Monitoring industrial equipment using audio
地域：
NJ
NJ Morris Plains
摘要
Systems, methods, and devices for monitoring industrial equipment using audio are described herein. One system includes two computing devices. The first computing device can receive, from an audio sensor, audio sensed during operation of industrial equipment, extract a plurality of features from the audio, determine whether any portion of the audio is anomalous, and send, upon determining a portion of the audio is anomalous, the anomalous portion of the audio to the second, remotely located, computing device. The second computing device can provide the anomalous portion of the audio to a user to determine whether the anomalous portion of the audio corresponds to a fault occurring in the equipment, and receive, from the user upon determining the anomalous portion of the audio corresponds to a fault occurring in the equipment, input indicating the anomalous portion of the audio corresponds to the fault to learn fault patterns in the equipment.
說(shuō)明書
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
The present application is a continuation of U.S. application Ser. No. 16/033,883, filed Jul. 12, 2018, entitled, “MONITORING INDUSTRIAL EQUIPMENT USING AUDIO,” which is hereby incorporated by reference.
The present disclosure relates generally to systems, methods, and devices for monitoring industrial equipment using audio.
A heating, ventilation, and air conditioning (HVAC) system can be used to control the environment within a facility (e.g., building), and is one example of an industrial system that may be associated with the facility. For example, an HVAC system can include a number of components (e.g., equipment) that can be used to control the air temperature of different zones (e.g., rooms, areas, spaces, and/or floors) of a facility, in order to keep the zones in a comfort state for their occupants. As an additional example in which the facility is a retail facility, the HVAC equipment may be used to continuously operate (e.g., cool) commercial refrigerators and/or freezers.
During operation of an industrial system such as an HVAC system, however, faults in the system (e.g., in the equipment of the system) may sometimes occur. Detecting and correcting faults in the equipment of the system can be important to provide and maintain a comfortable environment for the occupants of the facility, to prevent the fault from causing further damage to the system, to prevent downtime in the equipment that may lead to significant loss, and/or to avoid inefficient operation of the system which may result in higher energy consumption, for example.
權(quán)利要求
What is claimed is:1. A method for monitoring industrial equipment in a facility, comprising:receiving, from an audio sensor, an audio stream sensed by the audio sensor of an operation of industrial equipment in the facility; sending at least part of the audio stream to a trained machine learning module, wherein the trained machine learning module comprises an audio corpus of different classifications including at least one fault classification, and compares characteristics of at least part of the audio stream to the audio corpus of different classifications to identify one or more matching classifications, if any, the trained machine learning module is further configured to identify that the audio stream is indicative of a fault in the industrial equipment when the trained machine learning module identifies one or more matching classifications that also correspond to one or more of the fault classifications; and providing an alert to a user when the machine learning module determines that the audio stream is indicative of a fault in the industrial equipment. 2. The method of claim 1 , wherein the machine learning module has a training mode and a test mode. 3. The method of claim 1 , wherein at least one of the different classifications corresponds to a non-fault classification. 4. The method of claim 1 , wherein when one or more matching classifications are not identified, sending the audio stream for classification, and then incorporating the audio stream and the resulting classification to the audio corpus. 5. The method of claim 1 , wherein the alert identifies the one or more matching fault classifications that correspond to one or more of the fault classifications. 6. The method of claim 1 , wherein the audio corpus of different classifications comprises a database of audio files obtained from industrial equipment located in a plurality of different facilities. 7. A method for monitoring industrial equipment in a facility, comprising:receiving, from an audio sensor, an audio stream sensed by the audio sensor of an operation of industrial equipment in the facility; sending at least part of the audio stream to a machine learning module; the machine learning module comparing the received audio stream to an audio corpus to attempt to classify the audio stream into at least one of a plurality of different classifications including at least one fault classification; when the audio stream is classified by the machine learning module into at least one fault classification, provide an alert to a user; and when the audio stream cannot be classified by the machine learning module:receiving a classification of the audio stream; and adding at least one or more characteristics of the audio stream and the resulting classification to the audio corpus. 8. The method of claim 7 , wherein at least one of the plurality of different classifications corresponds to a non-fault classification. 9. The method of claim 7 , wherein the alert identifies the fault classification in the alert. 10. The method of claim 7 , wherein the audio corpus comprises a database of audio files obtained from industrial equipment located in a plurality of different facilities. 11. A method for monitoring industrial equipment in a facility, the method comprising:receiving, from an audio sensor, an audio stream sensed by the audio sensor of an operation of industrial equipment in the facility; sending at least part of the audio stream to a machine learning module, wherein the machine learning module accesses an audio corpus that includes a database of audio files obtained from industrial equipment, wherein each of the audio files is classified into one or more of a plurality of classifications; the machine learning module comparing at least part of the audio stream to the audio corpus to identify one or more matching classifications, if any; and when a matching classification is identified by the machine learning module, and when the matching classification is a fault classification, providing an alert to notify a user of a possible fault in the industrial equipment. 12. The method of claim 11 , wherein:when a matching classification is not identified by the machine learning module, receiving a classification of the audio stream, and then adding at least one or more characteristics of the audio stream and the resulting classification to the audio corpus. 13. The method of claim 11 , wherein the database of audio files includes auto files that are obtained from industrial equipment located in a plurality of different facilities, and wherein the machine learning module is hosted on a cloud based service. 14. The method of claim 13 , further comprising:processing the audio stream sensed by the audio sensor to determine an anomalous portion of the audio stream; and wherein the at least part of the audio stream that is sent to the machine learning module includes at least part of the audio stream that is determined to be anomalous, and wherein at least part of the audio stream that is not determined to be anomalous is not sent to the machine learning module. 15. The method of claim 14 , wherein the audio stream is processed to determine the anomalous portion of the audio stream via a controller located in the facility.
意見反饋
使用該功能遇到問(wèn)題
該功能需要專業(yè)版企業(yè)版VIP權(quán)限，您可以：
您也可以聯(lián)系官方QQ: 2157717237 電話: 13264338900