Abstract: In various example embodiments, techniques are provided for efficient and reliable anomaly detection and evaluation in a water distribution system (e.g., a smart water distribution system) using both flow and pressure time series data from sensors of the system. The techniques may implement a multi-step workflow that involves decomposing the time series data to remove seasonality and rendering the time series data stationary, detecting outliers of the stationary time series data, classifying sensor events in response to flow or pressure of detected outliers exceeding high or low thresholds for at least a given number of time steps, classifying anomaly events by correlating one or more sensor events related to flow with one or more sensor events related to pressure or by clustering a plurality of sensor events in temporal proximity, and determining a quantitative score for each of the detected anomaly events that indicates a level of significance or importance.

Abstract: In example embodiments, an enhanced deep belief learning model with an extended Kalman filter (EKF) is used for training and updating a deep belief network (DBN) with new data to produce a DBN model useful in making predictions on a variety of types of datasets, including data captured from infrastructure-attached sensors describing the condition of the infrastructure. The EKF is employed to estimate operation parameters of the DBN and generate the model's output covariance. Further, in example embodiments, the configuration of the DBN model may be optimized by a competent genetic algorithm.

Abstract: In various example embodiments, techniques are provided for training and/or using a semantic deep learning model, such as a segmentation-enabled CNN model, to detect corrosion and enable its quantitative evaluation. An application may include a training dataset generation tool capable of semi-automatic generation of a training dataset that includes images with labeled corrosion segments. The application may use the labeled training dataset to train a semantic deep learning model to detect and segment corrosion in images of an input dataset at the pixel-level. The application may apply an input dataset to the trained semantic deep learning model to produce a semantically segmented output dataset that includes labeled corrosion segments. The application may include an evaluation tool that quantitatively evaluates corrosion in the semantically segmented output dataset, to allow severity of the corrosion to be classified.

Abstract: In various example embodiments, techniques are provided for crack detection, assessment and visualization that utilize deep learning in combination with a 3D mesh model. Deep learning is applied to a set of 2D images of infrastructure to identify and segment surface cracks. For example, a Faster region-based convolutional neural network (Faster-RCNN) may identify surface cracks and a structured random forest edge detection (SFRED) technique may segment the identified surface cracks. Alternatively, a Mask region-based convolutional neural network (Mask-RCNN) may identify and segment surface cracks in parallel. Photogrammetry is used to generate a textured three-dimensional (3D) mesh model of the infrastructure from the 2D images. A texture cover of the 3D mesh model is analyzed to determine quantitative measures of identified surface cracks. The 3D mesh model is displayed to provide a visualization of identified surface cracks and facilitate inspection of the infrastructure.

Abstract: In an example embodiment, an analysis application is used to optimize sensor placement by minimizing information entropy or maximizing total modal energy. These objectives are achieved by implementing a two-part optimization procedure, involving generating an evaluation database that stores an information matrix, and using the evaluation in conjunction with a genetic algorithm to produce an optimized sensor location set.

Abstract: In one example embodiment, an analysis application is used to optimize water quality sensor placement in a water distribution network by implementing a two-part optimization solution procedure, involving building an impact database, and determining an optimized water quality sensor location set using the impact database. The optimized sensor location set may indicate locations that maximize a length of pipes where water quality variations are detectable by at least one water quality sensor. Pipe wall reaction coefficients may be used as calibration parameters, with water quality indicated to be detectable at a possible sensor location when a change in its pipe wall reaction coefficients leads to a change in water quality at the possible sensor location that is greater than a threshold.

Abstract: In one embodiment, a technique is provided for optimizing selection of hydrants for flow test in a water distribution system. An impact database is built that indicates whether a flow test at each hydrant in the water distribution system has an impact on each pipe of the water distribution system. When a user supplies a number of hydrants to be subject to flow test, a hydrant selection solver application may search for an optimized set of hydrants for flow test that includes the user-supplied number, the searching to include generating candidate sets of hydrants and evaluating the candidate sets of hydrants based on fitness values calculated using the impact database, each fitness value to evaluate performance of a candidate set of hydrants based on flow velocity or hydraulic gradient change in pipes.

Abstract: In an example embodiment, a dynamic feature extraction tool receives a data set from a SHM system that includes a plurality of sensors affixed to a structure (e.g., a bridge, dam, building, etc.), the data set including at least one of ambient vibration data or earthquake vibration data. A solution method is selected from among, for example, time domain analysis, frequency domain decomposition or eigensystem realization analysis. The dynamic feature extraction tool guides a user to select at least one parameter value used in the selected solution method from a subset of determined-effective parameter values computed by the software tool. The dynamic feature extraction tool then automatically performs the selected solution method on the data set using the selected at least one parameter value to determine dynamic features (e.g., frequencies or modal shapes), and displays a graphical representation of the dynamic features in a UI.

Abstract: In one example embodiment, an analysis application is used to optimize sensor placement by implementing a two-part optimization solution procedure, involving generating a contribution database, and determining an optimized sensor location set using the contribution database. The optimized sensor location set may indicate locations that maximize coverage of dynamic integrity, which is quantified by as a ratio of detectable damage scenarios to all damage scenarios used by the analysis application.

Abstract: A method and system for performing a criticality analysis of a water distribution network is provided. The method and system provides for segmentation of the system which allows a user to determine the set of elements that comprise segments, which in turn are the smallest portion of a water distribution system that can be isolated by valving. Isolating valves are included as elements in the set of elements that are used by an associated hydraulic solver engine to segment the water distribution network. Once the network has been segmented, a criticality analysis is performed whereby a hydraulic simulation is run for an outage of one or more segments, and the shortfall in demand supplied to other segments is calculated. The system provides for a linking of the ability to automatically identify segments with a hydraulic analysis model to enable a user not only to identify segments, but to rank their importance based on a variety of user defined metrics.

Abstract: In one embodiment, a hydraulic simulation model corresponding to a real-world hydraulic network is loaded in a hydraulic modeling and simulation application executing on a computer system. The hydraulic simulation model represents leakages as pressure dependent emitter flow at selected nodes (leakage nodes). Optimization criteria include a specified maximum of possible leakage nodes. A genetic algorithm (GA) generates trial solutions for an optimization, each trial solution representing locations for leakage nodes and corresponding emitter coefficients. A hydraulic analysis is performed for the trial solutions to generated model-simulated results. The model-simulated results are compared to field-observed data for the real-world hydraulic network to generate goodness-of-fit values. The process is repeated until a particular goodness-of-fit value is achieved or a maximum number of iterations is reached.

Abstract: In one embodiment, a technique is disclosed for calculating a relative pump speed factor for attaining a prescribed hydraulic head or for pumping a prescribed amount of flow. A hydraulic model of a water distribution or collection system is defined to include link elements and node elements. At least one of the node elements represents a fixed-flow variable speed pump (VSP) that delivers a desired amount of flow, a variable speed pump battery (VSPB) that represents multiple VSPs operating in parallel with each other, a VSP with a tank located on the VSP's discharge side, or a VSP with a tank located on the VSP's suction side.

Abstract: A computer software program provides an algorithm that solves for unknown demands (and junction pressures) within a modeling system that uses a generalized, unified loop-node formulation. The program can be used to calculate the available demand (i.e., the amount of water that is to be supplied) according to the nodal pressure. Both nodal heads and flows are simultaneously solved using a gradient algorithm, which allows, in accordance with the present invention, the model to simulate situations where a change in pressure affects the quantity of water used. Criticality analyses for segments of a system in such pressure dependent scenarios can also be performed using the software program of the present invention.

Abstract: A water distribution model calibration technique is provided that allows a user to design a calibration model by selecting several input parameters desired to be used for the calibration of a model that allows an engineer to collect a complete set of data to represent the overall system conditions at any given time of day. For example, several parameters may be chosen including link status, the pipe roughness coefficient, junction demand, and pipe and valve operational status. Trial solutions of the model calibration are generated by a genetic algorithm program. A hydraulic network solver program then simulates each trial solution. A calibration module runs a calibration evaluation program to evaluate how closely the model simulation is to the observed data.

Abstract: A system and method of automatically calibrating a water distribution model is provided that allows a user to design a calibration model by selecting several input parameters desired to be used for the calibration. For example, several parameters may be chosen including the pipe roughness coefficient, junction demand, and pipe and valve operational status. Trial solutions of the model calibration are generated by a genetic algorithm program. A hydraulic network solver program then simulates each trial solution. A calibration module runs a calibration evaluation program to evaluate how closely the model simulation is to the observed data. In doing so, the calibration evaluation program computes a “goodness-of-fit” value, which is the discrepancy between the observed data and the model data, for each solution. This goodness of fit value is then assigned as the “fitness” for that solution in the genetic algorithm program.