Abstract: A computing system determines if an event has occurred. A first window is defined that includes a subset of a plurality of observation vectors modeled as an output of an autoregressive causal system. A magnitude adjustment vector is computed from a mean computed for a matrix of magnitude values that includes a column for each window of a plurality of windows. The first window is stored in a next column of the matrix of magnitude values. Each cell of the matrix of magnitude values includes an estimated power spectrum value for a respective window and a respective frequency. A second matrix of magnitude values is updated using the magnitude adjustment vector. Each cell of the second matrix of magnitude values includes an adjusted power spectrum value for the respective window and the respective frequency. A peak is detected from the next column of the second matrix of magnitude values.

Abstract: A computing system detects an event. (A) A frequency spectrum is computed using a Fourier transform. (B) (A) is repeated a predefined plurality of times with successive windows of observation vectors. Each window of the successive windows includes a subset of the observation vectors. The successive windows include successive subsets selected sequentially in time. (C) An average frequency spectrum is computed from the frequency spectrum computed the predefined plurality of times. (D) A plurality of segmented average frequency spectra is computed from the computed average frequency spectrum. Each segmented average frequency spectrum of the plurality of segmented average frequency spectra is computed for a frequency band of a plurality of predefined frequency bands. (E) When an event has occurred is determined based on the computed plurality of segmented average frequency spectra and a predefined threshold value.

Abstract: A frequency spectrum is computed using a Fourier transform a predefined plurality of times with successive windows of observation vectors, wherein each window of the successive windows includes a subset of observation vectors, wherein the successive windows include successive subsets selected sequentially in time. An average frequency spectrum is computed. A plurality of segmented average frequency spectra is computed, wherein each segmented average frequency spectrum is computed for a frequency band of predefined frequency bands. A distance value is computed using a trained support vector data description model with the segmented average frequency spectra. When an event has occurred is determined based on a comparison between the distance value and a predefined threshold. A vector is computed from the segmented average frequency spectra using t-stochastic neighbor embedding. When an event has occurred based on the comparison, an event indicator is presented on a graph using the vector.

Abstract: A computing system detects an event. (A) A frequency spectrum of observation vectors is computed using a Fourier transform. Each observation vector includes a sensor value. (B) (A) is repeated a predefined plurality of times with successive windows of the observation vectors. Each window of the successive windows includes a subset of the observation vectors. The successive windows include successive subsets selected sequentially in time. (C) An average frequency spectrum is computed from the frequency spectrum computed the predefined plurality of times. (D) A predefined noise filter is applied to the average frequency spectrum to define a filtered frequency spectrum. (E) A distance value is computed between the filtered frequency spectrum and a predefined reference spectrum using a distance computation function. (F) When an event has occurred is determined based on a comparison between the computed distance value and a predefined distance threshold.

Abstract: The operation of a machine can be controlled by performing reconciliation using a cluster of nodes. In one example, a node can receive parent timestamped data from a parent dataset and child timestamped data from child datasets that are children of the parent dataset in a hierarchical relationship. The parent timestamped data and the child timestamped data can relate to an operational characteristic of the machine. The node can generate computer processing-threads. Each computer processing-thread can solve one or more respective reconciliation problems between a parent data point that has a particular timestamp in the parent timestamped data and child data points that also have the particular timestamp in the child timestamp data to generate a reconciled dataset. An operational setting of the machine can then be adjusted based on the reconciled dataset.