Monitoring system
According to one embodiment, a monitoring system includes a processor. The processor accepts first data output from a first detector. The first detector detects a signal caused by equipment. The processor performs a first determination when a first value is in a first state. The first value indicates a state of the first detector or an environment where the equipment is provided. The first determination determines a condition of the equipment by using a first model and the first data. The processor performs a second determination when the first value is in a second state different from the first state. The second determination determines the condition of the equipment by using a second model and the first data. The second model is different from the first model.
1. A monitoring system comprising a processor accepting first data output from a first detector detecting a signal caused by equipment,
the processor performing a first determination when a first value is in a first state, the first value indicating a state of the first detector, the first determination determining a condition of the equipment by using a first model and the first data, the first model including a neural network, and the first model being trained using a first detection data acquired by the first detector in a state where the condition of the equipment is good,
the processor performing a second determination when the first value is in a second state different from the first state, the second determination determining the condition of the equipment by using a second model and the first data, the second model being different from the first model and including another neural network.
2. The system according to claim 1 , wherein
when the first value is in the first state, the processor outputs first information indicating the first model, and
when the first value is in the second state, the processor outputs second information indicating the second model.
3. The system according to claim 1 , wherein
when the first value is in the first state, the processor outputs a first evaluation value indicating the condition of the equipment calculated using the first data and the first model, and
when the first value is in the second state, the processor outputs a second evaluation value indicating the condition of the equipment calculated using the first data and the second model.
4. The system according to claim 1 , wherein when the first value is in the second state, the processor generates, before the second determination, the second model by using third data output from the first detector.
5. The system according to claim 1 , wherein
when the first value is in the first state, the processor outputting a first determination result indicating the condition of the equipment determined using the first model and the first data,
when the first value is in the second state, the processor outputting a second determination result indicating the condition of the equipment determined using the second model and the first data.
6. The system according to claim 1 , wherein the first detector is a microphone or a vibration sensor.
7. The system according to claim 1 , wherein the first value is detected by a 3-axis acceleration sensor.
8. The system according to claim 1 , wherein
the first state corresponds to a first posture of an attached state of the first detector, and
the second state corresponds to a second posture of the attached state to which the attached state of the first detector has changed from the first posture.
9. The system according to claim 1 , wherein the neural network is a recurrent neural network.
10. The system according to claim 1 , wherein the first model is an autoencoder.
11. The system according to claim 1 , further comprising a second detector, wherein
the second detector is mounted to the first detector, and
the second detector outputs the first value.
12. The system according to claim 2 , wherein the processor outputs the first value and one of the first information or the second information.
13. The system according to claim 3 , wherein
when the first value is in the first state, the processor determines, in the first determination, the condition of the equipment based on the first evaluation value, and
when the first value is in the second state, the processor determines, in the second determination, the condition of the equipment based on the second evaluation value.
14. The system according to claim 3 , wherein when the first value is in the second state, the processor refers to a reference evaluation value and corrects the second evaluation value based on the reference evaluation value, the reference evaluation value being calculated using second data and the first model, the second data being output from the first detector before the second determination.
15. The system according to claim 9 , wherein the recurrent neural network has a Long Short Term Memory structure.
16. The system according to claim 11 , wherein
the second detector includes a 3-axis acceleration sensor, and
the 3-axis acceleration sensor outputs the first value.
17. The system according to claim 12 , wherein the processor further outputs the first data.