Energy management method and system for aviation hybrid power system

Abstract

The invention belongs to the technical field of new energy aircraft power systems, and provides an energy management method and system for an aviation hybrid power system, and the method comprises the steps: S1, constructing a power flow calculation model of the hybrid power system; s2, constructing a multi-objective optimization reward function on the basis of the fuel consumption quality, the pollutant emission quality and the battery health state; and S3, based on the constructed power flow calculation model of the hybrid power system and the multi-objective optimization reward function, outputting an energy management strategy of the hybrid power system under a TD3 algorithm architecture. The method has the advantages that the award value corresponding to each action value is calculated based on the multi-target TD3 algorithm, and the action value enabling the multi-target award function to be converged to the maximum is output, so that the energy management strategy capable of acting on each component of the hybrid power system is obtained, the fuel economy and reliability of the hybrid power system are improved, and the energy management efficiency is improved. And the endurance of the aircraft is increased.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy aircraft power systems, and in particular to an energy management method and system for an aviation hybrid power system. Background Art

[0002] At present, the pure electric power used by electric aircraft only supports a flight time of tens of minutes, which is far from meeting the needs of various long-duration flight missions. In order to solve the endurance problem, the aviation turbine generator hybrid system combines the advantages of aviation turbine engines' high energy density and electrification of power transmission pathways. Compared with the pure electric propulsion system, it can greatly increase the aircraft's load and flight time. It has broad equipment application prospects and has attracted the attention of various electric aircraft manufacturers around the world.

[0003] Energy management technology is a key technology to achieve the reasonable allocation of power requirements of different energy sources in aviation turbine power generation hybrid systems under complex flight conditions and to ensure the stable operation of the power system. It plays an important role in improving energy conservation and emission reduction benefits, increasing energy utilization, and extending the service life of power batteries. At present, energy management technology has been maturely applied in hybrid vehicles, but it is still in its infancy in the aviation field.

[0004] Energy management technology can be divided into rule-based strategies, optimization-based strategies, and learning-based strategies. Chinese patent CN117585170A uses a rule-based approach to design an energy management strategy. The disadvantage is that the rule setting is too dependent on manual experience and cannot guarantee the optimal working performance of each component of the hybrid system. In addition, the energy management strategy designed based on rules will produce a charge and discharge state of charge (SOC) boundary. When the hybrid system switches at this boundary, it is easy to cause the power distribution result to jump or even oscillate, affecting the normal operation of the hybrid system. Chinese patent CN115983109A uses a large amount of flight condition data to train a power prediction neural network model, and then uses a dynamic programming optimization strategy to obtain the optimal energy distribution strategy for the hybrid system under the flight range. The disadvantage is that deep learning is only applied to demand power prediction, and the dynamic programming energy management strategy, as a global optimization strategy, requires the prediction of all operating conditions of the flight range, and has poor adaptability. In addition, the strategy design only focuses on reducing the consumption of equivalent fuel, and does not consider the multi-objective optimization of the performance of other components at the same time. Summary of the invention

[0005] The object of the present invention is to provide an energy management method and system for an aviation hybrid power system. By adopting a deep reinforcement learning framework for hybrid power energy management based on TD3, and constructing a competitive Actor-Critic network architecture, flight condition variables are introduced into the state variables in the deep reinforcement learning network to increase the adaptability of the energy management strategy in the flight envelope, so as to solve the above problems.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows: An energy management method for an aviation hybrid power system, comprising the steps of: S1. Construct a power flow calculation model of the hybrid power system. The power flow calculation model includes a power flow model of the turbogenerator system and an equivalent circuit model of the power battery. The power flow model of the turbogenerator system represents the conversion relationship between power and torque between the aero-turbine engine and the generator in the hybrid power system; S2. Construct a multi-objective optimization reward function based on the fuel consumption mass, pollutant emission mass, and battery health state; S3. Based on the constructed power flow calculation model of the hybrid power system and the multi-objective optimization reward function, output the energy management strategy of the hybrid power system under the TD3 algorithm architecture.

[0007] Among them, the fuel consumption mass is formed by a real-time fuel consumption prediction model, and the real-time fuel consumption prediction model is obtained by training with a lightweight gradient boosting machine method based on a preset aero-turbogenerator system model.

[0008] The present invention also provides an energy management system for an aviation hybrid power system. The system includes a memory and at least one processor. The memory stores a computer program, and the processor is used to execute the computer program to implement the above-mentioned energy management method for the aviation hybrid power system.

[0009] Compared with the prior art, the present invention has at least the following beneficial effects: (1) The present invention establishes a power flow calculation model of the hybrid power system as the environment model of the TD3 optimization algorithm, and sets the state of charge (SOC) of the power battery, the required power (Preq), the altitude (H) and Mach number (Ma) of the aero-turbine engine as state variables, increasing the adaptability of the energy management optimization method in the flight envelope; (2) Design a multi-objective reward function fusion mechanism, and establish a real-time fuel consumption prediction model, a battery health state (SOH) degradation estimation model, a pollutant emission mass estimation model, etc. applicable to the aero-turbogenerator hybrid power system. Map heterogeneous objectives such as fuel consumption mass, pollutant emission mass, and battery performance to a learnable reward weight space, realizing multi-objective optimization considering the performance of other components, and making the energy management strategy take into account energy conservation, emission reduction performance and reliability. (3) Iteratively solve the multi-objective optimization problem through the TD3 algorithm to calculate the reward value corresponding to each action value. Its output obtains the action value that makes the multi-objective reward function converge to the maximum, thereby obtaining the energy management power distribution instruction that can act on each component of the hybrid power system, thereby improving the fuel economy and reliability of the hybrid power system and increasing the endurance of the aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0011] Figure 1 It is a schematic diagram of the hybrid power system and related system structures in the embodiments provided by the present invention; Figure 2 It is a schematic diagram of the step architecture of the energy management method for an aviation turbogenerator hybrid power system in the embodiments provided by the present invention; Figure 3 It is a schematic diagram of the component-level model of an aviation turbine engine in the embodiments provided by the present invention; Figure 4 It is a flowchart of the steps of the energy management optimization method based on the TD3 architecture in the embodiments provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] The technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0013] The following are specific embodiments of the present invention, and the technical solutions of the present invention will be further described in conjunction with the drawings, but the present invention is not limited to these embodiments.

[0014] In the present embodiment, the hybrid power system and related system structures corresponding thereto are as Figure 1 shown. The energy management controller obtains the required power of the propulsion system through the flight controller, executes the energy management strategy, and sends a power distribution instruction to the hybrid power system.

[0015] The hybrid power system includes a turbine engine, a generator, an AC / DC inverter, a battery, and a DC / DC converter. The turbine engine drives the generator to convert mechanical energy into electrical energy. The AC / DC inverter converts the alternating current generated by the generator into direct current and aggregates it into the DC power grid. The battery realizes "peak shaving and valley filling" of the propulsion system power through a bidirectional DC / DC converter. When the power demand of the propulsion system is greater than the efficient working power of the turbine power generation system, the battery provides supplementary power; when the power demand of the propulsion system is less than the efficient working power of the turbine power generation system, the battery stores the excess power of the turbine power generation system for charging, and the turbine power generation system does not need to frequently control the engine speed change.

[0016] On this basis, as Figure 2 shown, this embodiment proposes an energy management method for an aviation hybrid power system, including the steps of: S1. Construct a power flow calculation model of the hybrid power system. The power flow calculation model includes a turbine power generation system power flow model and a power battery equivalent circuit model. The turbine power generation system power flow model represents the conversion relationship between power and torque between the aviation turbine engine and the generator in the hybrid power system; S2. Construct a multi-objective optimization reward function based on fuel consumption mass, pollutant emission mass, and battery health state; S3. Based on the constructed power flow calculation model and multi-objective optimization reward function of the hybrid power system, output the energy management strategy of the hybrid power system under the TD3 algorithm framework.

[0017] Among them, in step S1, according to the component power flow principle, establish the relationship between power and torque from the aviation turbine engine to the generator. The turbine power generation system power flow model is: ; Among them, T eng is the engine shaft torque, P eng is the engine shaft power, T gen is the generator shaft torque, P gen is the generator shaft power, N is the engine and generator speed, and η gen is the generator efficiency.

[0018] The power battery equivalent circuit model is: In the discharge mode: ; In the charging mode: ; Battery SOC change: ; Among them, P bat is the power battery power. When in the discharge mode, P bat >0, and when in the charging mode, Pbat < 0; P req is the total required power of the hybrid system, P gen is the shaft power of the generator, η AC / DC is the efficiency of the AC / DC component, η DC / DC is the efficiency of the bidirectional DC / DC component; V oc is the open-circuit voltage of the battery, R0 is the internal resistance of the battery, Q ini is the initial battery charge, Q0 is the nominal total battery charge, I b is the battery current, t is the time.

[0019] In this embodiment, by establishing a power flow calculation model of the hybrid system as the environmental model of the TD3 optimization algorithm, the adaptability of the energy management optimization method in the flight envelope is increased.

[0020] In step S2, a multi-objective optimization reward function is established, comprehensively considering fuel consumption, pollutant emissions, and the state of health (SOH) of the battery. The form of the multi-objective optimization reward function is as follows: ; where α, β, γ are the weight coefficients of their respective corresponding terms, is the fuel consumption mass, ΔSOH is the percentage loss of the battery state of health, is the pollutant emission mass.

[0021] Furthermore, for calculating the fuel consumption of the aero-turbine power generation system, it first constructs an aero-turbine power generation system model, and generates a dataset for training a real-time prediction model of fuel consumption based on this model. The dataset covers the entire flight envelope operating conditions.

[0022] Thus, as Figure 3 shown, with the input quantities being altitude H, Mach number Ma, engine speed N, and the demand power P of the turbine discovery system e , and the output quantity being the fuel consumption rate W f of the aero-turbine engine component-level model is solved to obtain the fuel consumption mass of the aero-turbine power generation system.

[0023] The solution principle of the aero-turbine engine component-level model is as follows: ① Inlet It calculates the static temperature and static pressure of the undisturbed cross-section at the engine inlet according to altitude H and Mach number Ma, and further calculates the total temperature, total pressure, and air flow velocity at the inlet cross-section of the inlet, and finally obtains the total temperature T t2 and total pressure at the outlet of the inlet.

[0024] ② Compressor The function of the compressor is to increase the pressure of the air flow from the inlet duct to the combustion chamber so that the air flow can reach the combustion conditions. The calculation of the thermodynamic parameters of the air flow in the compressor depends on the component characteristic diagram. According to the current engine speed N and the total temperature T t2 the current reduced shaft speed N c can be calculated. Then, by introducing the R line in the compressor component characteristic diagram and based on the initial guess value R Line the current compressor efficiency η HPC , the converted flow rate W c2 and the pressure ratio π HPC can be obtained by using the two-dimensional interpolation method. Further calculations can obtain the total pressure at the compressor outlet and the actual flow rate W3. At the same time, based on the enthalpy h2 at the inlet section and the actual enthalpy h3 at the compressor outlet, the power consumption P HPC of the compressor is obtained as: P HPC = W3(h3 - h2).

[0025] ③ Combustion chamber The combustion chamber is a component where the high-pressure gas compressed by the high-pressure compressor is mixed and burned with the injected fuel, converting the chemical energy of the fuel into heat energy and kinetic energy. After adding fuel to the combustion chamber, the flow rate becomes W4 = W3 + W f , where W4 is the flow rate at the combustion chamber outlet.

[0026] ④ Turbine The turbine is a component that converts the high-temperature and high-pressure air flow into kinetic energy and generates work. The high-pressure and high-temperature air flow flowing out from the combustion chamber impacts the turbine, and the expansion work generated drives the compressor and the load (high-speed generator). Similar to the working principle of the compressor, based on the air flow characteristic parameters at the combustion chamber outlet, combined with the initial guess value of the pressure drop ratio π HPT of the turbine and the turbine characteristic diagram, the thermodynamic calculation of the turbine can be carried out, which will not be elaborated here. Finally, the work power P HPT of the turbine is obtained as: P = W5(h5 - h4). Where h4 is the enthalpy at the turbine inlet section and h5 is the actual enthalpy at the turbine outlet.

[0027] ⑤ Nozzle The turbine converts the work of the engine into the kinetic energy of the rotor, and the outlet air flow is discharged to the atmosphere through the nozzle and the exhaust device. Assuming that the fluid of the model is a compressible, subsonic unsteady flow, and the fluid does not do work or transfer heat when flowing in the flow channel with variable area, the flow rate W8 is calculated according to the outlet area.

[0028] ⑥ Controller The controller adopts a closed-loop proportional-integral (PI) controller. The inputs are the commanded speed N cmd and the actual speed N, and the output is the fuel consumption rate W f input to the engine model:[[]] ; where kp is the proportionality coefficient, k i is the integral coefficient, and z is the integral term.

[0029] ⑦ Simultaneous working equations and solutions The simultaneous working equations of the engine component-level model include three flow balance equations and one rotor dynamics equation. The flow balance equations are shown in Table 1.

[0030] Table 1 Specific correspondence of the flow balance equations

[0031] In the table, W C2 is the actual corrected flow rate of the compressor, and W C2map is the corrected flow rate of the compressor obtained by interpolation through the compressor characteristic map; W C4 is the actual corrected flow rate of the turbine, and W C4map is the corrected flow rate of the turbine obtained by interpolation through the turbine characteristic map; W0 is the inlet flow rate at the inlet of the intake duct.

[0032] The rotor dynamics equation is as follows: ; In the formula, J is the moment of inertia of the engine shaft, and P HPT is the work power of the turbine, and P HPC is the power consumed by the compressor, and P e is the power extracted by the generator from the engine shaft.

[0033] During the actual operation of an aero-turbine engine, the flow rates at each section (including the inlet flow rate W0 at the inlet of the intake duct) are difficult to measure directly. In order to understand the component characteristics of the engine through modeling, several initial guess values need to be assumed first. In this way, there are two calculation methods for the flow rates at each section. One is the actual W i (i = 1, 2,..., n) directly calculated from the inlet flow rate W0 at the inlet of the intake duct through the component flow balance, and the other is the W imap (i = 1, 2,..., n) obtained from the initial guess values according to the component characteristic map. For each flight condition, all the simultaneous equations are iteratively solved by numerical methods. When the error ε i (i = 1, 2,..., n) is large, the initial guess values are adjusted based on the numerical method. When the errors of all equations converge or reach the maximum number of iterations, that is, when W i is basically equal to W imap and the rotor acceleration of the engine shaft is 0, it is judged that the model converges to the steady state, and the fuel consumption rate at this time is output.

[0034] Then, based on the above aviation turbine power generation system model, a dataset for training the real-time fuel consumption prediction model is generated. During the generation of the dataset, based on the altitude H, Mach number Ma, engine speed N, and the required power P of the turbine power generation system e as input quantities, the thermodynamic parameters of each component are calculated through the above thermodynamic equations. Then, based on the results calculated by the common working equation, the initial guess value is updated, and the thermodynamic parameters of each component are recalculated using the updated initial guess value. When the constraint value ε i (i = 1, 2, 3, 4) is less than or equal to the preset value, the loop ends, and thus the predicted fuel consumption rate W under multiple flight conditions is obtained f to form a dataset.

[0035] Finally, the LightGBM (Lightweight Gradient Boosting Machine) method is adopted. Through the histogram algorithm to accelerate the calculation, data and feature processing are optimized using gradient-based one-sided sampling and exclusive feature bundling, and an efficient training is achieved by combining the decision tree strategy of growing by leaf. The real-time fuel consumption prediction model trained based on the dataset can be expressed as: ;

[0036] where is the fuel consumption mass.

[0037] In step S2, a degradation model characterizing the state of health (SOH) of the battery is constructed. The discrete transient change expression of SOH is as follows: ; where c is the battery discharge rate, ΔSOH is the percentage loss of the battery state of health, N k is the remaining total number of battery cycles; I is the magnitude of the power battery current, ΔC n is the percentage of capacity loss, is the pre-exponential factor positively correlated with the battery discharge rate, R is the ideal gas constant, Q t is the cumulative throughput of electricity, T b is the internal temperature of the battery, E b is the battery activation energy, which is negatively correlated with the discharge rate c. The more times the power battery working mode switches, the greater the transient discharge rate, which will cause the battery state of health (SOH) to degrade faster.

[0038] In step S2, the calculation principle of pollutant emission mass is as follows: The initial electric energy stored in the battery is regarded as equivalent fuel. To make the pollutant emissions caused by electric energy and fuel comparable, this embodiment tracks the emissions and consumption of these two energy sources throughout their life cycles. This means that the fuel consumption includes not only the emissions after combustion but also the emissions throughout the life cycle from production to consumption. The calculation of the initial power consumption also takes into account the same factors.

[0039] Thus, the calculation expression of pollutant emission mass is as follows: ; where, is the pollutant emission mass, is the CO2 emission mass, is the NO x emission mass, is the fuel consumption mass, is the electric energy consumed by the power battery.

[0040] In addition, in step S2, the weight coefficients are obtained through the normalization method to make each objective comparable.

[0041] In this embodiment, by designing a multi-objective reward function fusion mechanism, heterogeneous objectives such as fuel consumption mass, pollutant emission mass, and battery performance are mapped into a learnable reward weight space, realizing multi-objective optimization considering the performance of other components, and making the energy management optimization method take into account both energy conservation and emission reduction performance and reliability.

[0042] In step S3, this embodiment uses the TD3 algorithm architecture to design an energy management optimization method for an aero-turbine power generation hybrid system. Among them, the input state s and action a of the TD3 algorithm architecture are respectively defined as: ; a = P eng ; where, SOC is the state of charge of the power battery, H and Ma are respectively the altitude and Mach number of the aero-turbine engine, P req is the total demand power of the hybrid system, and P eng is the engine shaft power.

[0043] where, SOC is the state of charge of the power battery, H and Ma are respectively the altitude and Mach number of the aero-turbine engine, P req is the total demand power of the hybrid system, and P eng is the engine shaft power.

[0044] As Figure 4 shown, the specific calculation process of the energy management optimization method for the hybrid system based on the TD3 algorithm architecture is as follows: 1. Initialize the network Initialize the six network parameters of the TD3 algorithm architecture, namely the policy Actor network , the policy Critic1 network , the policy Critic2 network , the target Actor network , the target Critic1 network , the target Critic2 network , and initialize the buffer D for storing interaction data.

[0045] 2. Data sampling and storage According to the current state s t , the target Actor network generates an action , where ξ~N(0,σ) is Gaussian noise, and the noise range is clipped by clip(ξ,−c,c) (c represents the boundary value for truncating the action noise), and the action a is executed t , obtaining the reward r t and the next state s t+1 , and store (s t ,a t ,r t ,s t+1 ) into the buffer D.

[0046] 3. Policy dual Critic network update Sample a batch of data from the buffer D. When calculating the target Q value, take the minimum value of the dual target Critic network: ; where λ is the discount factor ≤1, and the larger λ is, the more attention is paid to long-term interests.

[0047] Minimize the mean square error of Critic: ; At the same time, use the stochastic gradient descent method to update the parameters of the current value network (θ i ).

[0048] 4. Delayed policy Actor network update Every d = 2 steps, update the policy Actor network by maximizing to avoid premature optimization due to the non-convergence of the policy Critic network: ; where ▽ is the gradient operation symbol.

[0049] 5. Soft update of the target network The target network is updated slowly with a soft update factor coefficient τ≪1 to improve training stability: 。

[0050] Based on the above energy management optimization method of the TD3 algorithm architecture, repeat processes 2 to 5 until the policy converges or reaches the maximum number of training rounds to obtain the energy management policy that maximizes the reward function for multi-objective optimization in step 2.

[0051] In this embodiment, the TD3 algorithm architecture combining dual Critic networks, delayed updates, and target policy smoothing is used to iteratively solve the multi-objective optimization problem. Based on the multi-objective TD3 algorithm, the reward value corresponding to each action value is calculated, and the action value that makes the multi-objective reward function converge to the maximum is output, so as to obtain the energy management power distribution instruction that can act on each component of the hybrid power system, thereby forming an energy management policy, which improves the fuel economy and reliability of the hybrid power system and increases the endurance of the aircraft.

[0052] This embodiment also provides an energy management system for an aviation hybrid power system. The system includes a memory and at least one processor. The memory stores a computer program, and the processor is configured to execute the computer program to implement the energy management method for an aviation hybrid power system as described above.

[0053] The specific embodiments described herein are merely illustrative of the spirit of the present invention. Those skilled in the art of the present invention can make various modifications or supplements to the described specific embodiments or use similar methods to replace them, but will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

[0054] It should be noted that all directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention are only used to explain the relative position relationship and movement conditions between components in a specific posture (as shown in the drawings). If the specific posture changes, the directional indications will also change accordingly.

[0055] In addition, in the present invention, descriptions such as "first", "second", and "one" are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0056] In the present invention, unless otherwise clearly specified or limited, terms such as "connection" and "fixation" shall be understood in a broad sense. For example, "fixation" may be a fixed connection, a detachable connection, or integrated; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two elements or the interaction relationship between two elements, unless otherwise clearly limited. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

Claims

1. An energy management method for an aviation hybrid power system, characterized in that, Including the steps: S1. Construct a power flow calculation model for the hybrid power system. The power flow calculation model includes a power flow model of the turbogenerator system and an equivalent circuit model of the power battery. The power flow model of the turbogenerator system represents the conversion relationship between power and torque between the aero-turbine engine and the generator in the hybrid power system; S2. Construct a multi-objective optimization reward function based on the fuel consumption mass, pollutant emission mass, and battery health state; S3. Based on the constructed power flow calculation model of the hybrid power system and the multi-objective optimization reward function, output the energy management strategy of the hybrid power system under the TD3 algorithm framework; Among them, the fuel consumption mass is formed by a real-time fuel consumption prediction model, and the real-time fuel consumption prediction model is obtained by training with a lightweight gradient boosting machine method based on a preset aero-turbogenerator system model.

2. The energy management method for an aviation hybrid power system according to claim 1, characterized in that In step S1, the power flow model of the turbogenerator system is: ； Where: T eng is the engine shaft torque, P eng is the engine shaft power, T gen is the generator shaft torque, P gen is the generator shaft power, N is the engine and generator speed, η gen is the generator efficiency.

3. The energy management method for an aviation hybrid power system according to claim 2, characterized in that, In step S1, the equivalent circuit model of the power battery is: In the discharge mode: ; During charging mode: ; Battery SOC change: ; Among them, P bat is the power of the power battery, P req is the total required power of the hybrid system, P gen is the shaft power of the generator, η AC / DC is the efficiency of the AC / DC component, η DC / DC is the efficiency of the bidirectional DC / DC component; V oc is the open-circuit voltage of the battery, R0 is the internal resistance of the battery, Q ini is the initial battery charge, Q0 is the nominal total battery charge, I b is the battery current, and t is the time.

4. The energy management method for an aviation hybrid power system according to claim 1, characterized in that In step S2, the component-level model of the aero-turbine engine with the input quantities of altitude H, Mach number Ma, engine speed N, and the required power P of the turbogeneration system e , and the output quantity of specific fuel consumption W f is solved to obtain the fuel consumption mass of the aero-turbogeneration system.

5. A method for energy management of an aviation hybrid power system according to claim 4, characterized in that, The steps to obtain the real-time fuel consumption prediction model are: Construct an aero-turbogenerator system model; Generate a dataset for training the real-time fuel consumption prediction model, and the dataset includes all working conditions of the flight envelope; Train based on the dataset using the lightweight gradient boosting machine method to obtain the real-time fuel consumption prediction model.

6. A method for energy management of an aviation hybrid power system according to claim 1, characterized in that, In step S2, the reward function of the pollutant emission mass is obtained according to the following formula, which is expressed as: ； Among them, is the pollutant emission mass, is the CO2 emission mass, is the NO x emission mass, is the fuel consumption mass, is the electricity consumed by the power battery.

7. A method for energy management of an aviation hybrid power system according to claim 1, characterized in that In step S2, the battery health state is obtained based on the degradation model of the battery health state, which is expressed as follows: ； where c is the battery discharge rate, ΔSOH is the percentage loss of battery health state, N k is the remaining total battery cycle times; I is the current of the power battery, ΔC n is the percentage of capacity loss, is the pre-exponential factor positively correlated with the battery discharge rate, R is the ideal gas constant, Q t is the cumulative throughput of electricity, T b is the internal temperature of the battery, E b is the activation energy of the battery.

8. A method for energy management of an aviation hybrid power system according to claim 1, characterized in that, In step S3, the input state s and action a of the TD3 algorithm framework are respectively defined as: ； a = P eng ; Among them, SOC is the state of charge of the power battery, H and Ma are the altitude and Mach number of the aero-turbine engine respectively, and P req is the total required power of the hybrid system, and P eng is the shaft power of the engine.

9. The energy management method for an aviation hybrid power system according to claim 8, characterized in that, The TD3 algorithm framework includes a dual Critic network, adds noise to the target Actor network, and delays the update of the Actor network to achieve training and obtain the energy management strategy of the hybrid power system.

10. An energy management system for an aviation hybrid power system, characterized in that, The system includes a memory and at least one processor. The memory stores a computer program, and the processor is used to execute the computer program to implement the energy management method for the aero-hybrid power system according to any one of claims 1 to 9.

Images