Abstract:To improve the efficiency of the traditional genetic algorithm when the agile satellite observes large-scale ground target points and increase the solution efficiency of intelligent optimization algorithms, the traditional genetic algorithm was improved, and a tabu search-simulated annealing genetic hybrid algorithm was proposed. First, considering the time constraints and attitude orbit dynamics constraints of spacecraft in observing ground target points, the corresponding fitness function was established. The proposed fitness function could guarantee high observation gains and low observation energy consumption, and reflect the observation requirements of practical engineering problems. Subsequently, to improve the mutation process of the traditional genetic algorithm, a tabu search-simulated annealing mutation method was proposed. This mutation method introduced the tabu search method and Metropolis rule in the process of individual mutation optimization. As a result, the tabu search-simulated annealing mutation method could improve the probability of obtaining the optimal global solution, and accelerate the convergence speed of the algorithm. Compared with the traditional genetic algorithm, simulation results showed that the tabu search-simulated annealing genetic hybrid algorithm saved about 40% of the running time. The operating efficiency of the algorithm was also higher than that of other improved genetic algorithms such as simulated annealing genetic algorithm and tabu search genetic algorithm. The results verified the efficiency of the tabu search-simulated annealing genetic hybrid algorithm in solving the mission planning problem of agile observation satellite.

Abstract:To meet the large demand of hyperspectral images for hyperspectral anomaly detection, a hyperspectral image generation method was proposed based on the hyperspectral characteristic simulation data and background characteristic data of the target. The flow and heat transfer model, infrared radiation characteristic model, and hyperspectral image simulation model were studied. The aircraft reflectivity measured by experiment was taken as the input for the calculation of target characteristics. Combined with the actual observed background undulating image, hyperspectral images with different pixel abundances and signal-to-noise ratios were generated under the conditions such as specific spectral response characteristics of remote sensor, relative calibration error of remote sensor, and random additive noise. The abnormal pixels of the simulation image were detected by RX algorithm and CEM algorithm. Results show that the model could generate hyperspectral subpixel simulation images based on the performance parameters and target abundance requirements of the remote sensor. The image could reflect the impact of target aircraft pixel abundance and signal-to-noise ratio on the detection results. By adjusting the input parameters, the hyperspectral simulation image for subpixel anomaly detection could be efficiently constructed. When the RX algorithm was used to detect hyperspectral simulation image, the detection results were largely affected by the noise. When the signal-to-noise ratio was as low as 10 dB, it was difficult to detect the abnormal pixels by RX algorithm with the abundance less than 0.4. While the CEM algorithm based on spectral matching detection could detect anomalies and improve the detection probability under low pixel abundances and signal-to-noise ratios.

Abstract:To reduce the bow wave disturbance on hose-drogue during the docking operation of aerial refueling and improve the docking success rate of aerial refueling, the motion characteristics of hose-drogue under the influence of bow wave were simulated and analyzed. First, based on the lumped parameter method, a dynamic model of hose-drogue assembly with hose elastic, bending restoring moment, and variable-length was built. Then, a theoretical model for the bow wave of Rankine half-body was introduced to simulate the bow wave field conforming to the nose shape of the receiver aircraft. Finally, through the numerical simulation of turbulence, gravity, and tanker wake, the motion characteristics of hose-drogue under bow wave were obtained. The effects of docking height, flight speed, hose material, docking speed, docking method, and nose shape of the receiver on the motion characteristics of hose-drogue under bow wave were analyzed. Results show that when the receiver aircraft was close to the drogue, the drogue would deviate away from the receiver under the bow wave effect and then swing back. High docking altitude and low flight speed reduced bow wave disturbance. Reasonable setting of docking speed, docking method, hose material, section type, and nose shape of aircraft could also reduce the disturbance of bow wave to drogue. Simulation results show that the bow wave disturbance could be reduced by optimizing the refueling equipment and docking process.

Abstract:To study the influence of initial parameters on the cavitation and motion characteristics of revolution bodies entering water in parallel, the realizable k-ε turbulence model, the volume of fluid (VOF) multiphase flow model, and the Schnerr and Sauer cavitation model were employed based on the finite volume method. The overlapping grid technique was used to simulate the process of water entry in parallel with different initial speeds, initial clearance distances, and cross-flow speeds. First, a numerical calculation model of high-speed water entry in parallel of revolution bodies was established, and the validity of the calculation method was verified. Then, based on this model, numerical calculations of parallel water entry with different initial parameters were performed to obtain the flow field and motion characteristics under different parameters. Finally, combined with the calculation results, the variations of the cavitation morphology and characteristic size, as well as the lateral and yaw movements of the moving bodies under different parameters were analyzed. Results show that with the increase of the initial velocity of water entry, the inner side cavitation phenomenon was more severe. With the outer polar radius and the limit length of the cavitation at the same dimensionless moment increased, the dimensionless lateral displacement and yaw angle of the revolution body were larger. As the initial clearance distance decreased, the limit length of the cavitation at the same dimensionless moment was smaller, and the dimensionless lateral displacement and yaw angle of the revolution body first increased and then decreased. Under the action of small cross-flow, the radial dimensions of the outer cavitation on the outer side of the revolution body in forward and leeward directions were basically the same with that of the single revolution body, while the length was slightly larger. With the increase of the cross-flow speed, the cavity size differences of parallel moving bodies were greater than that of single body. The radial size differences of the inner cavitation were large, where the polar radius of the inner cavitation of face-flow revolution body was greater, and with the increase of the cross-flow speed, the maximum value of the polar radius maintained around d+D/2. When the cross-flow speed was small, the two revolution bodies tended to be close to the head and far away from the tail; when the cross-flow speed was large, the two revolution bodies tended to be close to the tail and far away from the head.

Abstract:When an unmanned combat aerial vehicle (UCAV) is making the decision of autonomous maneuver in air combat, it faces large-scale calculation and is susceptible to the uncertain manipulation of the enemy. To tackle such problems, a decision-making model for autonomous maneuver of UCAV in air combat was proposed based on deep reinforcement learning algorithm in this study. With this algorithm, the UCAV can autonomously make maneuver decisions during air combat to achieve dominant position. First, based on the aircraft control system, a six-degree-of-freedom UCAV model was built using MATLAB/Simulink simulation platform, and the appropriate air combat action was selected as the maneuver output. On this basis, the decision-making model for the autonomous maneuver of UCAV in air combat was designed. Through the relative movement of both sides, the operational evaluation model was constructed. The range of the missile attack area was analyzed, and the corresponding advantage function was taken as the evaluation basis of the deep reinforcement learning. Then, the UCAV was trained by stages from the easy to the difficult, and the optimal maneuver control command was analyzed by investigating the deep Q network. Thereby, the UCAV could select corresponding maneuver actions in different situations and evaluate the battlefield situation independently, making tactical decisions and achieving the purpose of improving combat effectiveness. Simulation results suggest that the proposed method can make UCAV choose the tactical action independently in air combat and reach the dominant position quickly, which greatly improves the combat efficiency of the UCAV.

Abstract:For improving the accuracy and robustness of the impact-angle-control guidance of air-to-surface missiles to enlarge the penetration ability against the air-defense system and the impact effect, based on the analytical solution of pure proportional navigation (PPN) guidance law against stationary targets, a three-dimensional PPN-based impact-angle-control guidance law (3D-PPNIACG) was proposed. First, based on the analytical solution of PPN against stationary targets, the guidance performance of two-dimensional PPN-based impact angle constraint guidance law (2D-PPNIACG) was analyzed, including maximum commanded acceleration, velocity increment, and capture region. Then, based on the 2D-PPNIACG and the orthogonal decomposition method of three-dimensional guidance, and with the help of the spatial geometric relationship between missile and target, 3D-PPNIACG was put forward, which can guarantee the impact-angle-control in vertical plane as well as horizontal plane. Finally, the guidance performance of 3D-PPNIACG was analyzed through numerical simulation cases, and the effectiveness and robustness of this guidance law were verified. The study and simulation results show that the proposed 3D-PPNIACG has the advantages of simple structure, ease of implementation, and good robustness, which can realize the impact-angle-control in both vertical and horizontal planes and hence has good application prospect.

Abstract:Integrating the advantages of infrared and visible images by image fusion is an effective means to enhance the applicability of optical images in low illumination conditions. Despite the wide application of sparse representation (SR) theory in the field of infrared and visible image fusion, the drawbacks including detail loss and low toleration with mis-registration caused by the local patch representation nature of SR have never been effectively solved. Different from SR, the global representation capability of the recently emerged convolutional sparse representation (CSR) model reveals huge potential to overcome the above mentioned deficiencies. Drawing on the convolutional neural network (CNN) architecture, a multi-layer CSR model was designed for pixel level image fusion. The image fusion model was constructed with five layers in a forward-feeding manner: the first two layers are CSR layers which acquire sparse coefficient maps with response to the pre-learned dictionary filter sets; the third layer is fusion layer which obtains fused results of the sparse coefficient maps; the last two are reconstruction layers which reconstruct the fused image step by step, and the fusion results are thus obtained. Experimental results indicate that the image fusion method proposed in this paper can effectively overcome the two drawbacks of SR. The method outperforms SR, CSR, and CNN in the aspect of objective assessment metrics, and outperforms SR and CNN in terms of computation complexity and computation time.

Abstract:To study the unsteady effects of rotorcraft on the aerodynamic performance of parachute, a numerical simulation method for the unsteady compound flowfield of rotorcraft-parachute system was established. First, the pressure implicit split operator (PISO) algorithm and Reliazable k-ε model were applied to improve the efficiency of transient calculation and the accuracy of viscosity calculation, and the detail changes in the wake vortex were accurately captured. Subsequently, an efficient dynamic mesh updating model was established, and the Diffusion Smoothing and Remeshing methods were adopted to classify meshes with different deformation scales. Then, the unsteady wake characteristics of the rotorcraft-parachute system and the aerodynamic characteristics of the parachute were studied. Results show that the rotor increased the length of the frontal wake area. The influences of the wake on the parachute were increasing, and the flowfield structure at the entrance of the parachute was asymmetrically distributed. The negative vorticity area at the tail of the rotorcraft gradually moved up and connected with the negative vorticity area at the entrance of the canopy, which promoted the separation of the vorticity from the canopy. The number of vortices in the wake increased significantly. On the other hand, the rotor disrupted the vortex distribution around the rotorcraft, which formed a rotating vortex region. The range of shedding vortex in the wake became smaller. The vorticity magnitude entering the canopy was weakened due to the vorticity viscous dissipation. With increasing rotor speed, the outer pressure of canopy remained the same, but the inner pressure and pressure coefficient gradually decreased, and the pressure difference was reduced, so the average drag coefficient of the parachute gradually decreased.

Abstract:To improve the take-off and landing safety of shipborne helicopters, a numerical method based on one-way coupling strategy was developed for the study of ship-helicopter dynamic interface, with which the effect of different active flow control configurations on helicopter shipboard landing was analyzed under crosswind condition. First, the ship airwake data under different jet control conditions was obtained by using detached eddy simulation (DES). Then, the unsteady airloads and control margins during the shipboard landing were obtained by coupling the ship airwake data with the flight dynamics model. The effect of the active flow control at different locations on ship-helicopter dynamic interface under crosswind condition was investigated, in terms of unsteady loading levels and control characteristics of shipborne helicopter. Results show that the two active jet control configurations could both reduce the unsteady loading levels during the lateral translation of the helicopter. For the configuration that the jet outlet located at hangar vertical edge, the decoupling of unsteady characteristics and spatial characteristics could be realized, and the helicopter control margin was not reduced. While the configuration that the jet outlet located at the horizontal edge could increase the lateral velocity component of ship airwake and reduce the pedal margin. In addition, with the jet velocity increased, the control ability of jet decreased under the first configuration, and it increased first and then decreased under the second configuration.

Abstract:Current swarm intelligence evacuation models only consider single classic swarm intelligence, which is insufficient to describe the complex behavior characteristics of crowd evacuation. In addition, these models rarely take into consideration the impact of crowd chaos on crowd evacuation. In order to study the swarm intelligence evacuation model describing the evacuation behaviors of different groups, by integrating various swarm intelligence algorithms, and taking into account the impact of crowd chaos on evacuation, a crowd evacuation model based on hybrid artificial bee colony-bat algorithm with entropy correction was proposed. Firstly, the density-based spatial clustering of applications with noise (DBSCAN) algorithm was used for group partition. The evacuees were divided into group leader, group members, and disorganized people. Next, according to the characteristics of each type of evacuees, the group leader was described based on bat algorithm, the group members based on artificial bee colony algorithm, and the disorganized people based on particle swarm optimization (PSO). Finally, the evacuation entropy that quantitatively describes the degree of crowd chaos was introduced to correct the position of the group leader, and the evacuation model based on hybrid artificial bee colony-bat algorithm with entropy correction was thus constructed. Simulation results show that the model could well simulate group evacuation, which was basically consistent with the real shape of group evacuation, and the evacuation efficiency was improved by means of group evacuation to some extent. With the introduction of the evacuation entropy for correction, the group leader could guide the group members to avoid the chaotic area ahead, prevent the excessive concentration of evacuees, and enhance the safety and rapidity of evacuation.

Abstract:In recent years, many scene text detection methods based on generic object detection framework have been proposed. These methods usually predict the entire bounding box of the text directly, while it is difficult for them to detect long text effectively due to the limit of receptive field. To solve such problem, a scene text detection method based on short edge vertices regression network was proposed. The method divides the text region into three kinds of regions, namely regions near two short edges and middle region, where separate text regions are firstly predicted and then combined, whereas the entire bounding box of the text is not predicted directly. Specifically, three kinds of regions were segmented on a residual network combined with multi-scale features, and two vertices of a short edge were predicted at each pixel in the region near the short edge. Then the regions near the two short edges were combined on the basis of the adjacent relationship between middle region and short edge regions in the post process, and vertices of short edges predicted by the two regions near short edges were combined to generate complete and accurate detection results. Finally, experiments were performed on a long text detection dataset and several public scene text detection datasets such as MSRA-TD 500, ICDAR 2015, and ICDAR 2013. The proposed method outperformed most of existing methods in accuracy and speed. Experimental results demonstrate that the method has advantages in text detection, especially for long text.

Abstract:To realize the prediction of strongly coupled sound radiation and the optimization of noise in both internal and external sound fields, the finite element/boundary element coupling equation was established, and two methods for solving the sensitivity of sound power were proposed. First, the coupled finite element equations of the structure and the internal sound field were given. Considering the interaction of the external acoustic medium to the structure, the finite element/boundary element coupling equation was established according to the continuity of interface force and normal velocity. Then, in view of the difficulty of decoupling in element sensitivity analysis in sound power topology optimization, the sound power was converted into an expression with structural displacement as variable, and the adjoint equation was modified to realize the extension of the method to the strongly coupled optimization problem. The direct derivation method of sound power sensitivity was proposed considering the complicated derivation process of the adjoint variable method (AVM). Finally, the linearized stiffness method was adopted, which takes the relative density of structural elements as continuous design variables, and the optimization of different structural materials and acoustic media was studied. Comparison results show that the sound power calculated by the coupled model was consistent with that of the finite element method without reflection boundary. Both direct derivation method and AVM could quickly achieve convergence, but the derivation process of the direct derivation method was more concise and efficient. Numerical optimization proves that the optimization algorithm has good applicability and is effective in sound power optimization design.

Abstract:To address the problem that the traditional angle tracking loop based on rate gyros has difficulties in tracking modern high-speed and large maneuvering targets, a novel control method of angle tracking system based on pre-filter was proposed by introducing the concept of Kalman filter. Firstly, the mathematical model was established in body LOS coordinates, and tracking state equations of the angle tracking system were derived. The target maneuver was described by the “current” statistical model, and the misalignment angle was selected as the only observed quantity according to the actual measurable conditions of the seeker. Then the internal factors of low accuracy and early divergence of the traditional angle tracking control method were analyzed, and the improved method of pre-filter was proposed. The discrete state estimator of Kalman filter was designed, and the online real-time iterative solution of seeker was realized. Finally, considering the scenarios that the target adopts different evasive maneuvering strategies accordingly under the attacks of the missile from three different angles, the tracking results were compared. Results show that the improved pre-filter control method could reduce the shock caused by target maneuvering, delay the divergence time of the angle tracking system, further increase the terminal control distance of the seeker, enhance the capability of the seeker against large maneuvering targets, and thus effectively improve the hit probability of the missile.

Abstract:To ensure the high precision output of micro-electro-mechanical system (MEMS) accelerometers, the real-time compensation calibration needs to be implemented in the application process. In this study, an acceleration self-calibration model between the measured value and the real value was established, and multiple sets of static observation samples of the accelerometer at different positions were filtered. The Levenberg-Marquardt (LM) algorithm was combined with the least squares method to calculate the model parameters, which solved the problem of initial value dependence of the LM algorithm. For the static output of the accelerometer at different positions, the posture data that can be used for the least squares method were selected after filtering, which were applied to modify part or all the kth iteration model parameters, being the initial value of the (k+1)th iteration; other posture data were used to train the (k+1)th model parameters by the LM algorithm, realizing the closed-loop and real-time calibration of the accelerometer in application. Taking the application of intelligent insoles as an example, the calibration results of traditional 12-position method,ellipsoid fitting method, LM algorithm, and LM & least squares method were compared. Experimental results show that in the long-term use of intelligent insoles, the proposed LM & least squares method could eliminate the inaccurate calculation of model parameters due to the setting of the initial value of LM algorithm. Besides, it could realize the real-time acquisition, calculation, and calibration of the target, and achieve the precision of the same magnitude as the traditional calibration methods.

Abstract:To improve the overall detection performance of MEMS gyroscope, the impact of phase error in demodulated reference signal on the detection of MEMS gyroscope was analyzed, and a phase correction method based on vector inner product phase detector was proposed. Firstly, the impact of phase error on demodulation results was theoretically analyzed. Then, the simulation models of three typical MEMS gyroscope detection circuits were established, including open-loop demodulation, electrostatic feedback quadrature error cancellation plus open-loop demodulation, and electrostatic feedback quadrature error cancellation plus closed-loop demodulation, and simulation of phase error in demodulated reference signal was carried out. Results show that among the three demodulation methods, the impacts of phase error on angular velocity and quadrature error were basically the same, while those on the dynamic range of gyroscope were different. Under the same input conditions, the larger the phase error was, the more the demodulated angular velocity and quadrature error were deviated from the real value, so it is necessary to carry out phase correction to eliminate such influence. On this basis, the vector inner product phase detection method was introduced into the MEMS gyroscope detection circuit, and a general phase correction method which can be applied in three cases was proposed and simulated. Simulation results indicate that this method can eliminate phase error, calibrate quadrature error coefficient and angular velocity demodulation error, and improve the dynamic range of electrostatic feedback quadrature error cancellation plus open-loop demodulation circuit without affecting the start-up time of MEMS gyroscope, which has good versatility.

Abstract:To improve the computing efficiency of the computational fluid dynamics (CFD) simulation of reactor pressure vessel (RPV), a novel CFD simulation scheme for RPV was proposed by combining the multiple RANS model (MRANS) with CFD simulation. The accuracy of the CFD simulation scheme using MRANS for RPV was guaranteed by optimizing the domain division technology and the data transmission method of traditional MRANS scheme which is used in rod bundle simulation. First, a CFD simulation of the integrated RPV model was carried out, and the validity of the CFD method was verified by comparing the calculated results with the experimental data. Then, based on the domain overlapping approach, the domain division model of downcomer (DC) and lowerplenum (LP) was built, and the modeling of multi-domain CFD simulation was carried out by using the two-way pressure-velocity data transfer method. Next, the performance of different turbulent models in each domain was compared and analyzed aiming to select the appropriate turbulent model. Finally, taking the computing efficiency of CFD simulation as the optimization objective, two MRANS schemes (RSM-SKE and RSM-RKE) were proposed. Results show that compared with the traditional domain division simulation scheme, the velocity and pressure calculated by the two-way domain division simulation scheme based on overlapping had better consistency with the integrated simulation results. The computing efficiency of the optimized MRANS scheme increased by 28.75% compared with the traditional single RANS model scheme, which proved that the optimized MRANS scheme has high efficiency in CFD simulation of RPV. The research results provide a new way of improving the efficiency of CFD simulation for RPV.

Abstract:This paper aims to study the influence of motion parameters on the power-extraction efficiency of oscillating hydrofoils in tidal current, investigate the distributions of the motion parameters of hydrofoils at high power-extraction efficiency, and improve the power-extraction efficiency of oscillating hydrofoils in tidal power-extraction devices. A numerical calculation model of oscillating hydrofoils was established based on the overset grid technique, and the effectiveness of the grid was verified. By modifying the motion parameters of a hydrofoil including reduced frequency f^*, pitching amplitude θ₀, and heave amplitude h₀, the hydrodynamic characteristics of two-dimensional oscillating hydrofoil and the corresponding power-extraction efficiency were investigated. In addition, the influence of vortex shedding on the power-extraction efficiency during the motion of the hydrofoil was analyzed. The pressure distribution when the wing-in-ground effect was formed between two hydrofoils was also studied. The hydrodynamic characteristics of the hydrofoil and the influence of the wing-in-ground effect on the power-extraction efficiency were analyzed under this circumstance. Results show that reasonable planning of the reduced frequency, pitching amplitude, heaving amplitude, and shaft position of the oscillating hydrofoil could all improve the power-extraction efficiency. The generation of vortex shedding during the oscillating motion of the hydrofoil could affect the power-extraction efficiency. The wing-in-ground effect between the parallel symmetrical double hydrofoils would form a high-pressure area in the flow field between the two hydrofoils and increase the lift of the hydrofoils. Increasing the lift-to-drag ratio was beneficial to improve the power-extraction efficiency. When the minimum distance between the two hydrofoils reduced, the power-extraction efficiency of the hydrofoil increased due to the enhancement of ground effect.

Abstract:The current research on the hydrodynamic derivatives in open water is relatively mature, but the research on the hydrodynamic derivatives under ice-water coupling is still in infancy. In order to solve the hydrodynamic derivatives of a ship in oblique motion after ice-water coupling under brash ice conditions, the DEM particle of the CFD software STAR-CCM+ was used to simulate brash ice particles. The momentum and energy exchange was performed to achieve the coupling effect between ice and water by turning on the two-way coupling mode under the DEM module. The small drift angles of 0°, 2°, 4°, 6°, and 8° were selected for the numerical simulation of oblique motion. The effect of the free surface was ignored. The lateral force and turning moment of the ship under open water conditions and brash ice conditions were calculated respectively. The hydrodynamic derivatives were obtained by fitting the dimensionless force and moment at each drift angle. Considering the randomness of the brash ice interference, the maximum and minimum values of the force and moment after the interference were used to solve the corresponding hydrodynamic derivatives, so as to form a hydrodynamic fluctuation interval, which can more accurately predictthe hydrodynamic derivatives of the ship under ice-water coupling. Calculation results show that the hydrodynamic derivatives of the ship at each part under open water conditions were not significantly different from those calculated by the statistical model. The end values of the fluctuation interval of some hydrodynamic derivatives after ice-water coupling were positive and negative. Hydrodynamic derivatives were more random under brash ice-water coupling.

Abstract:In order to realize the adjustable stiffness and strength of compressor impeller and reduce the mass of the impeller, a new lightweight compressor impeller was designed based on the conventional cubic lattice structure. Based on the comparative calculation of three models of cantilever beam, it has been proved that the asymptotic homogenization (AH) method has good accuracy in large-scale lattice calculation. On this basis, the AH method was used to calculate the static performance of lattice impeller with different cell filling rates under the inertial load of 80 000 r/min. Results show that for the studied path, the deformation and stress of the lattice impeller were between those of the unfilled impeller and the solid impeller, and the difference (range) between the maximum and minimum circumferential deformation of the outer edge of the large blade was smaller than those of the unfilled impeller and the solid impeller. When the filling rate was 0.4, the circumferential deformation range of the lattice impeller was 23.25% lower than that of the solid impeller, 55.46% lower than that of the unfilled impeller, and the mass of the lattice impeller could be reduced by 17.08% compared with solid impeller. It indicates that in comparison with the traditional compressor impeller, lattice impeller can not only greatly reduce the impeller mass, but also has better circumferential distortion resistance and higher work efficiency. Meanwhile, the blade edge of lattice impeller has smaller axial stress than those of unfilled impeller and solid impeller, so it can provide strength guarantee for the design of compressor impellers with higher speed. The design of this paper also provides a new design idea for the lightweight design of impeller structures.

Abstract:The 24th Winter Olympic Games will be held in China in February 2022. The main venue is located in Chongli, Hebei Province. The outdoor daily average temperature during the competition is estimated to be-17 ℃－-2 ℃. To effectively improve the thermal comfort and viewing experience of outdoor audiences, the analysis of the cold tolerance characteristics of outdoor audiences is very important. In this study, eight subjects (four males and four females) from a university in Beijing participated in the cold tolerance experiment. The thermal response characteristics of different parts of the human body, the maximum permissible time, and the physiological limit thresholds of the subjects under cold exposure conditions (-14 ℃－-16 ℃) were investigated in a commercial cold store in Beijing. Results show that when the subject was exposed to a cold environment of -14 ℃ to -16 ℃ and the thermal insulation of the clothing ensembles was 2.1 clo, the skin temperature of the fingers and toes had the highest rate of decline and the cold sensation was the strongest; the skin temperature of trunk and limbs decreased less, and the cold sensation was weaker. Under the conditions of hand exposure and hands in pockets, the average maximum permissible time was 51 min and 67 min respectively. The key factors affecting the thermal sensation and permissible time of the subjects were the skin temperature of the fingers and toes, with the average tolerance of human body in extreme cold environments and the construction of corresponding local thermal physiological limit thresholds of 11℃ and 16 ℃ respectively, and no significant difference was found between male and female (P＞0.05). This study can provide basic data reference for the research on the coldcomfort environment.