A method, device and system for joint scheduling optimization of hydrogen-electric energy route network for fuel cell ships

Abstract

The application discloses a kind of hydrogen-electric energy road network combined scheduling optimization method, device and system for fuel cell ship, comprising the following steps: identifying the working condition parameter in the process of ship running, generating working condition label, determining the value range and constraint condition of variable under current working condition, for dynamically switching the scheduling model boundary condition and strategy selection in hydrogen system of fuel cell;The basic identification of the working condition parameter in the process of ship running, the influence of coupling external speed and load and establishing steady-state tidal flow relationship;Based on the working condition parameter in the process of ship running identified, the prediction of load in future time of ship is carried out and the determination of optimal strategy of ship in current state is carried out;The scheduling optimal model of the least hydrogen consumption of ship under different operating conditions is constructed, the scheduling optimal model is solved, and the output power of the hydrogen system and the power system of the ship is obtained.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of new energy ship power, and relates to a hydrogen-electric energy path network joint scheduling optimization method, device and system for a fuel cell ship. BACKGROUND

[0002] With the increasingly stringent global emission reduction requirements for the shipping industry, traditional diesel-based ship power systems are facing major challenges in green and low-carbon transformation. Hydrogen fuel cells are widely considered as one of the important technical paths for future green ships due to their zero emissions, high efficiency, and low noise. Under this trend, the research and application of fuel cell ships are gradually accelerating, including inland passenger ships, marine traffic boats, and coastal cargo ships, etc. However, fuel cell ships face challenges such as complex multi-source energy regulation, large dynamic load fluctuations, and frequent changes in operating conditions during actual operation, and there is an urgent need to build a hydrogen-electric collaborative energy management system with high responsiveness and high intelligence.

[0003] Existing fuel cell ships mostly adopt a hybrid power architecture of fuel cells + lithium batteries, but their energy scheduling mostly relies on fixed thresholds or control strategies based on empirical rules, which are difficult to adapt to the flexible demand for energy supply under different operating conditions. For example, there is a sudden high power climb at the start of the voyage, and the output stability needs to be maintained during the cruising stage, while during the approach or low-speed standby period, the system should preferentially use shore power or energy storage power supply to reduce hydrogen consumption. Under this background, whether the operating condition characteristics can be dynamically identified according to the sailing state, and the power output of the fuel cell and energy storage system can be adjusted adaptively combined with load prediction, is the key to determine the economy and safety of the system operation.

[0004] On the other hand, the hydrogen supply process of the fuel cell system itself has complex fluid transport characteristics, especially during the change of the ship's motion state, the pressure, flow rate, and density of the hydrogen network present obvious dynamic characteristics. If an accurate modeling and scheduling mechanism for the hydrogen network cannot be established, problems such as pressure overshoot, insufficient flow, and hydrogen waste may occur, which may even affect the stable operation of the fuel cell. Therefore, it is necessary to include the hydrogen network as a dynamic subsystem into the overall scheduling optimization model to realize coordinated control of the whole process of "hydrogen supply-hydrogen use".

[0005] Existing researches mostly simplify the hydrogen system as a static supply module without considering the dynamic energy coupling characteristics between the hydrogen system and the power system; at the same time, the response scheduling research on the change of ship operating conditions is limited, and there is a lack of intelligent energy management methods integrating machine learning, predictive optimization, and strategy self-learning capabilities. In addition, the existing scheduling strategies are difficult to adapt to the uncertainty of the operating environment, lack of rolling optimization and long-term performance evolution mechanism, causing problems such as large system operation redundancy, high hydrogen consumption, and low battery charging and discharging efficiency.

[0006] Therefore, there is an urgent need for a new ship energy management method that can integrate hydrogen-electric energy network modeling, dynamic scheduling optimization, and artificial intelligence self-learning control. SUMMARY

[0007] To solve the above problems, the technical scheme adopted by the present application is as follows: a joint scheduling optimization method for a hydrogen-electric energy network of a fuel cell ship, comprising the following steps:

[0008] S1: identifying the working condition parameters in the ship running process, generating a working condition label, determining the value range and constraint conditions of the variables in the current working condition, and dynamically switching the scheduling model boundary conditions and strategy selection of the fuel cell in the hydrogen system;

[0009] S2: identifying the basic working condition parameters in the ship running process, coupling the influence of external speed and load, and establishing a steady-state flow relationship;

[0010] S3: based on the identified working condition parameters in the ship running process, predicting the load at the future time of the ship and determining the optimal strategy of the ship in the current state;

[0011] S4: constructing a scheduling optimization model with the minimum hydrogen consumption of the ship under different operating conditions, solving the scheduling optimization model, and obtaining the output power of the hydrogen system and the power system of the ship.

[0012] Further: the basic working condition parameters in the ship running process are used to couple the influence of external speed and load as follows:

[0013] A one-dimensional distributed parameter model is used to model the hydrogen flow, and the speed and load disturbance source term is considered:

[0014]

[0015]

[0016] Wherein: v g The flow rate of hydrogen in the pipeline, the density of hydrogen, the pressure of hydrogen, H i,t is the hydrogen consumption, FC is the index set of fuel cell equipment, D is the pipe diameter, S1 is the source term function for coupling the influence of external speed and load, and is defined as:

[0017]

[0018] Wherein: k1 is a correction coefficient, reflecting that when the ship accelerates or the load suddenly changes, the flow and pressure response of the hydrogen pipeline network will change rapidly, and the hydrogen storage valve and the fuel cell load need to be adjusted.

[0019] Further, the basic identified ship running process parameters, while introducing the speed and load dynamic disturbance term, establish a steady-state power flow relationship using the node admittance method, and the specific expression is as follows:

[0020]

[0021] Wherein: P i,t The active output of the i-th fuel cell unit at time t, θ m， θ n The phase angle of the node voltage of the power system, Y mn The conductance between nodes m and n, η v The gain of the speed disturbance on the power system, η p The gain of the load change disturbance on the power system, The load prediction value, Pit is the useful power output of the fuel cell unit.

[0022] Further, the scheduling optimization model includes an optimization objective function and a constraint condition;

[0023]

[0024] Wherein H i,t : hydrogen consumption, And : start-stop state, c i , α i , β i : optimization objective function weight,

[0025] α i represents the hydrogen consumption cost weight, c i represents the battery life loss weight, β i represents the equipment start-stop cost weight; with the condition label w t Dynamic assignment, w t ∈{port, departure, cruising, acceleration, deceleration};

[0026] The constraint condition includes: fuel cell output boundary, lithium battery charge and discharge power boundary, hydrogen pressure and flow boundary, and SOC range adjustment requirement;

[0027] The fuel cell output boundary includes: departure condition requirement: allowing to run close to maximum rated power;

[0028] Cruising condition requirement: limited to high efficiency interval; berthing / standby requirement: limited to zero or minimum output;

[0029] The lithium battery charge / discharge power boundaries include: acceleration condition: allowing high-rate discharge; deceleration condition: prioritizing charging to recover energy.

[0030] The hydrogen pressure and flow rate boundaries include: high load conditions: the lower pressure limit is relaxed; low load conditions: the pressure is maintained at a higher level for immediate response.

[0031] SOC range adjustment: When demand is high, the lower limit of SOC is allowed to be lowered, while when docking, the SOC is kept at a high value to ensure subsequent departure.

[0032] Furthermore, the formula used to predict the ship's load at future time points is as follows:

[0033]

[0034] Where: X t =[v t , P load,t , qt, pt, P, SOC t [,…], state input vector; f class (): Working condition identification function; f pred (): Load prediction function; ω t Operating condition labels: ∈{berthing, departure, cruise, acceleration, deceleration}; P load,t+1:t+H The predicted load sequence at time H;

[0035] Xt: System eigenvector at time t; Xt:t N: From time t The set of all feature vectors from N to time t, where Wt is the ship's operating condition label at the current time; Qt is the hydrogen mass flow rate; Pt is the hydrogen system pressure; P is the system's rated power or current operating power; and SOCt is the lithium battery's state of charge.

[0036] The optimization objective for determining the optimal strategy for the ship in its current state is expressed by the following formula:

[0037]

[0038] Where: s t Current system status (including operating condition labels, predicted load, hydrogen pressure, battery SOC, etc.); π (s t ) represents the optimal policy under the current state; π represents the family of policy functions; γ represents the reward discount factor; r t+k Instant reward at time t+k.

[0039] Furthermore: the ship's different operating conditions include berthing / standby, during which port shore power is connected, low load, and lithium battery power is given priority;

[0040] Sailing, high power ramping, fuel cell dominant;

[0041] Cruising, steady load, hydrogen and electricity synergy;

[0042] Acceleration or gear shifting, significant load fluctuation;

[0043] Deceleration and berthing, load reduction, hydrogen system unloading.

[0044] Further: the scheduling model boundary conditions of the fuel cell are as follows:

[0045]

[0046] Where: P i is the power of the i-th fuel cell; is the minimum value of the fuel cell power, is the maximum value of the fuel cell power;

[0047] Ramp rate R constraint:

[0048] Where: is the lower limit of the ramp rate, is the upper limit of the ramp rate;

[0049] Lithium battery energy storage power P bat Charge-discharge mutual exclusion logic:

[0050]

[0051]

[0052] Where: Pch is the lithium battery charging power Pdis is the lithium battery discharging power, which is the charge-discharge mutual exclusion logic;

[0053] Lithium battery energy storage SOC dynamic constraint:

[0054]

[0055]

[0056] Where: SOC is the state of charge of the lithium battery, is the pure power supply efficiency of the lithium battery, is the discharging efficiency, t is the time term.

[0057] Further, the hydrogen consumption H i,t of the i-th fuel cell is related to its electric power output P i,t

[0058] ​

[0059] wherein: η FC is the fuel cell efficiency, LHV H2 is the hydrogen lower heating value, δ H (ω t ) is the operating condition correction term.

[0060] A joint scheduling optimization device for a hydrogen-electric energy route network of a fuel cell ship, comprising:

[0061] An intelligent ship operating condition recognition module: for recognizing operating condition parameters in the ship driving process, generating an operating condition label, determining the value range and constraint conditions of variables in the current operating condition, and for dynamically switching the scheduling model boundary conditions and strategy selection of fuel cells in the hydrogen system;

[0062] A hydrogen-electric network calculation module: the hydrogen-electric network calculation module comprises:

[0063] A hydrogen network module: for basic recognition of operating condition parameters in the ship driving process, coupling the influence of external speed and load;

[0064] A power network power flow module: for recognizing operating condition parameters in the ship driving process, establishing a steady-state power flow relationship;

[0065] An artificial intelligence auxiliary module: for predicting the load of the ship at a future time based on the recognized operating condition parameters in the ship driving process and determining the optimal strategy of the ship in the current state;

[0066] A joint scheduling optimization module: for constructing a scheduling optimization model of the ship under different operating conditions to minimize hydrogen consumption, solving the scheduling optimization model, and obtaining the output power of the hydrogen system and the power system of the ship.

[0067] A hydrogen-electric energy route network system for a fuel cell ship, comprising

[0068] A hydrogen battery system: for providing power to the ship based on hydrogen fuel cells;

[0069] A power system: for storing and converting electrical energy into mechanical energy for ship driving based on the electrical energy provided by the hydrogen battery system;

[0070] A sensor system: for collecting operating condition parameters in the ship driving process;

[0071] Based on the operating condition parameters in the ship driving process transmitted by the sensor system, a joint scheduling optimization device for a hydrogen-electric energy route network of a fuel cell ship is used to realize output power control of fuel cells in the hydrogen battery system and lithium battery packs in the power system under different operating conditions of the ship.

[0072] The application provides a joint scheduling optimization method, device and system for a hydrogen-electric energy route network of a fuel cell ship, which has fine energy flow coupling capability, real-time sensing, predictive decision-making and strategy adaptive capability, so as to effectively cope with the multi-objective operation requirements of the fuel cell ship under complex navigation conditions, and realize an efficient, safe and low-carbon green ship propulsion system. Based on the cooperative coupling of the hydrogen network, the power network and the artificial intelligence scheduling model, the optimal joint scheduling of hydrogen energy distribution, electric power coordination, energy storage control and strategy learning of the ship under different navigation conditions is realized, the operation efficiency of the ship is improved, the energy consumption and hydrogen consumption are reduced, and the condition recognition, load prediction, real-time regulation and self-learning capability are realized, so as to realize the cooperative optimization scheduling and economic operation of the hydrogen-electric multi-energy system. BRIEF DESCRIPTION OF DRAWINGS

[0073] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0074] Figure 1 is a flow chart of the system of the present application;

[0075] Figure 2 is a flow chart of the method of the present application. DETAILED DESCRIPTION

[0076] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict, and the present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0077] In order to make the purpose, technical scheme and advantages of the embodiments of the present application more clear, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. The description of the at least one exemplary embodiment is actually only illustrative, but not as any limitation on the present application and its application or use. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0078] A joint scheduling optimization method for a hydrogen-electric energy route network of a fuel cell ship, comprising the following steps:

[0079] S1: identifying the working condition parameters in the ship running process, generating a working condition label, determining the value range and constraint conditions of the variables in the current working condition, and used for dynamically switching the scheduling model boundary conditions and strategy selection of the fuel cell in the hydrogen system;

[0080] S2: identifying the working condition parameters in the ship running process, coupling the influence of external speed and load, and establishing a steady-state flow relationship;

[0081] S3: based on the identified working condition parameters in the ship running process, predicting the load of the ship at a future time and determining the optimal strategy of the ship in the current state;

[0082] S4: constructing a scheduling optimal model of the ship under different operating conditions to minimize hydrogen consumption, solving the scheduling optimal model, and obtaining the output power of the hydrogen system and the power system of the ship.

[0083] The steps S1 / S2 / S3 / S4 are sequentially executed;

[0084] The basic identified working condition parameters in the ship running process are used to couple the influence of external speed and load in the following process:

[0085] A one-dimensional distributed parameter model is used to model hydrogen flow, and the speed and load disturbance source term is considered:

[0086]

[0087]

[0088] Wherein: v g Hydrogen flow rate in the pipeline, ρ Hydrogen density, p Hydrogen pressure, H i,t Hydrogen consumption, FC Fuel cell equipment index set, D Pipe diameter, S1 Source function for coupling the influence of external speed and load, defined as:

[0089]

[0090] Wherein: k1 is a correction coefficient, reflecting that when the ship accelerates or the load suddenly changes, the flow and pressure response of the hydrogen pipeline network will change rapidly, and the hydrogen storage valve and fuel cell load need to be adjusted.

[0091] The basic identified working condition parameters in the ship running process simultaneously introduce the speed and load dynamic disturbance term, and the node admittance method is used to establish the steady-state flow relationship, and the specific expression is as follows:

[0092]

[0093] Wherein: P i,tActive output of the i-th fuel cell unit at time t, θ m， θ n Phase angle of node voltage of the power system, Y mn Conductance between nodes m and n, η v Gain of the power system to the speed disturbance, η p Gain of the power system to the load change disturbance, Load prediction value, P it Useful work output of the fuel cell unit;

[0094] The scheduling optimization model comprises an optimization objective function and a constraint condition;

[0095]

[0096] Wherein H i,t : hydrogen consumption, And : start-stop state, c i , alpha i , beta i : optimization objective function weight,

[0097] Alpha i represents the hydrogen consumption cost weight, c i represents the battery life loss weight, beta i represents the equipment start-stop cost weight; with a working condition label w t Dynamic assignment, w t E {port, sailing, cruising, acceleration, deceleration};

[0098] The constraint condition comprises: fuel cell output boundary, lithium battery charging and discharging power boundary, hydrogen pressure and flow boundary, and SOC range adjustment requirement;

[0099] The fuel cell output boundary comprises: sailing working condition requirement: allowing to run close to maximum rated power;

[0100] Cruising working condition requirement: limited in high efficiency interval; berthing / standby requirement: limited to zero or minimum output;

[0101] The lithium battery charging and discharging power boundary comprises: acceleration working condition: allowing high rate discharge; deceleration working condition: preferentially charging to recover energy;

[0102] The hydrogen pressure and flow boundary comprises: high load working condition: lower limit of pressure is relaxed; low load working condition: pressure is kept higher to respond at any time;

[0103] SOC range adjustment: allow SOC lower limit to decrease when high demand, keep high value when in port to ensure subsequent sailing.

[0104] The formula for predicting the load of the ship at a future time is as follows:

[0105]

[0106] Where: X t = [v t , P load,t , qt, pt, P, SOC t ,…], state input vector; f class (): operating condition identification function; f pred (): load prediction function; ω t ∈{port, sailing, cruising, acceleration, deceleration} operating condition label; P load,t+1:t+H is the predicted load sequence at future H time;

[0107] X t : system feature vector at time t; X t : t N: all feature vectors from time t N to time t, W t is the current operating condition label of the ship; Qt is the hydrogen mass flow rate, Pt is the hydrogen system pressure, P is the rated power or current operating power of the system, SOC t is the state of charge of the lithium battery;

[0108] The optimization objective of the optimal strategy of the ship in the current state is as follows:

[0109]

[0110] Where: s t current system state (including operating condition label, predicted load, hydrogen pressure, battery SOC, etc.); π (s t ) is the optimal strategy under the current state; π is the policy function family; γ is the reward discount factor; r t+k is the immediate reward at time t+k; this formula is the policy optimization objective of reinforcement learning, suitable for DQN, PPO, etc.

[0111] The different operating conditions of the ship include berthing / standby, when the port shore power is connected, the load is low, and the lithium battery is preferentially powered; sailing, when the high power climbs, the fuel cell dominates; cruising, when the load is stable, hydrogen and electricity are coordinated; acceleration or variable speed, when the load fluctuation is significant; deceleration and berthing, when the load is reduced, and the hydrogen system is unloaded.

[0112] The scheduling model boundary conditions of the fuel cell are as follows:

[0113]

[0114] Where: P i is the power of the i-th fuel cell; is the minimum value of the fuel cell power, is the maximum value of the fuel cell power;

[0115] The ramp rate R constraint:

[0116] Where: is the lower limit of the ramp rate, is the upper limit of the ramp rate;

[0117] The lithium battery energy storage power P bat Charge-discharge exclusion logic:

[0118]

[0119]

[0120] Where: Pch is the lithium battery charging power Pdis is the lithium battery discharging power, which is the charge-discharge exclusion logic;

[0121] Lithium battery energy storage SOC dynamic constraint:

[0122]

[0123]

[0124] Where: SOC is the state of charge of the lithium battery, is the pure power supply efficiency of the lithium battery, is the discharging efficiency, t is the time term.

[0125] The hydrogen consumption H i,t of the i-th fuel cell and its electric power output P i,t are related as:

[0126]

[0127] Where: η FC is the fuel cell efficiency, LHV H2 is the low heat value of hydrogen, δ H (ω t ) is the working condition correction term.

[0128] A joint scheduling optimization device for a hydrogen-electric energy path network of a fuel cell ship, comprising:

[0129] A ship working condition intelligent identification module: used for identifying working condition parameters in the ship running process, generating working condition labels, establishing scheduling boundaries and weights under the current working condition, and used for dynamically switching scheduling model boundary conditions and strategy selection of the fuel cell in the hydrogen system; the working condition identification result determines the output boundary, start-stop state and dynamic weight of the target function of each device in the system; the ship working condition intelligent identification module dynamically identifies the current running working condition ωt based on multiple parameter information such as speed, load change rate and acceleration, such as starting, cruising, berthing, accelerating and decelerating;

[0130] A hydrogen-electric network calculation module: the hydrogen-electric network calculation module comprises:

[0131] A hydrogen network module: used for basic identification of working condition parameters in the ship running process, coupling the influence of external speed and load; the hydrogen mass flow in the hydrogen network is determined by the actual output of the fuel cell, and constitutes a hydrogen consumption source term;

[0132] A power network power flow module: used for identifying working condition parameters in the ship running process, establishing a steady-state power flow relationship;

[0133] An artificial intelligence auxiliary module: used for predicting the load of the ship at a future time based on the identified working condition parameters in the ship running process and determining the optimal strategy of the ship in the current state;

[0134] A joint scheduling optimization module: used for constructing an optimization objective function of the minimum hydrogen consumption of the ship under different running working conditions, solving the optimization objective function, obtaining the output power of the hydrogen system and the power system of the ship, and directly feeding the strategy and prediction result output by the artificial intelligence auxiliary + joint scheduling optimization module into the optimization model, and correcting the mathematical model, i.e. the hydrogen network model (gas flow balance, pressure constraint, hydrogen consumption and power relationship), the power network model (power flow equation, power balance, line constraint), the coupling constraint (fuel cell power Hydrogen consumption SOC Battery charge and discharge limit); target function (multi-objective weight changes with working condition); does not include the artificial intelligence auxiliary module.

[0135] A hydrogen-electric energy path network system for a fuel cell ship, comprising

[0136] A hydrogen battery system: used for providing power to the ship based on a hydrogen fuel cell;

[0137] A hydrogen system: comprising a high-pressure hydrogen storage tank, a delivery pipeline and a fuel cell module;

[0138] The high-pressure hydrogen storage tank delivers hydrogen to the fuel cell module through the delivery pipeline;

[0139] Power system: for providing electrical energy through lithium battery energy storage and converting electrical energy into mechanical energy for ship travel based on the hydrogen battery system provided; the load demand of the power system affects the energy storage scheduling and fuel cell output, indirectly changing the hydrogen consumption;

[0140] Power system: including lithium battery, electric propulsion system, shore power interface, converter and shipboard load;

[0141] Sensor system: for collecting working condition parameters during ship travel;

[0142] Based on the working condition parameters of the ship during travel transmitted by the sensor system, a joint scheduling optimization device for a hydrogen-electric energy route network of a fuel cell ship is used to realize the output power control of the fuel cell in the hydrogen battery system and the lithium battery in the power system under different operating conditions of the ship.

[0143] Embodiment 1:

[0144] The hydrogen-electric energy route network system of the present application is applied to a medium-sized fuel cell propulsion ship, such as Figure 1 The system mainly consists of the following modules:

[0145] Hydrogen system: including high-pressure hydrogen storage bottle, multiple branch pipes, pressure reducing valve, mass flow meter, cooling unit, control solenoid valve and multiple fuel cell systems (PEMFC);

[0146] Power system: including multi-stage lithium battery, energy management unit (BMS), motor driver, shore power interface module, converter (DC / DC, DC / AC);

[0147] Sensor network: including speed sensor, GPS, load meter, voltage / current sensor, pressure sensor, SOC monitoring unit;

[0148] Control and communication system: for collecting data, transmitting model parameters, executing scheduling results and feeding back state information;

[0149] Scheduling optimization control platform: deployed in the shipboard control unit, embedded with hydrogen-electric network modeling, optimization scheduling algorithm and artificial intelligence scheduling model.

[0150] A joint scheduling optimization device for a hydrogen-electric energy route network of a fuel cell ship, comprising:

[0151] Ship working condition intelligent identification module: used for identifying working condition parameters in the ship running process, generating working condition labels, determining scheduling boundaries and weights under the current working condition (i.e. determining the value range and constraint conditions of the variables under the current working condition), and used for dynamically switching the scheduling model boundary conditions and strategy selection of the fuel cell in the hydrogen system; the working condition identification result determines the output boundary, start-stop state and target function dynamic weight of each device in the system;

[0152] This module establishes the scheduling boundary and weight under the current working condition based on machine learning identification of working conditions. Machine learning identification is based on the following line for pre-update: working condition → boundary / weight → optimized

[0153] Working condition identification (classifier): input: historical time window features Xt N:t, output: working condition label {berthing, sailing, cruising, acceleration / deceleration, …}

[0154] Load prediction Pload;

[0155] Generate scheduling boundaries (fuel cell output / climbing limit, battery SOC and charge / discharge power limit, hydrogen pressure / flow safety domain) from working conditions

[0156] Generate target weights from working conditions;

[0157] Hydrogen-electric network calculation module: the hydrogen-electric network calculation module comprises:

[0158] Hydrogen network module: used for basic identification of working condition parameters in the ship running process, coupling the influence of external speed and load; the hydrogen mass flow in the hydrogen network is determined by the actual output of the fuel cell, constituting a hydrogen consumption source term;

[0159] Power network flow module: used for identifying working condition parameters in the ship running process, establishing a steady-state flow relationship;

[0160] Artificial intelligence auxiliary module: used for predicting the load of the ship at a future time based on the identified working condition parameters in the ship running process and determining the optimal strategy of the ship under the current state;

[0161] Joint scheduling optimization module: used for constructing a scheduling optimization model of the ship under different operating conditions to minimize hydrogen consumption, solving the scheduling optimization model to obtain the output power of the hydrogen system and the power system of the ship, and the strategy and prediction results output by the intelligent scheduling module are directly input into the optimization model and the mathematical model is corrected in the reverse direction through running feedback.

[0162] Based on the output power, update the relevant parameters in the above hydrogen-electric network calculation module, such as actual power, hydrogen consumption, pressure, SOC, temperature, etc.

[0163] (1) Working condition identification module

[0164] The input variables of the module include the ship speed v t , load fluctuation rate γ t , acceleration a t , sea state, etc. The convolutional neural network + GRU structure is used to realize the classification of ship operation conditions, and the output ω t ∈{berthing / standby, sailing, cruising, speed change, port speed reduction}.

[0165] (2) Hydrogen-electric network modeling module

[0166] a) Hydrogen network module: a one-dimensional distributed parameter model is used to model hydrogen flow, and the speed and load disturbance source term is considered:

[0167]

[0168]

[0169] Wherein: v g The flow rate of hydrogen in the pipeline, the density of hydrogen, the pressure of hydrogen, H1,t is the hydrogen consumption, FC is the fuel cell device index set, D is the pipe diameter, S1 is the source function, which is used to couple the external speed and load influence, and is defined as:

[0170]

[0171] k1 is the correction coefficient, this module reflects that when the ship accelerates or the load suddenly changes, the flow and pressure response of the hydrogen pipeline network will change rapidly, and the hydrogen storage valve and fuel cell load need to be adjusted.

[0172] b) Power network flow module

[0173] The power module uses the node admittance method to establish the steady-state power flow relationship, and introduces the speed and load dynamic disturbance term, which is described as follows:

[0174]

[0175] Wherein: P i,t The active output of the i-th fuel cell power supply unit at time t, θm, θn is the phase angle of the node voltage of the system, Ymn is the conductance between nodes m and n. η v The gain of speed disturbance on the power system, η p The gain of load change disturbance on the power system.

[0176] The scheduling optimal model includes an optimization objective function and constraint conditions;

[0177]

[0178] wherein H i,t : hydrogen consumption, and : start-stop state, c i , a i , b i : optimization objective function weight,

[0179] a i represents the hydrogen consumption cost weight, c i represents the battery life loss weight, b i represents the device start-stop cost weight; with a working condition tag w t dynamic assignment, w t e {port, departure, cruising, acceleration, deceleration};

[0180] In the objective function, ci, ai, bi are optimization objective function weights, which are weight coefficients in the optimization objective function, indicating the priority settings of each optimization objective under different working conditions.

[0181] The weight distribution under different working conditions can be different:

[0182] departure → low hydrogen consumption priority (power guarantee is the main)

[0183] cruising → high hydrogen consumption priority (economic operation is the main)

[0184] port → high battery life weight (reduce frequent cycling)

[0185] acceleration → low device start-stop weight (fast response is the main)

[0186] In this way, the model will automatically make different scheduling strategies under different working conditions.

[0187] The constraint conditions include: fuel cell output boundary, lithium battery charging and discharging power boundary, hydrogen pressure and flow boundary, and SOC range adjustment requirement;

[0188] The fuel cell output boundary includes:

[0189] Departure working condition requirement: allow to run close to maximum rated power; (for example 0.8~1.0 Pmax);

[0190] Cruising working condition requirement: limit to high efficiency interval (for example 0.4~0.7 Pmax); port / standby requirement: limit to zero or minimum power;

[0191] The lithium battery charging and discharging power boundary includes: acceleration working condition: allow high rate discharge; deceleration working condition: prefer to charge to recover energy;

[0192] The hydrogen pressure and flow rate boundaries include: high load conditions: the lower pressure limit is relaxed; low load conditions: the pressure is maintained at a higher level for immediate response.

[0193] SOC range adjustment: During periods of high demand, the lower limit of SOC is allowed to be lowered (e.g., from 20% to 15%), while the SOC remains high when docked to ensure subsequent departure. The "dispatch boundary" refers to the upper and lower limits of physical quantities such as power, flow rate, pressure, and SOC for equipment such as fuel cells, energy storage, grid interfaces, and hydrogen systems, and these upper and lower limits are dynamically adjusted according to operating conditions.

[0194] The formula used to predict the ship's load at future time points is as follows:

[0195]

[0196] Where: X t =[v t , P load,t , qt, pt, P, SOC t [,…], state input vector; f class (): Working condition identification function; f pred (): Load prediction function; ω t Operating condition labels: ∈{berthing, departure, cruise, acceleration, deceleration}; P load,t+1:t+H The predicted load sequence at time H;

[0197] Xt: System eigenvector at time t; Xt:t N: From time t The set of all feature vectors from N to time t, Wt is the label of the ship's operating condition at the current time; Qt is the hydrogen mass flow rate; Pt is the hydrogen system pressure; P is the system rated power or current operating power; SOCt is the state of charge of the lithium battery.

[0198] The optimization objective for determining the optimal strategy for the ship in its current state is expressed by the following formula:

[0199]

[0200] Where: s t Current system status (including operating condition labels, predicted load, hydrogen pressure, battery SOC, etc.); π (st) represents the optimal policy under the current state; π represents the family of policy functions; γ represents the reward discount factor; r t+k The immediate reward at time t+k; this formula represents the policy optimization objective of reinforcement learning, and algorithms such as DQN and PPO are used to solve for the optimization objective of the optimal policy.

[0201] Different operating conditions of the ship include berthing / standby, when the port shore power is connected, low load, and lithium battery priority power supply;

[0202] Sailing, when high-power climbing, fuel cell dominant;

[0203] Cruising, when the load is stable, hydrogen and electricity are coordinated;

[0204] Acceleration or gear change, when the load fluctuation is significant;

[0205] Deceleration and berthing, when the load is reduced, and the hydrogen system is unloaded.

[0206] The boundary conditions of the scheduling model of the fuel cell are as follows:

[0207]

[0208] Where: P i is the power of the i-th fuel cell; is the minimum value of the fuel cell power, is the maximum value of the fuel cell power;

[0209] Climbing rate R constraint:

[0210] Where: is the lower limit of the climbing rate, is the upper limit of the climbing rate;

[0211] Lithium battery energy storage power P bat Charge-discharge exclusion logic:

[0212]

[0213]

[0214] Where: Pch is the charging power of the lithium battery Pdis is the discharging power of the lithium battery, which is the charge-discharge exclusion logic;

[0215] Lithium battery energy storage SOC dynamic constraint:

[0216]

[0217]

[0218] Where: SOC is the state of charge of the lithium battery, is the pure power supply efficiency of the lithium battery, is the discharging efficiency, and t is the time term.

[0219] Further: the relationship between the hydrogen consumption Hi,t of the i-th fuel cell and its electric power output Pi,t is:

[0220]

[0221] wherein: η FC is the fuel cell efficiency, LHV H2 is the hydrogen low heating value, δ H (ω t ) is the operating condition correction term.

[0222] Figure 2 is the system startup, complete state acquisition and model initialization;

[0223] machine learning identifies the operating condition, establishes the scheduling boundary and scheduling weight under the current operating condition;

[0224] roll forward to predict future load trends, input to the optimization module;

[0225] the optimization module solves the scheduling scheme, and outputs the power value of each device;

[0226] the control system executes the scheduling instruction, and feeds back the actual running data;

[0227] the reinforcement learning module collects feedback and updates the model weight.

[0228] Typical scheduling scenario examples include the following:

[0229] Sailing acceleration operating condition: after identifying the high load acceleration state, the scheduling preferentially increases the fuel cell output, and part of the battery is used for collaborative power supply, while the SOC limit is restricted to ensure the subsequent cruising margin;

[0230] Berthing deceleration operating condition: automatically switch to shore power supply mode, fuel cell shutdown, and battery supplement short-time power demand;

[0231] Cruise operating condition: the fuel cell is scheduled to work in the optimal efficiency interval, and the battery is lightly adjusted to ensure system stability.

[0232] Operation performance improvement results

[0233] Energy utilization efficiency is improved. Through dynamic collaborative scheduling of hydrogen-electric system, the fuel cell always runs in the high efficiency interval, and the overall energy efficiency of the system can be improved by about 6%-12%.

[0234] Hydrogen consumption is significantly reduced. Hydrogen consumption model and optimization strategy linkage adjustment effectively avoids excessive output and redundant hydrogen supply, and the average single trip consumption of hydrogen can be reduced by 8%-15%.

[0235] Battery life is extended. The dynamic boundary regulation and charge-discharge mutual exclusion strategy of energy storage SOC reduces the frequent high-power cycle of the battery, and the battery cycle life is expected to be improved by 10%-20%.

[0236] Scheduling response and stability are improved

[0237] The response time of working condition switching is shortened, the working condition recognition based on machine learning, and the scheduling boundary adjustment time of the ship switching from "standby" to "sailing" or "acceleration" state is more than 70% faster than the artificial rule.

[0238] The load fluctuation suppression capability is enhanced, the prediction module predicts the load change 10-20 steps in advance, the scheduling response is intervened in advance, and the load fluctuation rate can be reduced by more than 25%.

[0239] The shore power and the on-board power source are coordinated to improve the efficiency, the system intelligently switches the shore power main supply and unloads the fuel cell and battery load when the ship is in port, the shore power access efficiency is improved by about 15%, and the redundant battery discharge is avoided.

[0240] The intelligent and adaptive results are as follows

[0241] The reinforcement learning scheduling strategy optimization effect is obvious, and the scheduling strategy trained by using the DQN or PPO algorithm can reduce the system comprehensive operation cost by about 9%-18% compared with the static strategy in multiple navigation cycles.

[0242] The adaptability of multi-working condition full life cycle scheduling is improved, the modular modeling + AI boundary response makes the system automatically adapt to five typical navigation working conditions, without manual switching of the control mode, and supports all-weather operation.

[0243] The scheduling system self-learning ability forms a closed loop, and the scheduling system can automatically update the strategy parameters according to the feedback indicators (such as hydrogen consumption deviation and SOC change rate), so as to realize scheduling self-iteration and long-term optimization.

[0244] The system model structure is suitable for 200kW-2MW fuel cell hybrid power ship platform, and is suitable for various types of ships such as inland ships, ferries and port ships.

[0245] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A joint scheduling optimization method for hydrogen-electricity circuit networks used in fuel cell ships, characterized in that: Includes the following steps: S1: Identify the operating parameters during the ship's operation, generate operating condition labels, determine the value range and constraints of variables under the current operating condition, and use them to dynamically switch the boundary conditions and strategy selection of the scheduling model for fuel cells in the hydrogen system. The boundary conditions for the scheduling model of the fuel cell are as follows: Where: P i Let i be the power of the i-th fuel cell; This represents the minimum power output of the fuel cell. P represents the maximum power of the fuel cell; i,t It is the useful work output of the i-th fuel cell unit at time t; Slope rate R constraint: in: This represents the lower limit of the gradient rate. This represents the upper limit of the gradient rate. Lithium battery energy storage power P bat Charge / discharge mutual exclusion logic: Where: Pch is the lithium battery charging power and Pdis is the lithium battery discharging power, which is a charging and discharging mutual exclusion logic; Lithium-ion battery energy storage SOC dynamic constraints: Where: SOC refers to the state of charge of the lithium battery. For the pure power supply efficiency of lithium batteries The discharge efficiency is given by t, which is the time term. S2: Based on the identified operating parameters during the ship's navigation process, the influence of external speed and load is coupled to establish a steady-state tidal current relationship; S3: Based on the identified operating parameters during the ship's operation, predict the ship's load at future moments and determine the optimal strategy for the ship in the current state. S4: Construct an optimal scheduling model that minimizes hydrogen consumption of the ship under different operating conditions, solve the optimal scheduling model, and obtain the output power of the ship's hydrogen system and electric system; The optimal scheduling model includes an optimization objective function and constraints. Where H i,t Hydrogen consumption and Start / stop status, c i α i β i : Optimize the weights of the objective function, α i c represents the hydrogen consumption cost weight. i β represents the weighting of battery life loss. i Indicates the weight of equipment start-up and shutdown costs; based on operating condition labels. w t Dynamic assignment, w t ∈{berthing, departure, cruising, acceleration, deceleration}; The constraints include: fuel cell output boundary, lithium battery charge and discharge power boundary, hydrogen pressure and flow boundary, and SOC range adjustment requirements; The fuel cell output limits include: start-up operating conditions: allowing operation close to the maximum rated power; Cruise operating conditions requirements: limited to the high-efficiency range; berthing / standby requirements: limited to zero or minimum output; The lithium battery charge / discharge power boundaries include: acceleration condition: allowing high-rate discharge; deceleration condition: prioritizing charging to recover energy. The hydrogen pressure and flow rate boundaries include: high load conditions: the lower pressure limit is relaxed; low load conditions: the pressure is maintained at a higher level for immediate response. SOC range adjustment: When demand is high, the lower limit of SOC is allowed to be lowered, while when docking, the SOC is kept at a high value to ensure subsequent departure.

2. The joint scheduling optimization method for hydrogen-electricity circuit network of fuel cell ships according to claim 1, characterized in that: The process of coupling the basic identified ship operating parameters during navigation with the influence of external speed and load is as follows: Hydrogen flow is modeled using a one-dimensional distributed parameter model, taking into account the disturbances from speed and load: in: v g Hydrogen flow rate in the pipe, ρ hydrogen density, p hydrogen pressure, H i,t For hydrogen consumption, the set of FC fuel cell device indices, D is the pipe diameter, and S1 is the source term function used to couple the effects of external speed and load, defined as: Where: k1 is a correction coefficient, reflecting the rapid changes in the flow and pressure response of the hydrogen pipeline network when the ship accelerates or experiences sudden load changes, requiring adjustment of the hydrogen storage valves and fuel cell load. Indicates ship speed.

3. The joint scheduling optimization method for hydrogen-electricity circuit network of fuel cell ships according to claim 1, characterized in that: The basic identified operating parameters of the ship during its navigation process are combined with dynamic disturbance terms of speed and load. The steady-state current relationship is established using the nodal admittance method, and the specific expression is as follows: Where: P i,t θ is the useful work output of the i-th fuel cell unit at time t. m， θ n Y is the node voltage phase angle of the power system. mn Let n be the electrical conductance between nodes m and n. η v For the gain of the power system due to speed disturbance, η p The gain of the electrical system due to load change disturbances This is the load forecast value.

4. The joint scheduling optimization method for hydrogen-electricity circuit network of fuel cell ships according to claim 1, characterized in that: The formula used to predict the ship's load at future times is as follows: Where: X t =[v t , P load,t , qt, pt, P, SOC t [,…], state input vector; f class (): Working condition identification function; f pred (): Load prediction function; ω t Operating condition labels: ∈{berthing, departure, cruise, acceleration, deceleration}; P load,t+1:t+H The predicted load sequence at time H; Xt: System eigenvector at time t; Xt:t N: From time t The set of all feature vectors from N to time t, where Wt is the ship's operating condition label at the current time; Qt is the hydrogen mass flow rate; Pt is the hydrogen system pressure; P is the system's rated power or current operating power; and SOCt is the lithium battery's state of charge. The optimization objective for determining the optimal strategy for the ship in its current state is expressed by the following formula: Where: s t Current system status includes operating condition tags, predicted load, hydrogen pressure, and battery SOC; π (s t ) represents the optimal policy under the current state; π represents the family of policy functions; γ represents the reward discount factor; r t+k Instant reward at time t+k.

5. The joint scheduling optimization method for hydrogen-electricity circuit network of fuel cell ships according to claim 1, characterized in that: The ship's different operating conditions include berthing / standby, during which port shore power is connected, low load, and lithium battery power is given priority. At the start of the journey, high-power ramp-up and fuel cell dominance are key. Cruise mode, during which the load is stable and hydrogen and electricity work together; Acceleration or speed change causes significant load fluctuations. As the vehicle slows down and docks, the load decreases and the hydrogen system is unloaded.

6. The joint scheduling optimization method for hydrogen-electricity circuit network of fuel cell ships according to claim 1, characterized in that: The hydrogen consumption H of the i-th fuel cell i,t Its electrical power output P i,t The relationship is: in: For fuel cell efficiency, LHV H2 The lower heating value of hydrogen, δ H (ω t ) represents the operating condition correction item.

7. A joint scheduling optimization device for a hydrogen-electricity circuit network for fuel cell ships, characterized in that: include: Ship operating condition intelligent identification module: used to identify the operating condition parameters of the ship during operation, generate operating condition labels, determine the value range and constraints of variables under the current operating condition, and dynamically switch the boundary conditions and strategy selection of the scheduling model of fuel cells in the hydrogen system. The boundary conditions for the scheduling model of the fuel cell are as follows: Where: P i Let i be the power of the i-th fuel cell; This represents the minimum power output of the fuel cell. P represents the maximum power of the fuel cell; i,t It is the useful work output of the i-th fuel cell unit at time t; Slope rate R constraint: in: This represents the lower limit of the gradient rate. This represents the upper limit of the gradient rate. Lithium battery energy storage power P bat Charge / discharge mutual exclusion logic: Where: P ch Lithium battery charging power P dis This represents the discharge power of the lithium battery; this is a charge / discharge mutual exclusion logic. Lithium-ion battery energy storage SOC dynamic constraints: Where: SOC refers to the state of charge of the lithium battery. For the pure power supply efficiency of lithium batteries The discharge efficiency is given by t, which is the time term. Hydrogen-electric network calculation module: The hydrogen-electric network calculation module includes: Hydrogen network module: used for basic identification of ship operating parameters during navigation, coupled with the influence of external speed and load; Power flow module: Used to identify the operating parameters of the ship during its navigation process and establish steady-state power flow relationships; Artificial intelligence-assisted module: used to predict the ship's load at future moments and determine the optimal strategy for the ship in the current state based on the identified operating parameters during the ship's operation. Joint scheduling optimization module: used to construct the optimal scheduling model that minimizes hydrogen consumption of the ship under different operating conditions, solve the optimal scheduling model to obtain the output power of the ship's hydrogen system and electric system; The optimal scheduling model includes an optimization objective function and constraints. Where H i,t Hydrogen consumption and Start / stop status, c i α i β i : Optimize the weights of the objective function, α i c represents the hydrogen consumption cost weight. i β represents the weighting of battery life loss. i Indicates the weight of equipment start-up and shutdown costs; based on operating condition labels. w t Dynamic assignment, w t ∈{berth, departure, cruise, acceleration, deceleration}; The constraints include: fuel cell output boundary, lithium battery charge and discharge power boundary, hydrogen pressure and flow boundary, and SOC range adjustment requirements; The fuel cell output limits include: start-up operating conditions: allowing operation close to the maximum rated power; Cruise operating conditions requirements: limited to the high-efficiency range; berthing / standby requirements: limited to zero or minimum output; The lithium battery charge / discharge power boundaries include: acceleration condition: allowing high-rate discharge; deceleration condition: prioritizing charging to recover energy. The hydrogen pressure and flow rate boundaries include: high load conditions: the lower pressure limit is relaxed; low load conditions: the pressure is maintained at a higher level for immediate response. SOC range adjustment: When demand is high, the lower limit of SOC is allowed to be lowered, while when docking, the SOC is kept at a high value to ensure subsequent departure.

8. A hydrogen-electricity network system for fuel cell ships, characterized in that: include Hydrogen battery system: used to provide power to ships using hydrogen-based fuel cells; Power system: used for storing electrical energy provided by the hydrogen battery system through lithium battery packs and converting electrical energy into mechanical energy for ship propulsion; Sensor system: used to collect operating parameters of the ship during its operation; Based on the operating parameters of the ship during its voyage transmitted by the sensor system, the joint scheduling optimization device for the hydrogen-electric power circuit network of fuel cell ships as described in claim 7 is used to realize the output power control of the fuel cells in the hydrogen battery system and the lithium battery packs in the power system under different operating conditions of the ship.

Images