Lattice structures have become a research highlight in the fields of aviation, aerospace and automobile because of its high specific strength and stiffness. The development of additive manufacturing technology makes it possible to manufacture mechanical metamaterials with complex geometric shapes and special mechanical properties. However,traditional additive manufacturing processes are difficult to coordinate the contradiction between the cost, geometric size and accuracy of such structures. In view of this, a mechanical metamaterial structure concept of reversible assembly and the connection method of its basic voxels were proposed, which can effectively overcome the size limitation of manufacturing equipment, and can construct a macro lattice structure with rich mechanical properties through different combination sequences. Aiming at the basic voxel type of reversible assembly, the influence of voxel geometric parameters, relative density and voxel numbers on its mechanical properties through parametric modeling method was analyzed, and the experimental results were compared. The results show that the three basic voxels have great rigidity difference, and Poisson’s ratio covers the range from negative to positive. Therefore, different combination sequences of these three basic voxels can produce macro structures with different mechanical property distributions, which proves the rationality of using these three basic voxels as the basic voxels of the reversible-assembly structure. In addition, the mathematical relationship between geometric parameters and mechanical properties can provide guidance for the engineering application of such mechanical metamaterials.

An online measurement method of a dynamic unbalance force of the geotechnical centrifuge was carried out based on monitoring the change of the axial preload of anchor bolts. Firstly, the change model of the axial preload of the anchor bolt under different unbalanced force states was deduced theoretically, which proved the feasibility of indirectly measuring the unbalanced force of the centrifuge. Then, the typical testing case was carried out on the TLJ-500 centrifuge, and the law of the dynamic unbalance moment changing with the counterweight was obtained. By fitting the change of the axial force at the measuring point, the relation between the axial force at the measuring point and the unbalance mass and the running gravitational acceleration g value of the centrifuge was established. Finally, the relation between the change of the axial preload of anchor bolts, acceleration of the centrifuge and counterweight mass were obtained by measuring the preload change of one anchor bolt and fitting the response surface. The results show that it is feasible to obtain the unbalanced force of the centrifuge operation by monitoring the change of anchor bolts’ pre-tightening force，and the centroids of different counterweights should be kept consistent in the test. When calculating the counterweight mass to balance the rotating arm according to the fitting formula, a larger acceleration should be used as far as possible to reduce the error caused by the centrifuge not being leveled.

The crack tip constraint and failure assessment diagram （FAD） in both plate specimen and pipe structure were studied to find out the compatibility and correlation of between the sample and piple. Based on the finite element method,constraints at crack tip of central cracked plate （CCP） specimen and cracked pressure pipe were obtained. It was found that crack tip constraint of CCP specimen under uniaxial loading is better matched with thick-walled pipeline under axial tensile loading, while CCP specimen under biaxial loading is better matched with thin-walled pipeline under internal pressure. In addition,the failure assessment diagram of CCP specimen and pipeline were obtained. It was found that FAD of CCP specimen is less related with thickness, when the thickness of pipe is within the range of thick wall or thin wall; was also pointed out that stress parallel to the crack direction increases the failure assessment curve（FAC） of the structure. Finally, it is found that FAC of the central crack plate specimen is higher than that of the pressure bearing pipeline, for the two structures with the same constraint level，and crack tip mixed mode degree and loading mode. That is, it is unsafe to use FAD of CCP specimen to evaluate pipeline.

In order to clarify the influence of different working conditions on the wear of tapered roller bearings, the wear state of tapered roller bearings was carried out. A quasi-statics model of tapered roller bearings including roller balance equation and inner ring balance equation was established, and a wear depth calculation model suitable for wear analysis of tapered roller bearings was derived based on Archard wear theory. The wear distribution characteristics of bearing inner and outer raceways and roller elements under light load, medium load and heavy load conditions were comparative analyzed. The results show that the wear depths of the tapered roller and the inner/outer raceways on the contact line show bimodal distribution characteristics, and the wear distribution of the outer raceway is closely related to the positions of the rolling elements. For light load condition, the peak wear depth of bearing raceways and roller elements increases by 20% as bearing speed increases 25%, and for heavy load condition, the peak wear depth of bearing raceways and roller elements increases more than 17% as bearing load increases 16%. The comparison of bearing wear peaks under the three working conditions shows that the product of the equivalent load and the speed of the bearing is not the determining factor controlling bearing wear，and the load change has a significant impact on the bearing wear.

In order to solve fatigue fracture problem of the transverse stabilizer bar link of a pickup truck, taking the measured load spectrum as support, the fatigue analysis and structural improvement of bar link was carried out based on multi-body dynamics load decomposition method. The load spectrum acquisition test of stabilizer bar link was carried out, and the correlation analysis between sensor signals was conducted, and the validity of the collected signals was verified. Through the kinematics compliance（KC） test, the accuracy of suspension dynamics model was verified, by using virtual iteration method, the accuracy of the vehicle dynamic model is verified by the measured load spectrum signal as excitation. The critical load of bar link was obtained by means of multi-body dynamics load decomposition method, compression rod instability theory method and bench test method respectively, the strength of link was checked and compared with the measured load value, and the accuracy of the three methods was evaluated. The results show that multi-body dynamics load decomposition method can accurately predict the critical load, the bench test method have some errors, and the the compression rod instability theory cannot be directly used for strength check of link. Fatigue simulation and optimization of link were carried out by combining nominal stress method and load spectrum of multi-body dynamics decomposition. The fatigue mileage predicted by simulation is 12 143 km, the relative error value with real life is 14%, which proves validity of simulation method. The structural improvement of the dangerous part of link body port eliminates stress concentration of it, and makes the design life of link meet the service requirements.

In order to investigate the relation between abrasion and crack initiation in the welded zone of rails and axle weight and friction coefficient, through the field measurement of a large number of uneven weld zones, two kinds of typical weld zone uneven data were fitted, and two kinds of typical uneven wheel-rail contact finite element models of upper convex and lower concave weld zones were established. Combined with the friction work model and Archard wear theory, the maximum wear cross section of the welded zone was predicted, and the crack initiation life of the welded zone was predicted based on the Jiang-Sehitoglu model. It is found that with the increase of axle load, the abrasion rate of both the upper convex and lower concave weld zone increases, and the abrasion rate of the upper convex weld zone increases significantly when the axle load reaches 16 t, while the abrasion rate of the lower concave weld zone increases significantly when the axle load reaches 18 t. When the friction coefficient increases from 0.2 to 0.35, the maximal abrasion amount of the two types of weld zones of the upper convex and the lower concave is 1.93 mm and 1.08 mm, respectively. The abrasion rate of the upper convex type increases significantly at a friction coefficient of 0.3, while the abrasion rate of the lower concave type increases significantly at a coefficient of 0.35. When the axle load increases from 12 t to 18 t, the service life of the upper convex weld zone decreases less, while the service life of the lower concave weld zone decreases more. In addition, when the friction coefficient is increased from 0.2 to 0.35, its effect on the service life of the upper convex weld zone is significantly smaller than that of the axle load （12-18 t）. However, when the friction coefficient is increased from 0.2 to 0.35, the effect on the service life of the lower concave weld zone is comparable to that of the axle load （12-18 t）. The results show that with the increase of axle load and friction coefficient, the influence of concave irregularity on the life of rail welding zone is more significant. In the process of engineering maintenance, we should pay attention to the appearance of concave welding zone and mark and repair it in time.