines the distribution ratio of the new data in the local distribution groups; The data volume determination module determines the historical data volume of the historical data to be optimized according to the following formula based on the distribution ratio of each new data item in the local distribution group and the distribution ratio of the historical data in the local distribution group and the data volume of each new data item;

[0032] Among them, H represents the amount of historical data involved in optimization, N represents the amount of new data, n represents the number of local distribution groups, K i Indicates the proportion of new data in the i-th local distribution group, L i represents the proportion of historical data in the i-th local distribution group, ΔD represents the preset difference ratio, τ i represents the influence weight of the i-th local distribution group, and C represents a custom constant.

[0033] The layer number determination module determines the number of layers of the neural network model, and determines the optimized number of layers according to the following formula based on the data difference information between the new data and the historical data;

[0034] Among them, s represents the number of optimized levels, R represents the number of levels of the neural network model, m represents the number of items of data, ΔE j represents the data difference between the i-th new data and historical data, β j represents the impact weight of the i-th new data, represents the mean reference total difference; The optimization module randomly selects historical participation data of historical data volume from historical data, forms a new optimization data group with historical participation data and new data, obtains the optimized number of layers to be optimized in order from the top layer to the bottom layer from the neural network model, optimizes the layers to be optimized based on the optimization data group, and realizes the optimization update of the neural network model.

[0035] In this embodiment, the global data distribution is divided into multiple local distribution groups based on data aggregation conditions. Based on all data types, the value range under each data type constitutes the local data of all data types as a distribution.

[0036] In this embodiment, the ratios of the local distribution groups are all between (0, 1), and the sum of the ratios of all the local distribution groups is 1.

[0037] In this embodiment, the greater the influence of the local distribution group on the model output result, the greater the corresponding influence weight.

[0038] In this embodiment, C is always greater than .

[0039] In this embodiment, the data difference values ​​between the new data and the historical data are unified and set within the range of (0, 1).

[0040] In this embodiment, the greater the influence of each item of new data on the model output result, the greater the corresponding influence weight.

[0041] In this embodiment, Always less than 1.

[0042] The beneficial effects of the above design scheme are: by determining the amount of historical data involved in the optimization based on the distribution ratio of each new data item in the local distribution group and the distribution ratio of historical data in the local distribution group, combined with the data volume of each new data item, the overfitting phenomenon caused by only using new data for model optimization and the inefficiency of model optimization caused by adding too much historical data are avoided by adding an appropriate amount of historical data. At the same time, based on the distribution ratio of each new data item in the local distribution group and the distribution ratio of historical data in the local distribution group, combined with the data volume of each new data item and the historical data volume of historical data involved in the optimization, an optimized number of layers to be optimized in order from the top layer to the bottom layer are obtained from the neural network model, and the layers to be optimized are optimized based on the optimized data group, so as to realize the optimization update of the neural network model, ensure the invariance of the underlying features, improve the efficiency and stability of the model optimization, realize the optimization of the model implementation, and help ship operators better understand and optimize the power system, thereby providing more accurate predictions and better control strategies.

[0043] The model practice unit embeds the neural network model into the PMS power management system and uses the neural network model to make real-time predictions of various energy consumption and power in the power system during the operation of the laying ship. The model practice unit includes: Model integration module, which embeds the verified neural network model into the PMS power management system of the laying vessel, and uses the neural network model to predict energy consumption and power in real time during the operation of the laying vessel; The real-time processing module transmits the data of the power system to the PMS power management system of the laying ship in real time, analyzes and predicts the real-time power data through the neural network model, and obtains the prediction results of various energy consumption and power in the power system; in this embodiment, the neural network model is integrated into the PMS power management system of the laying ship, which specifically includes the following steps: Define the input and output formats of the neural network model to ensure that the neural network model can easily interact with the power management system of the laying vessel; write code or script to load the trained neural network model, including: weight parameters and model structure and other information; input the processed data into the model for real-time prediction to obtain the predicted energy consumption and power; then visualize the prediction results to facilitate operators to observe and adjust the ship's operating strategy, thereby helping ship operators to better manage the ship's power system, thereby improving energy efficiency and reducing operating costs.

[0044] At the same time, it also includes: collecting the gap between the model prediction and the actual energy consumption, updating the parameters of the neural network model to make it closer to the actual situation; and continuously iterating and optimizing the neural network model based on new data to make it more accurately predict energy consumption and power; if the model's prediction effect deteriorates, the latest model can be rolled back to restore it to its previous good state.

[0045] The abnormal alarm module, when the neural network model predicts a certain abnormal situation, such as: too high fuel consumption or too low engine speed, the PMS power management system will immediately issue an alarm to the terminal held by the ship operator and provide corresponding suggestions; specifically, the corresponding abnormal threshold is set according to the prediction result of the neural network model. When the predicted value exceeds the defined threshold, it can be judged as an abnormal situation; combined with the actual operation data of the ship obtained by the PMS power management system, such as: fuel consumption, engine speed, etc., it is compared with the data predicted by the model to find out the abnormality; if an abnormal situation is detected, an alarm is immediately issued to the terminal held by the ship operator. The alarm content includes: the specific value of the abnormality, the type of abnormality and possible solutions. Solutions such as: changing the voyage plan, reducing fuel consumption, etc., help the ship operator solve the problem as soon as possible.

[0046] The power management unit takes power management actions immediately when there is a difference between the predicted energy consumption output by the neural network model and the actual energy consumption; such as adjusting the engine settings or changing the flight plan to achieve more accurate and intelligent power management; and summarizes and stores various management plans to form a strategy library; the power management unit includes: The strategy storage module records the actual energy consumption, predicted energy consumption and selected strategy information each time to form a strategy library, and regularly updates and optimizes the strategies in the strategy library to improve the accuracy of prediction and the precision of control; In one embodiment, the policy storage module regularly updates and optimizes the policies in the policy library including: Obtain the historical management data of the strategy library for the PMS power management of the laying ship, and obtain the newly added change data on the PMS power management of the laying ship; Determine the power management effect of the current strategy on the current laying ship situation based on the historical management data, and determine the first current strategy whose power management effect does not meet the preset effect and its corresponding power management situation; Obtaining attribute keywords of the newly added change data, obtaining relevant keywords about the attributes in the policy library, matching the attribute keywords with the relevant keywords, and retrieving a second current policy with a matching degree greater than a preset matching degree from the policy library according to the matching result; Establishing a difference sequence based on the difference between all power management situations of the first current strategy and the target management effect, establishing a power value management sequence corresponding to the difference sequence based on all power management situations, matching the difference sequence with the power value management sequence, and determining the corresponding relationship between the difference and the power value according to the sequence matching result; A first optimization strategy for optimizing a first current strategy under a power value based on the corresponding relationship; Inputting the change amplitude of the newly added change data into the neural network model to obtain the corresponding power prediction result, and based on the difference between the power prediction result and the standard power result, determining the optimization value of the second current strategy to obtain the second optimization strategy; Obtaining a first distribution of a first optimization strategy in the PMS power management of a laying ship, and a second distribution of a second optimization strategy in the PMS power management of a laying ship, and integrating, verifying and adjusting the first optimization strategy and the second optimization strategy based on distribution correlation characteristics between the first distribution and the second distribution to obtain a target optimization strategy; The first current strategy and the second current strategy in the strategy library are updated based on the target optimization strategy.

[0047] In this embodiment, based on the correspondence between the difference and the power value, the first current strategy is optimized to optimize the strategy to eliminate the management difference under the power value.

[0048] In this embodiment, the newly added change data include, for example, changes in characteristics of the laying vessel, changes in power management standards, and the like.

[0049] In this embodiment, the greater the difference between the power prediction result and the standard power result, the greater the corresponding optimization value.

[0050] In this embodiment, the distribution association characteristics between the first distribution and the second distribution include, for example, a cross strategy, a parallel strategy, a conflict strategy, and the like.

[0051] The beneficial effect of the above design scheme is: by obtaining the historical management data of the policy library on the PMS power management of the laying ship, and obtaining the new change data on the PMS power management of the laying ship, the strategies related to the policy library are optimized through the two aspects of historical management data and new change data. The optimization process is carried out based on the power management effect to ensure the real-time and accuracy of the finally updated policy library.

[0052] The automatic management module selects appropriate strategies from the pre-defined strategy library for execution based on the difference between the predicted energy consumption and the actual energy consumption, as well as the current ship operation status; for example, it can recommend switching to a backup engine or changing the speed when necessary, and feed back the adjustment results to the terminal held by the ship operator in real time to keep the laying ship in the optimal operating state. In this embodiment, the power adjustment of the laying ship power system includes the following solutions: First, energy-saving mode: according to different sailing conditions, the best combination of engine and propeller is selected to achieve the best energy-saving effect; specifically, through real-time monitoring of the speed, heading, draft and loading status of the laying ship, as well as meteorological and hydrological conditions and other factors, the engine and propeller speed, output shaft size, etc. can be dynamically adjusted to achieve the best economic operation state.

[0053] Second, emission optimization: through real-time monitoring of the emission of the laying ship, as well as related meteorological and hydrological factors, the fuel injection amount and injection angle of the combustion chamber are reasonably adjusted to achieve the best emission control target and reduce the pollution of the laying ship to the surrounding water environment.

[0054] Third, load control: The PMS power management system can monitor the various load conditions of the laying ship in real time, such as propulsion force, number of crew members, and cargo loading status, and adjust the load as needed to achieve the best energy efficiency ratio.

[0055] Fourth, energy management: by collecting and analyzing a large amount of laying ship operation data, the energy use of the laying ship can be comprehensively managed and controlled; for example: through energy metering equipment, the energy consumption of the laying ship can be monitored in real time, and the energy consumption can be quantitatively allocated, so as to help the laying ship achieve the best energy utilization.

[0056] The beneficial effects achieved by the above content: Through the above operations, it helps the ship operator to better manage the ship's power system, thereby realizing intelligent monitoring of the laying ship, optimizing energy management, saving energy and reducing emissions, reducing operating costs, and improving the safety of laying ship operation.

[0057] Working principle: By using a neural network model to train and test various historical data in the power system of the laying ship, a model that can predict the operating status of the power system is obtained; then it is embedded in the PMS power management system to enable it to monitor and predict various data of the laying ship power system in real time, thereby obtaining the prediction results of various energy consumption and power in the power system; at the same time, the energy consumption and power of the power system are adjusted according to the prediction results, thereby achieving the purpose of intelligent monitoring of the laying ship.

[0058] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "including", "having" or any other variations thereof are intended to cover non-exclusive possessing, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0059] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The PMS power management system for laying vessels is characterized by: include: The data collection unit collects the engine data, generator data, auxiliary engine data, energy consumption data, environmental data, fault alarm data and maintenance data of the laying vessel through the PMS power management system; The target clarification unit determines the target operating conditions of the power system according to the use requirements of the laying ship. The target operating conditions include: idle point, economic point and maximum continuous working time; A model building unit is used to build a neural network model based on various data and target working conditions, use the neural network model to describe the operating characteristics of the power system under different working conditions, and regularly optimize the neural network model; the optimization process includes: optimizing the neural network model based on a new optimization data group composed of various historical data and new data; The model practice unit uses a neural network model to make real-time predictions of various energy consumption and power of the power system; The power management unit takes power management actions when there is a difference between the predicted energy consumption and the actual energy consumption; aggregates various plans to form a strategy library, and regularly updates the optimization strategy library; the update and optimization process includes: updating the current newly added change data, and optimizing the historical strategies that do not meet the preset effects based on the newly added change data.

2. The PMS power management system for laying vessels according to claim 1, characterized in that: The data collection unit comprises: The data processing module cleans, fuses and removes noise from engine data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data; The data partitioning module divides the processed data into training sets and test sets according to the seven-part rule. The training set is used to train the neural network model, and the test set is used to evaluate the performance of the neural network model.

3. The PMS power management system for laying vessels according to claim 1, characterized in that: The model building unit comprises: Model building module, defining input layer, hidden layer, output layer, activation function and loss function. The input layer inputs data samples, the hidden layer analyzes the data samples and extracts feature information from the data samples, and the output layer outputs the prediction results. The activation function and loss function prevent the neural network model from overfitting. The model training module imports the processed data samples into the neural network model for training and testing, simulating the prediction results of the power system of the laying ship under different working conditions; and compares the prediction results with the actual results to verify the performance of the neural network model; The model optimization module organizes and arranges the new data in the ship's power system and incorporates it into the neural network model for optimization and updating.

4. The PMS power management system for laying vessels according to claim 3 is characterized by: The model optimization module comprises: A partitioning module determines the amount of new data in the power system of the laying ship, divides the global distribution of data into multiple local distribution groups based on data aggregation, and determines the distribution ratio of the new data in the local distribution groups; The data volume determination module determines the historical data volume of the historical data to be optimized according to the following formula based on the distribution ratio of each new data item in the local distribution group and the distribution ratio of the historical data in the local distribution group and the data volume of each new data item; ； Among them, H represents the amount of historical data involved in optimization, N represents the amount of new data, n represents the number of local distribution groups, K i Indicates the proportion of new data in the i-th local distribution group, L i represents the proportion of historical data in the i-th local distribution group, ΔD represents the preset difference ratio, τ i represents the influence weight of the i-th local distribution group, and C represents a custom constant.

5. The PMS power management system for laying vessels according to claim 4, characterized in that: The model optimization module further includes: The layer number determination module determines the number of layers of the neural network model, and determines the optimized number of layers according to the following formula based on the data difference information between the new data and the historical data; ； Among them, s represents the number of optimized levels, R represents the number of levels of the neural network model, m represents the number of items of data, ΔE j represents the data difference between the i-th new data and historical data, β j represents the impact weight of the i-th new data, represents the mean reference total difference; The optimization module randomly selects historical participation data of historical data volume from historical data, forms a new optimization data group with historical participation data and new data, obtains the optimized number of layers to be optimized in order from the top layer to the bottom layer from the neural network model, optimizes the layers to be optimized based on the optimization data group, and realizes the optimization update of the neural network model.

6. The PMS power management system for laying vessels according to claim 1, characterized in that: The model practice unit includes: Model integration module, which embeds the verified neural network model into the PMS power management system of the laying vessel, and uses the neural network model to predict energy consumption and power in real time during the operation of the laying vessel; The real-time processing module transmits various data of the power system to the PMS power management system of the laying ship in real time, analyzes and predicts the real-time power data through the neural network model, and obtains the prediction results of various energy consumption and power in the power system.

7. The PMS power management system for laying vessels according to claim 6, characterized in that: The model practice unit also includes: The abnormal alarm module enables the PMS power management system to immediately send an alarm to the terminal held by the ship operator and provide corresponding suggestions when the neural network model predicts a certain abnormal situation.

8. The PMS power management system for laying vessels according to claim 1, characterized in that: The power management unit comprises: The strategy storage module records the actual energy consumption, predicted energy consumption and selected strategy information each time to form a strategy library, and regularly updates and optimizes the strategies in the strategy library; The automatic management module selects appropriate strategies from the pre-defined strategy library for execution based on the difference between predicted and actual energy consumption and the current operating status of the ship, and feeds back the adjustment results to the terminal held by the ship operator in real time.

9. The PMS power management system for laying vessels according to claim 8, characterized in that: The strategy storage module regularly updates and optimizes the strategies in the strategy library, including: Obtain the historical management data of the strategy library for the PMS power management of the laying ship, and obtain the newly added change data on the PMS power management of the laying ship; Determine the power management effect of the current strategy on the current laying ship situation based on the historical management data, and determine the first current strategy whose power management effect does not meet the preset effect and its corresponding power management situation; Obtaining attribute keywords of the newly added change data, obtaining relevant keywords about the attributes in the policy library, matching the attribute keywords with the relevant keywords, and retrieving a second current policy with a matching degree greater than a preset matching degree from the policy library according to the matching result; Establishing a difference sequence based on the difference between all power management situations of the first current strategy and the target management effect, establishing a power value management sequence corresponding to the difference sequence based on all power management situations, matching the difference sequence with the power value management sequence, and determining the corresponding relationship between the difference and the power value according to the sequence matching result; A first optimization strategy for optimizing a first current strategy under a power value based on the corresponding relationship; Inputting the change amplitude of the newly added change data into the neural network model to obtain the corresponding power prediction result, and based on the difference between the power prediction result and the standard power result, determining the optimization value of the second current strategy to obtain the second optimization strategy; Obtaining a first distribution of a first optimization strategy in the PMS power management of a laying ship, and a second distribution of a second optimization strategy in the PMS power management of a laying ship, and integrating, verifying and adjusting the first optimization strategy and the second optimization strategy based on distribution correlation characteristics between the first distribution and the second distribution to obtain a target optimization strategy; The first current strategy and the second current strategy in the strategy library are updated based on the target optimization strategy.

10. The PMS power management system for laying vessels according to claim 2, characterized in that: The data collection unit further includes: The data storage module stores the processed data in the database and displays them using a directed graph structure; The data update module regularly updates the real-time collected engine data, generator data, auxiliary machine data, energy consumption data, environmental data, fault alarm data and maintenance data.

Images