One failure is never local.
Water supply, gas distribution and electricity networks consist of thousands of physically connected objects. A failure in one component can have consequences far beyond its location. Mantikon is designed to maintain a dynamic model of the network that joins geography, asset characteristics, repair history, incident reports and crew availability.
- Next waveDesigned as an application area; a demonstrated world model in this domain comes first.
- Who decidesUtility operator, asset management, field service
- Typical sourcesGIS, technical asset register, work orders, incident reports, permitted meter data
- 01
Read the signals together
Several sensors reporting an unusual change in pressure become one hypothesis about the network, checked against topology, incident history and consumption.
World model - 02
Compare the explanations
One agent weighs a leak, another a faulty instrument, a third a change in consumption. The brief lists the sections worth inspecting and the order in which to inspect them.
Alternatives - 03
Plan the renewal
Asset age, failure intensity, maintenance cost and the consequences of a failure are compared so that capital is allocated by total infrastructure risk, not by age alone.
Capital
A worked
example
Illustrative scenario, not a customer case.
In a city water system several sensors record an atypical change in pressure. Mantikon sets the signals beside the network topology, the history of failures, the results of previous repairs and consumption data.
One agent analyses the probability of a leak. Another tests the hypothesis of a faulty measuring device. A third examines a possible change in the consumption regime. The brief proposes a list of sections to survey and a sequence of diagnostic actions.
If a pipeline fault is confirmed, the platform helps the dispatcher estimate how many consumers are affected, what repair materials are available, what crews are needed and what the optimal work schedule looks like.
Scope and
measures
Example workflows
Anomaly triage · leak and fault localization · crew dispatch · renewal programme planning.
Evaluation metrics
Time to localize · repeat failures · consumers affected per incident · share of capital spend justified by risk rather than age.
Discuss a network operations use case