The odor exploration method explores an odor based on odor information generated using a plurality of sensor elements each outputting a detection signal according to the state of adsorption by indicating an adsorption reaction unique to each odor substance, and an arithmetic unit generating the odor information by quantifying each detection signal from the sensor elements. The method includes a detection step detecting a first gas containing odor substances, a generation step generating a first odor information group generated based on a detection result of the first gas, a preparation step preparing a second odor information group generated based on a detection of a second gas, and a calculation step calculating a sum of or a difference between the second odor information group and the first odor information group based on a sum of or difference between the odor information generated for the first gas and for the second gas using the same sensor element.

A crack detection device performs online identification of the occurrence and propagation of a crack in the surface of the barrel portion of a rolling roll. A rolling roll is provided with the crack detection function without any substantial modification of the rolling device and without any continuous disposition of multiple sensors in the rolling roll. The detection device is incorporated in a rolling device having a barrel portion and shaft portions extending as a unit from both ends of the barrel portion and includes the rolling roll where an AE sensor detecting elastic waves generated on a surface of the barrel portion and a calculation unit calculating a feature value of the elastic waves detected by the AE sensor are disposed in at least one of the shaft portions and a discrimination unit discriminating, from the feature value, elastic waves attributable to a crack occurring in the barrel surface.

Methods and systems for anomaly detection include training a neural network model to identify a form of data augmentation that has been performed on a waveform. Multiple forms of data augmentation are performed on a sample waveform to generate data augmentation samples. The data augmentation samples are classified with the neural network model. An anomaly score is determined based on the classification of the data augmentation samples.

A method for predicting a remaining life of a tool of a computer numerical control machine is provided. In the method, indirect measurement indicators of the tool are selected based on monitoring and analyzing a current state of the tool, a prediction model for the remaining life of the tool is established based on data de-noising, feature extraction and a multi-kernel W-LSSVM algorithm. Thereby, a method for predicting a remaining life of a tool of a computer numerical control machine is provided.

A strength inspection device for evaluating a tensile strength of a test body as a fiber reinforced composite material includes: an AE sensor that detects AE waves generated in the test body by a tensile load in a test period of application of the increasing tensile load to the test body, and generates waveform data of the AE waves; a target wave specifying unit that specifies, as target waves, the AE waves of duration longer than a time threshold, based on the waveform data; an arithmetic unit calculates a frequency center of gravity concerning each target wave; and an evaluation data generation unit generates strength evaluation data of association between the frequency center of gravity concerning each target waves and magnitude of the tensile load applied to the test body at a detection time point of the target wave.

An acoustic inspection apparatus includes a vibration sound source, a microphone group, and a processor. The vibration sound source emits a vibration sound to an inspection target object. The microphone group includes a first microphone arranged near the inspection target object and a second microphone arranged to have an interval with respect to the first microphone. The processor calculates a first impulse response between the first and second microphones, denoises a component corresponding to the vibration sound from the first impulse response, converts, into a frequency characteristic, a second impulse response obtained from the first impulse response, calculates acoustic energy between the first and second microphones based on the frequency characteristic, and determines an abnormal state of the inspection target object based on the acoustic energy.

A measuring method in an embodiment includes detecting elastic waves generated in a structural object formed of a solid material in a non-contact manner and estimating a position of an occurrence source of the elastic waves on the basis of information regarding an intensity of a signal of the detected elastic waves.

According to one embodiment, an inspection apparatus includes a vibration sensor, a microphone, and a processor. The vibration sensor detects a vibration of an inspection target object to which the vibration is excited. The microphone arranged near the inspection target object and collects a radiated sound from the inspection target object. The processor calculates an impulse response between the vibration sensor and the microphone. The processor denoises an unnecessary component from the impulse response. The processor converts the impulse response into a frequency characteristic. The processor calculates acoustic energy between the vibration sensor and the microphone based on the frequency characteristic. The processor determine the presence/absence of an abnormal state of the inspection target object.

Methods, systems, and computer-readable storage media for accurate time delay estimation using coherent averaging. A plurality of out-of-bracket acoustical impulses are generated in a pipe segment of a fluid distribution system. Signal data representing the acoustical impulses sensed at two locations along the pipe segment are recorded. Precise timings for the generation of the acoustical impulses are obtained, and the acoustical impulses in the signal data recorded from the first location are averaged based on the precise timings to produce a near-sensor average impulse. Similarly, the acoustical impulses in the signal data recorded from the second location are averaged based on the same precise timings to produce a far-sensor average impulse. A time delay between arrival of the plurality of out-of-bracket acoustical impulses at the first and second locations is estimated from the timing of the near-sensor average impulse and the far-sensor average impulse.

Embodiments of the invention include a system and method for detecting wood-boring species of insects in a structure (16), involving one or more primary (14) and reference (18) sensors and a signal conditioning and acquisition device (22) capable of being coupled to the sensors (14, 18). The system (10) also includes a processor (24) capable of being coupled to a non-transitory, computer-readable storage medium that includes program logic for execution by the processor (24). The program logic includes a logic module that receives signals originating from the sensors (14, 18) and discriminates between noise generated by any wood boring species in the structure and extraneous noise unrelated to the wood boring species of insects. The extraction of signal features based on pulse duration, signal spectra and signal envelope spectra can be used for insect pulse discrimination. A sound-suppressing sensor assembly (12) can be weighted to enhance the coupling of the primary sensor (14) with the structure (16) being tested.