Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining causal models for controlling environments. One of the methods includes repeatedly selecting control settings for the environment based on (i) a causal model that identifies causal relationships between possible settings for controllable elements in the environment and environment responses that reflect a performance of the control system in controlling the environment and (ii) current values of a set of internal parameters; and during the repeatedly selecting: monitoring environment responses to the selected control settings; determining, based on the environment responses, an indication that one or more properties of the environment have changed; and in response, modifying the current values of one or more of the internal parameters.

A controller and loop performance monitoring system is coupled to a controller, detects loop performance degradation in time, and diagnoses a cause of the loop performance degradation. If the cause of loop performance degradation is poor controller tuning, a re-tuning mechanism is triggered. If the cause of loop performance degradation is external to the controller (a disturbance acting on the loop, hardware malfunction etc.), an action defined in control strategy is taken, or the user is informed via alarm, user interface, or upper layer software that collects the performance measures. The monitoring itself is designed to be recursive and with low memory demands, so it can be implemented directly in the controller, without need for data transfer and storage. The monitoring is modular, consisting of oscillation detection and diagnosis part, performance indices part, internal logic part, and triggering part, easily extensible by other performance indices or parts (e.g. for overshoot monitoring). The oscillation detection and diagnosis part includes controller output oscillation monitoring, the performance indices part includes predictability index and offset index. The outputs of the controller and loop performance monitoring are overall loop performance together with loop diagnosis information, and overall controller performance together with controller diagnosis. The outputs of the controller and loop performance monitoring are used as parts of controller and loop performance monitoring user interface.

A system is provided for determining causes of faults in a manufacturing system. The system stores data associated with a processing system which includes machines and associated processes, wherein the data includes timestamp information, machine status information, and product-batch information. The system determines, based on the data, a topology of the processing system, wherein the topology indicates flows of outputs between the machines as part of the processes. The system determines information of machine faults in association with the topology. The system generates, based on the machine-fault information, one or more fault parameters which indicates frequency and severity of a respective fault. The system constructs, based on the topology and the machine-fault information, a system model which includes the one or more fault parameters, thereby facilitating diagnosis of the processing system.

A system and method are provided for determining a causal inference in a manufacturing process. During operation, the system can receive data associated with a processing system which includes a set of interconnected machines and an associated set of processes. The system can generate, based on the data, a graph indicating flows of outputs between the machines as part of the processes. The system can determine, based on a set of variables, one or more candidate clusters in the graph. The system can perform, based on one or more variables of interest, root cause analysis on the one or more candidate clusters by: applying an additive noise model to prune the one or more candidate clusters from the graph; and determining, based on the pruned graph, a candidate pathway likely to cause an issue in at least one process, thereby facilitating improved efficiency in the processing system.

A system for detecting faults in building equipment includes an integration fault detector, a kernel density fault detector, and a fault detector selector. The integration fault detector is configured to detect faults in the building equipment by analyzing time series data using an integration fault detection technique. The kernel density fault detector is configured to detect faults in the building equipment by analyzing the time series data using a kernel density estimation fault detection technique. The fault detector selector is configured to select the integration fault detector or the kernel density fault detector for use in detecting faults in the building equipment based on an attribute of the time series data.

A method for monitoring the operation of a turbomachine controlled by a digital control system including at least one component, includes acquiring operating state information relating to the state of at least one component; determining, depending on the state information acquired, a current degraded configuration in which at least one of the components has failed; determining a classification of the current degraded configuration using at least one classification table stored in a storage device, the classification tables associating with at least one degraded configuration one classification expressing the level of criticality of the degraded configuration, the tables being obtained by calculating a conditional probability of a predefined anticipated event from the probability of occurrence of elementary events relating to a failure of one of the components; and estimating an operating time permitted for the turbomachine depending on the classification determined for the current degraded configuration.

A diagnostic device includes a storage unit that stores first information including a first abnormal event which occurred in the past in a plant, a first attribution event that is a cause of the first abnormal event, and a first occurrence probability of the first attribution event, in which a causal relationship between the first abnormal and attribution events is indicated by a tree structure, and second information including a second abnormal event which is supposed to occur in the plant but has not yet occurred, a second attribution event that is a cause of the second abnormal event, and a second occurrence probability of the second attribution event, in which a causal relationship between the second abnormal and attribution events is indicated by a tree structure; and an estimation unit that estimates the cause of the sign of the abnormality, based on the first and second information.

A method for modelling technical systems having a plurality of technical components, including the step of assigning a component Markov chain to each component having a Markov chain for representing various states of the respective component, at least one input one failure mode for externally triggering a transition from one state of the Markov chain into another state of the Markov chain, and at least one output failure mode to each Markov chain for propagating failures to other components, is provided.

A method for generating a fault tree of a multi-component system is provided. The multicomponent system includes a logical-functional system layer and a physical system layer as different layers of abstraction. The physical system layer may correspond, for example, to software and/or hardware implementing the functional aspects of the logical-functional system layer. The method first provides a logical-functional fault tree for the logical-functional system layer and a physical fault tree for the physical system layer, the latter having elements corresponding to elements in the logical-functional fault tree. Next, a mixed-layer fault tree is generated by combining aspects of both fault trees in a systematic way. The disclosed is particularly relevant for analyzing safety-critical systems. However, the present concepts are not limited to these applications and may be applied to general use cases where fault tree analysis is applicable.

A disturbance source positioning method for positioning disturbance sources in a system including a plurality of nodes is provided. The method includes the following steps: grouping the plurality of nodes into a plurality of node groups based on an oscillation feature; establishing an in-group causality of the plurality of node groups based on a successive order of a coherent oscillation component; selecting at least one candidate group from the plurality of node groups based on the in-group causality; and positioning at least one disturbance source node in each candidate group.