After a strong earthquake, the first hour of a city's response often depends less on heavy rescue than on controlling utilities. Broken gas lines can ignite, ruptured water mains can flood basements and undermine foundations, and damaged power feeders can energize fallen cables. Operators are under pressure to shut things off quickly, but a blunt shutdown can also cut water to hospitals, fire hydrants, sewage pumping stations and cooling systems. The real question is not whether to isolate, but in what order and in which zones.
A utility digital twin links the network model to the city model. Gas and water pipes, valves, regulators, pumping stations and substations are stored with their connectivity, depth, material and installation age, while the 3D layer shows buildings, roads and ground conditions. When sensor feeds are available, such as pressure drops in district metering areas, flow anomalies or substation alarms, the twin can show which parts of the network have lost integrity and which remain sound. Operators then see not only a broken pipe but the set of valves that would isolate it and the customers who would lose supply.
The most useful work happens before any event. Teams can define isolation playbooks for common scenarios: a gas leak near a school, a burst main under a major junction, or a damaged feeder serving a treatment plant. Each playbook lists the valves to close, the order of closure, the expected pressure changes in neighbouring zones and the critical sites that need a bypass or temporary supply. Running these sequences against the network model reveals side effects, such as a closed valve trapping a low-pressure area that then draws air into the pipes, or a pumping station losing suction.
Scenario testing also exposes gaps in the data. A valve recorded as operable may be seized or buried under a later road resurfacing, and a pipe record may not match the real alignment. Before an exercise, the team can flag unverified assets and schedule field checks, which is far better than discovering the gap during a real shutdown. Priority lists need the same discipline. Hospitals, treatment plants, fire stations and shelters should appear together with their dependencies, so that closing one district does not silently remove power or water from a site meant to stay live.
During an incident, the twin works as a short, auditable decision aid rather than an automatic controller. The control room can compare the proposed sequence with live pressure and gas readings, mark zones as isolated or restored, and log each step for later review. Field crews receive the same map with an ordered valve list, which reduces the risk of two teams closing the same valve on different instructions. Afterwards, the logged sequence becomes evidence for improving the playbooks.
For TwinUp and similar projects, the lasting value lies in keeping these models current between emergencies. Asset registers, network changes and operating rules must be updated as the city changes, or the playbook will describe a network that no longer exists. A shutoff plan is only as reliable as the last verification of its valves and connections, so resilience becomes a routine maintenance task rather than a document kept for a single day of crisis.
