For ultra-reliable, low-latency communications (URLLC) applications in industrial wireless networks such as mission-critical industrial control, it is important to ensure the communication quality of individual packets. Prior studies have considered Probabilistic Per-packet Real-time Communications (PPRC) guarantees for single-cell, single-channel networks but they have not considered real-world complexities such as inter-cell interference in large-scale networks with multiple communication channels and heterogeneous real-time requirements. To fill the gap, we propose a real-time scheduling algorithm based on local-deadline-partition (LDP), and the LDP algorithm ensures PPRC guarantee for large-scale, multi-channel networks with heterogeneous real-time constraints. We numerically study the properties of the LDP algorithm and observe that it significantly improves the network capacity of URLLC, for instance, by a factor of 5-20 as compared with a typical method. Furthermore, the PPRC traffic supportable by the LDP algorithm is significantly higher than that of state-of-the-art comparison schemes. This demonstrates the potential of fine-grained scheduling algorithms for URLLC wireless systems regarding interference scenarios.