The fourth industrial revolution aims to realize industrial internet-of-things (IIoT) and automatic manufacturing systems. However, the stringent ultra-reliability low-latency communications (URLLC) requirements and severe mutual interference are two major challenges for industrial private networks. To achieve a broad range of industrial connections, we aim to investigate the method to realize the rate splitting multiple access (RSMA) for overloaded scenarios. A multi-antenna central controller (CC) is employed to serve multiple devices with the short blocklength regime. With the objective of maximizing weighted sum throughput over all devices, the cooperative beamforming and rate splitting schemes tailored for overloaded scenarios are joint optimized, subject to the constraints of resource allocation and diverse quality-of-service (QoS) requirements. To tackle the complexity of the fractional rate expression with channel dispersion, we propose a universal transform method, where tractable expressions for the original problem are efficiently obtained. The nonconvex optimization problem is solved iteratively using sequential convex approximation and interior-point methods. Simulation results confirm the effectiveness of the proposed method and show that RSMA can significantly improve performance in overloaded URLLC compared with other multi-antenna multiple access schemes.