To investigate the composite performance of the track slab and the self-compacting concrete filling layer in the China Railway Track System III slab track structure, six segmental specimens comprising either one or a pair of sleepers were fabricated by dividing a full-scale slab track structure. Fatigue tests in the transverse bending perpendicular to the rail were conducted on four slab track specimens supported on rubber plates simulating the subgrade, to a maximum of five million applied load cycles. It was observed in the fatigue tests that longitudinal cracks first appeared on the top surface of the track slab at the mid-span of the specimen. Splitting cracks appeared later from the anchorage ends of the transverse prestressed tendons to the top surface of track slab. Delamination of the track slab–filling layer interface developed with increasing number of cycles. Static three-point bending tests were also conducted on both fatigue-damaged and control composite plate specimens to evaluate the effect of fatigue damage on the composite performance of the track slab and the filling layer. In the static tests, significant deformation developed in the fatigue-damaged specimens from the start of loading, and the fatigue damage and interface delamination adversely affected the composite action of the track slab and the filling layer. Static loads corresponding to new cracking of the fatigue-damaged composite plates were reduced by 20–30%, and the load corresponding to the initial interlayer slip reduced by 25–62% compared with the control specimens. The interfacial bond strength between the track slab and the filling layer is the first line of defense against the interlayer slip, and the dowel action of the interlayer connection reinforcement is the second line of defense. Roughening by multitooth chiseling on the track slab and an increase in the ratio of interlayer connection reinforcements are both effective ways to enhance the composite action of the track slab and the filling layer.

Unsupervised graph representation learning (GRL) aims at distilling diverse graph information into task-agnostic embeddings without label supervision. Due to a lack of support from labels, recent representation learning methods usually adopt self-supervised learning, and embeddings are learned by solving a handcrafted auxiliary task (so-called pretext task). However, partially due to the irregular non-Euclidean data in graphs, the pretext tasks are generally designed under homophily assumptions and cornered in the low-frequency signals, which results in significant loss of other signals, especially high-frequency signals widespread in graphs with heterophily. Motivated by this limitation, we propose a multi-view perspective and the usage of diverse pretext tasks to capture different signals in graphs into embeddings. A novel framework, denoted as Multi-view Graph Encoder (MVGE), is proposed, and a set of key designs are identified. More specifically, a set of new pretext tasks are designed to encode different types of signals, and a straightforward operation is proposed to maintain both the commodity and personalization in both the attribute and the structural levels. Extensive experiments on synthetic and real-world network datasets show that the node representations learned with MVGE achieve significant performance improvements in three different downstream tasks, especially on graphs with heterophily.