Mountain environment-oriented power airship heavy-load logistics transportation system

Abstract

The invention discloses a power airship heavy-load logistics transportation system facing a mountain environment, and belongs to the technical field of aerospace engineering. The system comprises an active working lift system, an exoskeleton mechanical structure, a helium buoyancy lifting system and an intelligent control system. The active working lift system is used for providing a power supply for the airship and adjusting lift; the exoskeleton mechanical structure is used for supporting the periphery of the airship; the helium buoyancy lifting system is used for providing basic buoyancy lifting force and forms redundant supplement with the active working lift force system; and the intelligent control system is used for monitoring the airship operation state data and adjusting the airship operation state based on the monitored data. Working lift force is generated in the mode of the energy storage battery, the motor and the blades, the exoskeleton mechanical structure and the spiral blades are integrated, helium in the airship generates lift force to mainly balance the self weight, the spiral blades mainly play a role in balancing the lift force, and adaptability and economical efficiency of heavy-load transportation in the mountain environment are greatly improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of aerospace engineering, and particularly relates to a heavy-load logistics transportation system of a powered airship for mountainous environments. Background Art

[0002] With the continuous development of modern transportation and logistics technologies, the demand for heavy-load transportation in various fields is increasing day by day. Especially in mountainous and complex terrain areas, traditional transportation methods such as roads, railways, and helicopters are limited by geographical conditions, costs, and safety factors, and there is an urgent need to find alternative solutions. Among them, the material transportation solution using heavy drones has a certain feasibility, but it has deficiencies in load capacity and endurance. In recent years, airship transportation has received extensive attention due to its advantages such as vertical takeoff and landing, no runway requirement, and strong terrain adaptability. In the prior art, most airships mainly rely on conventional power systems and gas buoyancy to achieve basic load capacity. However, in mountainous environments with high altitudes and variable climates, airships often face the following technical problems: First, how to improve the working lift and ensure the safe loading and unloading of goods while ensuring their own stability; second, how to achieve the effective coordination of the power system and the gas buoyancy system so that the airship has higher dynamic regulation capabilities when carrying heavy loads; third, how to strengthen the wind and earthquake resistance performance of the overall airship structure without significantly increasing its own weight. The current technical solutions mostly focus on a single lift source, such as relying solely on filling helium or relying on an external fuel drive system, which cannot fully meet the actual needs of multi-working conditions and dynamic load changes in complex environments. Summary of the Invention

[0003] The present invention aims to solve the deficiencies of the prior art and proposes a heavy-load logistics transportation system of a powered airship for mountainous environments. By combining the respective advantages of drones and airships and introducing an exoskeleton mechanical support and a helical blade active adjustment structure, multi-level and composite lift distribution and regulation are achieved. The present invention can not only effectively balance the self-weight of the airship but also quickly match the buoyancy change when loading and unloading goods, thereby improving transportation efficiency and safety. At the same time, with the help of advanced materials and structural designs, the overall torsion resistance, vibration resistance, and fatigue resistance are enhanced, enabling the airship to still operate stably in extreme mountainous environments and meet the various strict requirements of modern heavy-load transportation.

[0004] To achieve the above object, the present invention provides the following solution: A heavy-load logistics transportation system of a powered airship for mountainous environments, comprising: an active working lift system, an exoskeleton mechanical structure, a helium buoyancy system, and an intelligent control system;

[0005] The active working lift system is used to provide power to the airship and adjust the lift;

[0006] The exoskeleton mechanical structure is used for the peripheral support of the airship;

[0007] The helium lifting system is used to provide basic lifting force, forming a redundant supplement with the active working lifting system;

[0008] The intelligent control system is used to monitor the operation state data of the airship and adjust the operation state of the airship based on the monitored data.

[0009] Further preferably, the active working lifting system includes: a lithium-ion energy storage battery, a battery management subsystem, and a motor-driven blade subsystem;

[0010] The lithium-ion energy storage battery is used to provide the main energy source for the airship;

[0011] The battery management subsystem is connected to the lithium-ion energy storage battery and is used to monitor the battery state, temperature, and current fluctuation of the lithium-ion energy storage battery;

[0012] The motor-driven blade subsystem is used to adjust the external lifting force of the airship.

[0013] Further preferably, the method for the battery management subsystem to obtain the battery state includes: using a dynamic multi-model fusion SOC algorithm to complete battery state monitoring;

[0014] The method for the dynamic multi-model fusion SOC algorithm to monitor the battery state includes:

[0015] Construct an electrochemical model based on the extended Kalman filter, initialize the state variables of the electrochemical model, and update the solid-phase diffusion equation based on the input current; calculate the residual between the observed voltage and the experimental voltage based on the solid-phase diffusion equation, adjust the Kalman gain based on the residual value, and output the corrected SOC EKF estimated value;

[0016] Construct a data-driven model, and the data-driven model learns the battery dynamic characteristics from historical data based on the long short-term memory network; the long short-term memory network includes: an input layer, a hidden layer, and an output layer; the input layer is used to receive historical current, voltage, and temperature sequences; the hidden layer includes: 2 layers of LSTM units, each layer has 64 neurons, and the activation function is tanh; the output layer is used to output SOC LSTM , and the loss function uses the mean square error;

[0017] Construct an equivalent circuit model, identify the parameters of the equivalent circuit model through the hybrid pulse power characteristics; update the OCV-SOC relationship table in real time; calculate SOC based on the ampere-hour integration method RC ;

[0018] By adjusting the weights, obtain the percentage of the remaining battery power SOC:

[0019] SOC = ω1·SOC EKF+ω2·SOC LSTM +ω3·SOC RC ,

[0020] In the formula, ω i represents the weight;

[0021] Among them,

[0022]

[0023] In the formula, λ represents the normalization factor for weight adjustment in the dynamic multi-model fusion algorithm; E i is the root mean square of the prediction errors of the dynamic multi-model in the most recent 10 times.

[0024] Further preferably, the method for the battery management subsystem to monitor the temperature includes:

[0025] A voltage dividing circuit is formed by an NTC thermistor and a fixed resistor, and the temperature of the battery is obtained based on the characteristic that the resistance value changes with temperature:

[0026]

[0027] In the formula, V T represents the output voltage of the voltage dividing circuit; V ref represents the reference voltage; R NTC represents the resistance value of the thermistor; R 固定 represents the resistance value of the fixed resistor;

[0028]

[0029] In the formula, T represents the current temperature; T ref represents the reference temperature; R ref represents the nominal resistance value at the reference temperature; B1 represents the material constant.

[0030] Further preferably, the motor-driven blade subsystem adjusts the rotational speed and output torque based on real-time load data and environmental disturbance data by adopting a closed-loop control algorithm to complete the lift adjustment.

[0031] Further preferably, the exoskeleton mechanical structure is optimized by the finite element analysis method, and the optimization method includes:

[0032] The structure is discretized into elements, the stiffness matrix [K] and the load vector {F} are established, and the displacement field {u} is solved:

[0033] [K]{u} = {F},

[0034] In the formula, {u} is the nodal displacement vector;

[0035] The material distribution is optimized by the variable density method:

[0036]

[0037] where ρ e ∈[0, 1] represents the unit density; p = 3; E0 represents the reference elastic modulus;

[0038] The trade - off design between lightweight and stiffness is achieved through the unit density ρ e

[0039] Further preferably, the intelligent control system adopts a distributed architecture, and the distributed architecture includes: an edge computing node layer, a central coordination layer, and a data management layer;

[0040] The edge computing node layer is used for flight control and power management;

[0041] The central coordination layer is used for global optimization;

[0042] The data management layer is used for environmental monitoring and fault diagnosis.

[0043] Further preferably, the method of flight control includes:

[0044] Control law design: Based on model predictive control, the state - space model is:

[0045]

[0046] y = C3x,

[0047] where x = [α β p q r] T , where α, β, p, q, and r respectively represent the angle of attack, sideslip angle, roll rate, pitch rate, and yaw rate; A represents the state matrix; B represents the input matrix; u represents the control input vector; C3 represents the output matrix; y represents the output vector;

[0048] Real - time guarantee: Through time - partitioned scheduling of the ARINC 653 standard, the control cycle is divided into: 50 μs: attitude loop; 200 μs: trajectory loop;

[0049] The method of power management includes:

[0050] The fuzzy PID algorithm is used to adjust the motor thrust, and the adjustment method is as follows:

[0051]

[0052] where ΔP represents the difference between the required power and the actual power; K p 、K i 、K d ​respectively represent the real-time response weight of the difference between the required power and the actual power, the compensation weight of the historical cumulative power deviation, and the prediction weight of the change trend of the power deviation. Compared with the prior art, the beneficial effects of the present invention are as follows:

[0053] The present invention generates working lift by adopting the method of "energy storage battery + motor + blade", combines with traditional helium floating, and realizes multi-energy complementarity; it can respond quickly when the load changes suddenly, breaking through the limitation of the single floating of traditional airships. The airship integrates an exoskeleton mechanical structure and a propeller blade. The helium gas in the airship generates lift mainly to balance its own weight (including the airship structure, battery motor, etc.), and the propeller blade mainly plays a role in lift balance, including the scheme of buoyancy matching for loading and unloading goods, improving the overall energy conversion efficiency, and ensuring continuous output under complex working conditions; strengthening the structural rigidity and anti-interference ability, and improving the stability of the airship under the condition of sudden change of wind speed; it is also convenient for later maintenance and module replacement; this design effectively resists the variable wind force and terrain interference in the mountain environment. Based on the adaptive control algorithm of real-time multi-sensor data, the coordinated operation of the motor, blade and helium system is realized; the system not only has high fault tolerance, but also can automatically switch the operation mode under extreme working conditions to ensure flight safety; the propeller blade adopts variable pitch technology, which can accurately control the local lift during the loading and unloading process, solve the safety hazards caused by buoyancy imbalance during the loading and unloading process of traditional airships, and greatly improve the transportation efficiency. Multiple redundant backups are fully considered in the design, such as spare solutions are provided in the battery, motor, helium and sensing control systems; even if a single module fails, other modules can quickly intervene to ensure the safe flight of the whole machine. Greatly improve the adaptability and economy of heavy-duty transportation in mountainous environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0055] Figure 1 It is a schematic structural diagram of the heavy-duty logistics transportation system of the power airship for mountain environment according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0056] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0057] To make the above objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] Embodiment 1:

[0059] Regarding the following defects of airships in the prior art: 1. Low energy utilization rate: The poor matching between the traditional energy storage and power system leads to low overall efficiency. 2. Insufficient structural stability: The flexible airbag structure is vulnerable to wind and temperature changes in complex mountain environments, resulting in poor stability. 3. Imprecise load regulation: A single lifting system is difficult to achieve rapid and precise lift matching during loading and unloading, posing safety hazards. 4. Poor environmental adaptability: Under conditions of high altitude, low temperature, and drastic wind speed changes, traditional airships cannot achieve effective lifting and lowering adjustments. The present embodiment of the invention provides a heavy-load logistics transportation system for a power airship facing mountain environments, as Figure 1 shown, including: an active working lift system, an exoskeleton mechanical structure, a helium lifting system, and an intelligent control system; the active working lift system is used to provide power to the airship and adjust the lift; among them, the active working lift system includes: a lithium-ion energy storage battery, a battery management subsystem, and a motor-driven blade subsystem; the lithium-ion energy storage battery is used to provide the main energy source for the airship; the battery management subsystem is connected to the lithium-ion energy storage battery and is used to monitor the battery state, temperature, and current fluctuations of the lithium-ion energy storage battery; the motor-driven blade subsystem is used to adjust the external lift of the airship. The exoskeleton mechanical structure is used for the peripheral support of the airship; the helium lifting system is used to provide basic lifting force and form a redundant supplement with the active working lift system; the intelligent control system is used to monitor the operation state data of the airship and adjust the operation state of the airship based on the monitored data.

[0060] Furthermore, in this embodiment, a high-energy-density lithium-ion energy storage battery is used as the core energy source, integrated with a battery management subsystem (BMS) to monitor battery status, temperature, and current fluctuations in real time, ensuring power supply stability and safety. The motor-driven blade subsystem utilizes a high-efficiency brushless DC motor, coupled with a digital servo drive. High-speed rotation generates localized airflow, enabling active regulation of external lift. Specifically, the motor-driven blade subsystem employs a closed-loop control algorithm to automatically adjust speed and output torque based on real-time load and environmental disturbance data, ensuring rapid and accurate lift response. Furthermore, the motor-driven blade subsystem complements traditional helium lift, enabling timely intervention and adjustment in the event of sudden load changes or wind speed fluctuations. The propeller blades utilize a variable-pitch design. Built-in high-precision sensors measure flight status and load conditions, automatically adjusting blade angle to achieve precise lift matching. During loading and unloading, precise control of propeller blade speed and pitch provides localized lift compensation, ensuring smooth cargo transfer, reducing vibration, and preventing safety hazards caused by sudden changes in the cargo hold.

[0061] Traditional SOC estimation relies on a single model (such as Kalman filtering or ampere-hour integration), and the error increases significantly (>5%) under complex operating conditions such as battery aging and low temperature. In this embodiment, the method for the battery management subsystem to obtain the battery status includes: using a dynamic multi-model fusion SOC algorithm to complete battery status monitoring; the method for monitoring the battery status using the dynamic multi-model fusion SOC algorithm includes:

[0062] An electrochemical model based on an extended Kalman filter (EKF) is constructed. The state variables of the electrochemical model are initialized with an initial SOC value of 50% and a uniform distribution of positive and negative electrode concentrations. The solid-phase diffusion equation is updated based on the input current. The residual between the observed voltage and the experimental voltage is calculated based on the solid-phase diffusion equation. The Kalman gain is adjusted based on the residual value and the corrected SOC is output. EKF Estimated value.

[0063] Construct a data-driven model, which learns battery dynamic characteristics from historical data based on a long short-term memory network without the need for a physical model; the long short-term memory network includes: an input layer, a hidden layer, and an output layer; the input layer is used to receive historical current, voltage, and temperature series (time window n = 10); the hidden layer includes: 2 layers of LSTM units, each layer has 64 neurons, and the activation function is tanh; the output layer: a fully connected layer maps to the SOC LSTM The loss function uses the mean square error (MSE). The model is trained with 10,000 battery test data sets covering -20°C to 60°C and 0.1C to 2C charge and discharge conditions.

[0064] Construct an equivalent circuit model RC network, use the resistor-capacitor (RC) circuit network to simulate the external characteristics of the battery, and balance the accuracy and computational complexity. The implementation steps include: identifying the parameters of the equivalent circuit model through the hybrid pulse power characteristics; updating the OCV-SOC relationship table in real time; calculating the SOC based on the ampere-hour integration method RC .

[0065] Adjust the weight in real time according to the working conditions to obtain the percentage of the remaining battery power SOC:

[0066] SOC = ω1·SOC EKF +ω2·SOC LSTM +ω3·SOC RC ,

[0067] In the formula, ω i represents the weight.

[0068] Among them,

[0069]

[0070] In the formula, λ represents the normalization factor for weight adjustment in the dynamic multi-model fusion algorithm, which needs to be calibrated through experiments according to different battery types to make the weight distribution more suitable for the actual working conditions, thereby improving the robustness and accuracy of battery SOC estimation in complex environments; E i is the root mean square of the last 10 prediction errors of the dynamic multi-model.

[0071] The method for the battery management subsystem to monitor the temperature includes:

[0072] Form a voltage dividing circuit through an NTC thermistor and a fixed resistor, and obtain the temperature of the battery based on the characteristic that the resistance value changes with temperature:

[0073]

[0074] In the formula, V T represents the output voltage of the voltage dividing circuit, which is used to reflect the battery temperature; V ref represents the reference voltage, which is the input voltage source of the voltage dividing circuit in this embodiment; R NTC represents the resistance value of the thermistor; R 固定 represents the resistance value of the fixed resistor.

[0075]

[0076] In the formula, T represents the current temperature, which is calculated through the resistance value R of the thermistor NTC ; T ref represents the reference temperature; R ref represents the nominal resistance value at the reference temperature (25 °C); B1 represents the material constant. Among them, R ref , Tref , B1 is usually provided by the manufacturer.

[0077] The method for the battery management subsystem to monitor current includes: connecting a low-resistance shunt resistor in series and measuring the voltage drop to calculate the current:

[0078]

[0079] In the formula, V shunt represents the voltage value of the shunt resistor; R shunt represents the resistance value of the shunt resistor.

[0080] Among them, R shunt needs temperature compensation and correction:

[0081] R shunt (T) = R0·(1 + mΔT + n(ΔT) 2 ),

[0082] In the formula, R0 represents the nominal resistance value of the shunt resistor at the reference temperature (usually 25°C or the standard laboratory temperature); ΔT represents the difference between the current ambient temperature and the reference temperature; m and n are the temperature coefficients of resistance.

[0083] In this embodiment, the blade pushes air through high-speed rotation and generates lift based on the momentum theorem and Bernoulli's principle; the formula of the momentum theorem is as follows:

[0084]

[0085] In the formula, represents the air mass flow rate; ρ air represents the air density, A1 represents the blade swept area, v avg represents the average flow velocity; v in and v out respectively represent the air inflow velocity in front of the blade and the air outflow velocity behind the blade.

[0086] The relationship between the lift coefficient and the angle of attack:

[0087] C L = 2πsinθ1 (approximate for small angles of attack),

[0088] In the formula, θ1 represents the angle of attack of the blade, and the lift coefficient C L increases linearly with the increase of the angle of attack.

[0089] The dynamic lift adjustment mechanism is as follows:

[0090] By adjusting the motor speed (ω) to change the blade tip speed (v tip = ωr, where r represents the blade rotation radius), and then controlling the lift:

[0091]

[0092] The lift is proportional to the square of the rotational speed. Rapid adjustment of ω can achieve active compensation of the lift.

[0093] The process of the closed-loop control algorithm includes:

[0094] 1. System modeling and problem description

[0095] For the motor dynamics equation, assuming the drive motor is a permanent magnet synchronous motor (PMSM), its torque output equation is:

[0096]

[0097] In the formula: T e represents the electromagnetic torque of the motor; p1 represents the number of pole pairs of the permanent magnet synchronous motor; ψ f represents the magnetic flux of the permanent magnet; L d and L q represent the direct-axis and quadrature-axis inductances respectively; i d and i q represent the direct-axis and quadrature-axis currents respectively.

[0098] (1) Load dynamics equation

[0099] Considering the external load and inertia, the system motion equation is as follows:

[0100]

[0101] In the formula: J represents the moment of inertia; ω represents the motor speed; T L represents the real-time load torque (which needs to be dynamically measured); B ω represents the viscous friction coefficient; T dist represents the environmental disturbance torque (such as wind resistance and mechanical vibration).

[0102] (2) Control objective

[0103] Design a closed-loop controller to make the speed ω track the target value ω ref , while suppressing the influence of the load T L and the disturbance T dist :

[0104]

[0105] 2. Closed-loop control algorithm

[0106] Design an extended state observer (ESO): Regarding the load and disturbance as a total disturbance f(t) unifiedly, the system motion equation is transformed into:

[0107]

[0108] In the formula: u = T e , B3 represents the damping coefficient of the system, reflecting the rotational resistance characteristics.

[0109] ESO structure: Expand the total disturbance f(t) into the system state and establish a third-order observer:

[0110]

[0111] In the formula, z1 represents the estimated rotational speed; z2 represents the estimated acceleration; z3 represents the estimated total disturbance f(t); β1, β2, and β3 are all observer gains, determined by pole placement.

[0112] Feedforward-feedback composite controller design:

[0113] Control law:

[0114] Feedforward term To compensate for the total disturbance, b represents the system control gain (input gain), and represents the amplification factor of the control input (such as the motor torque T 电机 ) on the system dynamics.

[0115] Feedback term: PID regulator to ensure tracking performance.

[0116] Frequency-domain analysis (stability verification):

[0117] The transfer function of the closed-loop system is:

[0118] In the formula, s represents the complex frequency variable; k p , k i , k d respectively represent the proportional, integral, and derivative gains of the controller, used to adjust the system tracking performance.

[0119] Select k p , k i , k d to make the system stable.

[0120] 3. Disturbance rejection and parameter adaptation mechanism

[0121] Real-time measurement of load torque:

[0122] Invert the load torque through the motor current and rotational speed:

[0123]

[0124] Among them, it is necessary to perform low-pass filtering on the rotational speed differential signal, and the cut-off frequency f c = 100 Hz.

[0125] Environmental disturbance suppression:

[0126] Frequency domain separation:

[0127] The disturbance spectrum analysis shows that T dist is concentrated in the high frequency band (>50 Hz), and a notch filter is adopted:

[0128]

[0129] where ω n represents the center frequency of the notch filter, corresponding to the main frequency of the environmental disturbance to be suppressed, taking 50 Hz; ω d represents the designed target frequency of the notch filter, used to suppress the extended frequency band of high frequency disturbances, taking 500 Hz; ξ represents the damping characteristic parameter of the notch filter, used to control the attenuation depth and bandwidth, taking 0.7.

[0130] Fuzzy adaptive PID:

[0131] According to the error e = ω ref -ω and its rate of change de / dt, the PID parameters are dynamically adjusted:

[0132] k p = k p0 +α|e|,

[0133] k i = k io e -β|e| ,

[0134] k d = k d0 (1 + γ|de / dt|),

[0135] where α, β, γ represent the proportional dynamic gain, integral term exponential gain, and derivative term sensitivity gain, which are calibrated through experiments; k p0 , k i0 , k d0 represent the proportional term reference gain (determining the initial response speed), integral term reference gain (determining the initial steady-state error elimination ability), and derivative term reference gain (determining the initial anti-disturbance ability) respectively.

[0136] In this embodiment, the method for the motor-driven vane subsystem to work in cooperation with the traditional helium levitation includes:

[0137] 1. Cooperative control architecture and mathematical model

[0138] 1) The system dynamics equation is as follows:

[0139] Total lift balance equation: L total = FHe +F motor -W total ,

[0140] In the formula: F He = ρ air V He g - ρ He V He g, representing the static buoyancy of helium; representing the dynamic lift of the motor blade; W total = m payload g + m structure g, representing the total weight. Among them, ρ air represents the air density; V He represents the volume of helium; g represents the acceleration due to gravity; ρ He represents the density of helium; m payload represents the payload mass; m structure represents the structural mass.

[0141] 2) Considering the kinematic and dynamic coupling of the airship, the dynamic coupling equation is as follows:

[0142]

[0143]

[0144]

[0145] In the formula, h represents the flight altitude; v z represents the vertical velocity; represents the aerodynamic drag; C D represents the aerodynamic drag coefficient, which determines the energy loss; A cross represents the cross-sectional area, which affects the drag calculation; m total represents the total mass, which needs to be balanced with the buoyancy; T motor represents the active thrust, which is used for altitude adjustment; T load represents the external disturbance, which needs to be dynamically suppressed.

[0146] 2. Lift Complementary Mechanism

[0147] 1) Helium Buoyancy Response Characteristics

[0148] Buoyancy adjustment method: By inflating and deflating the auxiliary airbag to change the volume of helium V He .

[0149] Response time model: ΔF He (t) = ΔV He ·(ρ air - ρ He )g·(1 - e -t / τ ), where ΔF He represents the change in helium buoyancy; ΔV He represents the change in helium volume; the typical value τ≈5 - 10s.

[0150] 2) Dynamic compensation of motor blades

[0151] Lift transient response:

[0152]

[0153] where τ motor = J / (B + K t K e ), represents the motor electromechanical time constant, the typical value τ motor ≈0.1 - 0.5s; K motor represents the comprehensive gain coefficient; K t represents the motor torque constant; K e represents the back - electromotive force constant.

[0154] 3) Cooperative control strategy

[0155] Lift distribution optimization problem:

[0156]

[0157] s.t. F He + F motor = W total + ΔW,

[0158] F He,min ≤F He ≤F He,max ,

[0159] 0 ≤ F motor ≤F motor,max ,

[0160] In the formula, F He represents the helium buoyancy; F He,ref represents the reference value of helium buoyancy; F He,min represents the minimum value of the physical limit of helium buoyancy; F He,max represents the maximum value of the physical limit of helium buoyancy; ΔW represents the sudden change in weight caused by external disturbances; α1 and β1 are both weight systems, α1 is much smaller than β1, and helium buoyancy is preferentially used for energy conservation.

[0161] The analytical solution is: In the formula, the superscript * represents the optimal solution.

[0162] 3. Cooperative response under sudden load conditions

[0163] 1) Dynamics of sudden load:

[0164] Assume that a weight mutation ΔW occurs at t = 0, and the helium buoyancy response: adjusts slowly according to an exponential curve; the motor lift response: completes compensation within t d <τ motor time.

[0165] Rewrite the motion equation as:

[0166]

[0167] where F He (t) is the slow response, F motor (t) is the fast response, and u(t) is the step function.

[0168] 2) Composite control law design:

[0169] Feedforward compensation: Anticipate the demand according to ΔW:

[0170]

[0171] where t m =τ motor / 2..

[0172] Feedback correction: Eliminate the residual error through PID:

[0173] F motor,fb =K p (h ref -h)+K i ∫(h ref -h)dt,

[0174] where h ref 、K p 、K i are the core parameters of the composite control law, which set the target, adjust the dynamic response, and eliminate the steady-state error respectively.

[0175] Total motor lift command:

[0176] F motor =sat(F motor,ff +F motor,fb ), where sat() is the actuator output limiter function.

[0177] In this embodiment, the exoskeleton mechanical structure adopts advanced composite materials and high-strength aluminum alloy. The structure is optimized through finite element analysis to ensure that the external forces can be effectively dispersed under the conditions of large loads and high wind speeds, thus enhancing the overall rigidity. The structural design takes into account torsional resistance, seismic resistance and fatigue performance, and introduces the modular design concept: each component can achieve rapid configuration adjustment and maintenance replacement in different transportation tasks through standardized interfaces and quick disassembly and assembly technologies. In addition, high-performance shock-absorbing materials and damping devices are used to effectively reduce the impact of vibration and shock on the core structure, ensuring long-term stable operation. Sensors are installed at each node to monitor the stress state in real time, providing data support for the intelligent control system to achieve dynamic compensation and safety warning.

[0178] 1. The methods for optimizing the structure through finite element analysis include:

[0179] Discretization equation: Discretize the structure into elements, establish the stiffness matrix [K] and the load vector {F}, and solve the displacement field {u}:

[0180] [K]{u} = {F},

[0181] where {u} is the nodal displacement vector.

[0182] Variable density method for optimizing material distribution:

[0183]

[0184] where ρ e ∈[0,1], representing the unit density; p is the penalty factor, and in this embodiment, p = 3; E0 represents the reference elastic modulus, which is the original elastic modulus of the material before density optimization, that is, the stiffness value of the material in the fully dense (density ρ = 1) state. The trade-off between lightweight and stiffness is achieved through the unit density ρ e

[0185] 2. Sensor network architecture design

[0186] 1) The deployment of multi-modal sensors is shown in Table 1:

[0187] Table 1

[0188]

[0189] 2) Data fusion architecture

[0190] Adopt the federated Kalman filter architecture:

[0191] Local filtering fusion:

[0192] Global fusion:

[0193] where P​i is the local estimation error covariance matrix, and H i is the observation matrix; z i represents the real-time observation data of the i-th sensor in the federated Kalman filter.

[0194] 3. Real-time calculation of the stress state

[0195] 1) Strain-stress conversion

[0196] Based on Hooke's law in general form, σ = C1·ε, where C1 represents the stiffness matrix; ε represents the linear strain.

[0197] For isotropic materials, it is simplified to:

[0198]

[0199] In the formula, E represents the elastic modulus, which is the stiffness of the material in the elastic deformation stage; v represents the Poisson's ratio, which is the ratio of the lateral contraction to the longitudinal elongation of the material; σ xx , σ yy , σ zz , τ xy , τ yz , τ zx respectively represent the normal stresses in the x, y, and z directions and the shear stresses in the xy, yz, and zx planes; ε xx , ε yy , ε zz , γ xy , γ yz , γ zx respectively represent the relative elongation or compression rates in the x, y, and z directions and the shear deformation amounts of the angles in the corresponding planes.

[0200] 2) Dynamic load reconstruction

[0201] Invert the inertial force through the acceleration data:

[0202]

[0203] In the formula, I is the inertia tensor, which is updated in real time through the finite element model; m represents the total mass of the system; a1 represents the linear acceleration.

[0204] 3. Dynamic compensation method

[0205] 1) Active damping control

[0206] The actuator output based on the LQR algorithm: u = -K·x1, K = R -1 B T P,

[0207] In the formula, P is the solution of the Riccati equation: A T P + PA - PBR-1 B T P + Q = 0. K represents the LQR feedback gain matrix, which is calculated by the optimal control algorithm; x1 represents the system state vector, including dynamic variables such as displacement and velocity; R represents the control input weight matrix, which adjusts the cost of control energy in the LQR algorithm.

[0208] 2) Strain compensation mechanism

[0209] Establish a strain prediction model:

[0210] Adjust the actuator displacement through feedforward compensation: In the formula, L is the length characteristic, and K act is the actuator stiffness.

[0211] 4. Safety warning model

[0212] 1) The multi-level warning thresholds are shown in Table 2:

[0213] Table 2

[0214]

[0215] 2) Fatigue damage accumulation

[0216] Improved Miner linear cumulative model:

[0217] Among them, α2 = 0.3 is the creep weight damage factor, n represents the actual number of cycles at the stress level; n i represents the actual number of cycles at the i-th stress level; S i represents the i-th stress level; N i represents the fatigue life at the stress S i ; t j represents the creep loading time at the j-th stress; σ j represents the j-th creep stress; T creep represents the creep failure time at the stress σ j .

[0218] 3) Crack growth prediction

[0219] Real-time integration based on Paris' law:

[0220]

[0221] In the formula, Y(a) represents the geometric correction factor, reflecting the influence of crack shape, size and boundary conditions; ΔK represents the stress intensity factor range; a2 represents the current crack length.

[0222] The Runge-Kutta method is used to iteratively solve the crack length a(t).

[0223] The helium floating system is filled with high-purity helium to provide the basic lifting force to balance the overall self-weight. The helium chamber is composed of multi-layer heat insulation, pressure resistance and anti-seepage materials to ensure the sealing and stable shape under extreme environments such as low temperature, strong wind at high altitude. Pressure, temperature sensors and automatic regulating valves are integrated in the system to monitor the helium state in real time; in case of extreme weather or sudden load changes, helium works together with the active working lifting system to quickly adjust the lifting force to achieve redundant supplement and safety guarantee. In this embodiment, an intelligent heat preservation system is also introduced, and the built-in electric heating device is used to prevent the performance decline of helium caused by low temperature to ensure long-term stable floating effect.

[0224] In this embodiment, the method for coordinated control of the helium floating system and the active working lifting system includes:[[]]

[0225] In extreme weather, the helium buoyancy works together with the active lifting system (such as a vector thruster):

[0226] L 总 = F 氦 + F 推进 ·sinθ1,

[0227] where θ1 is the thruster elevation angle, and the error e(t) = L ref - L actual is adjusted through a PID controller, where L ref represents the target lift set value, the benchmark for the control system to track; L actual represents the actual lift feedback value for closed-loop error correction.

[0228] Adjust the center of gravity position by changing the air volume of the auxiliary airbag:

[0229] m air = ρ air ·V 副气囊 .

[0230] Combine the Kalman filter to optimize the control command to achieve a center of gravity adjustment accuracy of ±0.5m.

[0231] 1. Helium low-temperature performance decline

[0232] 1) Temperature sensitivity of helium buoyancy

[0233] The helium buoyancy follows Archimedes' law, and its buoyancy formula is:

[0234] F 浮力 = (ρ air - ρ He )·V·g,

[0235] Among them, V represents the volume of helium displacing air; ρ He Is significantly affected by temperature. According to the international standard atmosphere model, for every 1°C decrease in temperature, the volume of helium shrinks by approximately 0.366% (thermal expansion coefficient β = 1 / 273), resulting in a buoyancy loss of approximately 3.66%. At extremely low temperatures (such as -50°C), the increase in helium density may reduce the buoyancy by more than 30%, directly affecting the airship's lifting ability.

[0236] 2) Influence of low temperature on materials and structures

[0237] In a low-temperature environment, airship skin materials (such as polyester film and carbon fiber composite materials) are prone to embrittlement, and at the same time, the helium molecule permeability increases, exacerbating the leakage risk.

[0238] 2. Thermodynamic model of the electric heating device

[0239] 1) Heat balance equation

[0240] The system needs to satisfy steady-state heat balance:

[0241] P 加热 = P 散热 + P 氦气温升 ,

[0242] Among them: P 加热 represents the electric heating power; P 散热 = k·A3·ΔT, where k is the skin thermal conductivity, A3 represents the effective area of heat conduction between the airship skin and the external environment, and ΔT is the temperature difference.

[0243] In the formula, c p is the specific heat capacity at constant pressure of helium, is the mass flow rate.

[0244] 3. Design description of the intelligent thermal insulation system

[0245] 1) System architecture

[0246] Perception layer: Distributed temperature sensors (such as fiber Bragg grating sensors) monitor the temperature gradient inside the helium gas bag with an accuracy of ±0.1°C.

[0247] Control layer: The PID controller dynamically adjusts the heating power with a response time <1 second and an overshoot <2%.

[0248] Execution layer: Conductive fabric heating film: Square resistance 2400 - 3600 Ω / sq, embedded in the composite skin, temperature resistant -60 to 150°C, power density 50 - 100 W / m 2 .

[0249] Zone heating module: Divide the helium gas bag into multiple independent temperature zones and preferentially heat areas with high leakage risk (such as joints).

[0250] 2) Energy efficiency optimization strategies

[0251] Adaptive power adjustment: Dynamically adjust the power according to the ambient temperature and flight altitude.

[0252] Phase change energy storage (PCM): Integrate paraffin-based phase change materials (latent heat ≥ 200 J / g), release the stored heat during low-temperature periods at night, and reduce the power consumption by 30%.

[0253] Waste heat recovery: Utilize the waste heat of the airship motor (efficiency ≥ 85%) for auxiliary heating to reduce the demand for independent energy supply.

[0254] 3) Safety and reliability design

[0255] Multi-level redundancy: When the main heating film fails, the standby carbon nanotube heating wire is automatically activated, and the switching time < 0.5 seconds.

[0256] Insulation protection: Adopt cast mica ceramic insulating tubes with a withstand voltage of 6.0 MPa and a breakdown voltage ≥ 10 kV / mm to prevent arc discharge.

[0257] Fault diagnosis: Predict the life of heating elements based on big data analysis (e.g., a resistance change rate > 5% triggers an early warning), and the MTBF (Mean Time Between Failures) ≥ 50,000 hours.

[0258] The intelligent thermal insulation system actively regulates the helium temperature through an electric heating device, solving the problems of buoyancy attenuation and material failure caused by low temperature. Its design integrates thermodynamic modeling, materials science, and intelligent control technology, achieving optimal energy efficiency and long-life operation while ensuring the floating performance.

[0259] The intelligent control system adopts a distributed architecture, consisting of multi-core embedded real-time processors, integrating functions such as flight control, power management, environmental monitoring, and fault diagnosis. A multi-parameter sensor array is deployed internally, including an anemometer, temperature and humidity sensors, pressure sensors, accelerometers, and gyroscopes, etc., to collect real-time data on the external environment and the internal state of the airship. The data is transmitted to the central processing unit through a high-speed data bus or wireless network. The central processing unit uses advanced PID and adaptive control algorithms to dynamically adjust the motor drive, blade pitch, and helium replenishment strategy to achieve optimal lift distribution. The system design has multiple fault tolerance mechanisms and automatic switching functions. When any subsystem fails, other modules can quickly take over some functions to ensure the safe flight of the whole machine; at the same time, it maintains data interaction with the ground command center through a wireless communication module to achieve full-process status monitoring and early warning.

[0260] In this embodiment, the distributed control architecture is based on the multi-agent cooperative control theory. The system adopts a hierarchical distributed architecture, including an edge computing node layer (flight control, power management), a central coordination layer (global optimization), and a data management layer (environmental monitoring and fault diagnosis). Each node realizes deterministic communication through Time-Triggered Ethernet (TTEthernet), and the communication delay is controlled at the microsecond level. The core equation is:

[0261] In the formula, x i represents the real-time state of the edge node, driving local control decisions; e i represents the control error, and feedback regulation is used to achieve tracking or cooperation.

[0262] The function modules are implemented as follows:

[0263] 1) Flight control subsystem

[0264] Control law design: Based on Model Predictive Control (MPC), the state space model is:

[0265]

[0266] y = C3x,

[0267] where x = [α β p q r] T , where α, β, p, q, and r represent the angle of attack, sideslip angle, roll rate, pitch rate, and yaw rate, respectively; A represents the state matrix; B represents the input matrix; u represents the control input vector; C3 represents the output matrix; y represents the output vector.

[0268] Real-time guarantee: Through the time-partitioned scheduling ARINC 653 standard, the control cycle is divided into: 50 μs: attitude loop (gyroscope data update) 200 μs: trajectory loop (GPS / INS fusion).

[0269] 2) Power management module

[0270] Dynamic power distribution: The fuzzy PID algorithm is used to adjust the motor thrust, and the mathematical model is:

[0271]

[0272] where ΔP is the difference between the required power and the actual power. K p , K i , K d represent the real-time response weight of the difference between the required power and the actual power, the compensation weight of the historical cumulative power deviation, and the prediction weight of the power deviation change trend, respectively.

[0273] Multi - mode switching: Preset cruise / climb / emergency power mode, and realize millisecond - level mode switching through a finite - state machine.

[0274] 3) Environmental monitoring network

[0275] Multi - sensor fusion: Integrate the air data system (static pressure, dynamic pressure), temperature sensors (PT1000), and vibration sensors (MEMS accelerometers), and use the federated Kalman filter to achieve:

[0276]

[0277] Achieved accuracy: Temperature ±0.5°C, pressure ±10 Pa.

[0278] Anomaly detection: Predict sensor data based on the long short - term memory network (LSTM), and trigger an alarm when the residual exceeds 3σ.

[0279] 4) Fault diagnosis and fault tolerance

[0280] The hierarchical diagnosis architecture is shown in Table 3:

[0281] Table 3

[0282]

[0283] Dynamic redundancy strategy:

[0284] Hot backup: Key channels (such as flight control computers) adopt dual redundancy, and the switching time <10 μs.

[0285] Cold backup: Non - critical modules (cabin pressure control) adopt a restart - after - failure mechanism.

[0286] In this embodiment, the working modes of the airship during flight and loading / unloading are as follows:

[0287] Take - off stage: Utilize the basic buoyancy provided by helium, first start the motor - driven blade subsystem, and gradually increase the active lift; the exoskeleton mechanical structure synchronously releases the pre - tension force, and through precise adjustment, ensure overall stability and gradually achieve low - speed and smooth take - off.

[0288] Cruise stage: In the stable flight state, the active lift and helium buoyancy act together, and are dynamically adjusted according to the real - time load and climate change; the intelligent control system continuously monitors the flight parameters and automatically adjusts to maintain the established altitude and speed, ensuring long - term stable cruise.

[0289] Loading / unloading process: In response to the lift fluctuation caused by the sudden change of goods in the loading / unloading cabin, the propeller blades immediately respond, achieve lift compensation by adjusting the pitch and rotational speed, and at the same time start the auxiliary support system to reduce vibration and ensure smooth loading / unloading.

[0290] Landing stage: In the predetermined landing area, the system gradually reduces the rotational speed of the motor-driven blades, and at the same time, intelligently adjusts the helium lifting amount to make the airship descend slowly; the exoskeleton structure and the intelligent control system work together to ensure precise control and safe landing in the final stage.

[0291] The embodiments described above are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A heavy-load logistics transportation system for a power airship facing mountainous environments, characterized in that, Including: An active working lift system, an exoskeleton mechanical structure, a helium floating lift system, and an intelligent control system; The active working lift system is used to supply power to the airship and adjust the lift; The exoskeleton mechanical structure is used for the peripheral support of the airship; The helium floating lift system is used to provide the basic floating lift, forming a redundant supplement with the active working lift system; The intelligent control system is used to monitor the operation state data of the airship and adjust the operation state of the airship based on the monitored data.

2. The heavy-load logistics transportation system of a powered airship for mountainous environments according to claim 1, characterized in that, The active working lift system includes: a lithium-ion energy storage battery, a battery management subsystem, and a motor-driven blade subsystem; The lithium-ion energy storage battery is used to supply the main energy source to the airship; The battery management subsystem is connected to the lithium-ion energy storage battery and is used to monitor the battery state, temperature, and current fluctuation of the lithium-ion energy storage battery; The motor-driven blade subsystem is used to adjust the external lift of the airship.

3. The heavy-load logistics transportation system of a powered airship for mountainous environments according to claim 2, characterized in that The method for the battery management subsystem to obtain the battery state includes: using a dynamic multi-model fusion SOC algorithm to complete the battery state monitoring; The method for the dynamic multi-model fusion SOC algorithm to monitor the battery state includes: Construct an electrochemical model based on the extended Kalman filter, initialize the state variables of the electrochemical model, and update the solid-phase diffusion equation based on the input current; calculate the residual between the observed voltage and the experimental voltage based on the solid-phase diffusion equation, adjust the Kalman gain based on the residual value, and output the corrected SOC EKF Estimated value; Build a data-driven model, which learns the battery dynamic characteristics from historical data based on a long short-term memory network; the long short-term memory network includes: an input layer, a hidden layer, and an output layer; the input layer is used to receive historical current, voltage, and temperature sequences; the hidden layer includes: 2 layers of LSTM units, with 64 neurons in each layer, and the activation function is tanh; the output layer is used to output the SOC LSTM , and the mean square error is used as the loss function; Construct an equivalent circuit model, and identify the parameters of the equivalent circuit model through the hybrid pulse power characteristics; update the OCV-SOC relationship table in real time; calculate the SOC based on the ampere-hour integration method RC ; By adjusting the weight, obtaining the percentage of the remaining battery power SOC: SOC = ω1·SOC EKF + ω2·SOC LSTM + ω3·SOC RC , where ω i represents the weight; Wherein, where λ represents the normalization factor for weight adjustment in the dynamic multi-model fusion algorithm; E i is the root mean square of the prediction errors of the dynamic multi-model in the last 10 times.

4. The dynamic airship heavy-load logistics transportation system for mountain environment according to claim 2, wherein The method for the battery management subsystem to monitor the temperature includes: By forming a voltage dividing circuit with an NTC thermistor and a fixed resistor, obtaining the temperature of the battery based on the characteristic that the resistance value changes with temperature; Where, V T represents the output voltage of the voltage dividing circuit; V ref represents the reference voltage; R NTC represents the resistance value of the thermistor; R 固定 represents the resistance value of the fixed resistor; Where T represents the current temperature; T ref represents the reference temperature; R ref represents the nominal resistance value at the reference temperature; B1 represents the material constant.

5. The dynamic airship heavy-load logistics transportation system for mountain environments according to claim 2, characterized in that The motor-driven blade subsystem adjusts the rotation speed and output torque based on the real-time load data and environmental disturbance data by adopting a closed-loop control algorithm to complete the lift adjustment.

6. The heavy-load logistics transportation system of a powered airship for mountainous environments according to claim 1, characterized in that The exoskeleton mechanical structure optimizes the structure through the finite element analysis method, and the optimization method includes: Discretizing the structure into elements, establishing the stiffness matrix [K] and the load vector {F}, and solving the displacement field {u}: [K]{u} = {F}, In the formula, {u} is the node displacement vector; Optimizing the material distribution by the variable density method: where ρ e ∈[0, 1] represents the unit density; p = 3; E0 represents the reference elastic modulus; Through the unit density ρ e Complete the trade-off design between lightweight and stiffness.

7. The heavy-load logistics transportation system of a powered airship for mountainous environments according to claim 1, wherein The intelligent control system adopts a distributed architecture, and the distributed architecture includes: an edge computing node layer, a central coordination layer, and a data management layer; The edge computing node layer is used for flight control and power management; The central coordination layer is used for global optimization; The data management layer is used for environmental monitoring and fault diagnosis.

8. The dynamic airship heavy-load logistics transportation system for mountain environment according to claim 7, wherein The method for the flight control includes: Control law design: Based on model predictive control, the state space model is: y = C3x, where x = [α β p q r] T , where α, β, p, q, and r respectively represent the angle of attack, sideslip angle, roll rate, pitch rate, and yaw rate; A represents the state matrix; B represents the input matrix; u represents the control input vector; C3 represents the output matrix; and y represents the output vector. Real-time guarantee: By the time partition scheduling ARINC 653 standard, the control period is divided into: 50μs: attitude loop; 200μs: trajectory loop; The method for the power management includes: Adopting a fuzzy PID algorithm to adjust the motor thrust, and the adjustment method is as follows: Among them, ΔP represents the difference between the required power and the actual power; K p , K i , K d respectively represent the real-time response weight of the difference between the required power and the actual power, the compensation weight of the historical cumulative power deviation, and the prediction weight of the change trend of the power deviation.

Images