Logic flowgraph methodology (LFM) is intended to provide a more efficient way of constructing failure models for use in a diagnosis oriented disturbance analysis system. The LFM approach represents a considerable development beyond previous methods and also may be useful in reliability and risk analysis applications. Like the digraph method, LFM produces process models in which the fundamental units of nodes and edges are used to represent process variables and causality relations, respectively. In LFM, however, a more extended set of representation rules allows one to achieve a greater level of modeling capability and flexibility. The LFM models hinge on the interconnection of two distinct networks, namely, the “causality network” and the “condition network.” In a formally defined and organized way the condition network represents the conditions whose occurrence can change or modify the course of process causality flow in the causality network. A test case demonstrates the applicability of LFM to situations of interest in nuclear power plant operation and also shows that once a suitable process flow graph model has been derived, it is possible to obtain any fault-tree structure whose top event can be expressed as a weak or strong perturbation on one of the variables constituting a flowgraph node. This fault-tree construction is performed automatically by a computer routine, accepting as input the logic flowgraph topology and the top event of the desired fault tree. In a disturbance analysis application, this routine also accepts as input a set of field instrumentation signals; using this information on line identifies within a fraction of a second the prime cause of the disturbance by logically developing only those tree branches that the instrumentation indicates as active. In reliability or risk analysis applications, on the contrary, the desired fault tree is developed to its full extent.