System, apparatus and method for automatic pipe type detection
A fluid flow meter can include a sensor capable of transmitting a transmit signal to propagate, at least partially, through a fluid in a pipe and receiving a respective receive signal. The fluid flow meter can include a memory storing computer code instructions and a plurality of pipe type signatures associated with a plurality of pipe types. Each pipe type signature of a respective pipe type of the plurality of pipe types can include one or more characteristics of receive signals associated with that pipe type. The fluid flow meter can also include a processor communicatively coupled to the sensor and to the memory. When executing the computer code instructions, the processor can determine one or more signal features of the receive signal, and identify a pipe type of the pipe based on the one or more signal features of the receive signal and the plurality of pipe type signatures.
1. A fluid flow meter comprising:
an ultrasonic sensor including a first ultrasonic transducer capable of transmitting a transmit signal to propagate, at least partially, through a fluid in a pipe and a second ultrasonic transducer capable of receiving a respective receive signal;
a memory storing computer code instructions and a plurality of pipe type signatures associated with a plurality of pipe types, each pipe type signature of a respective pipe type of the plurality of pipe types including one or more characteristics of receive signals associated with that pipe type; and
a processor, communicatively coupled to the sensor and to the memory, configured to, when executing the computer code instructions:
compute, using the receive signal, a relative energy parameter indicative of a relative pipe signal energy of a pipe signal compared to a total energy of the receive signal, the pipe signal representing a first portion of the receive signal corresponding to a respective first portion of the transmit signal propagating substantially on, or within, a wall of the pipe between the first ultrasonic transducer and the second ultrasonic transducer, and the fluid signal representing a second portion of the receive signal corresponding to a respective second portion of the transmit signal propagating substantially through the fluid in the pipe between the first ultrasonic transducer and the second ultrasonic transducer;
determine a first time instance at which a first specific fraction of an energy of the pipe signal is received at the second ultrasonic transducer;
determine a second time instance at which a second specific fraction of an energy of fluid signal is received at the second ultrasonic transducer; and
identify, using the relative energy parameter, the first time instance, the second time instance, and a classifier defined based on the plurality of pipe type signatures, a pipe type of the pipe.
2. The fluid flow meter of claim 1 , wherein the processor is configured, when executing the computer code instructions, to compute wavelet coefficients of the receive signal or a signal portion of the receive signal.
3. The fluid flow meter of claim 1 , wherein the processor is further configured, when executing the computer code instructions, to select one or more samples of the receive signal.
4. The fluid flow meter of claim 1 , wherein the processor is further configured to, when executing the computer code instructions, to compute an envelope function of the receive signal or a portion of the receive signal.
5. The fluid flow meter of claim 1 , wherein the classifier includes a Gaussian mixture model (GMM) classifier.
6. The fluid flow meter of claim 5 , wherein identifying a pipe type of the pipe includes:
computing, for each Gaussian mixture distribution defined by the GMM classifier, a probability of the relative energy parameter, the first time instance, and the second time instance belonging to that Gaussian mixture distribution; and
selecting the pipe type corresponding to the Gaussian mixture distribution associated with the highest probability.
7. The fluid flow meter of claim 1 , wherein the processor is a processor of the fluid flow meter or a processor of a computing device communicatively coupled to the fluid flow meter.
8. The fluid flow meter of claim 1 , wherein the plurality of pipe type signatures include values, ranges, or regions associated with the relative energy parameter, the first time instance, and the second time instance.
9. The fluid flow meter of claim 8 , wherein identifying a pipe type of the pipe includes comparing the relative energy parameter, the first time instance, and the second time instance to respective values, ranges or regions in the plurality of pipe type signatures.
10. The fluid flow meter of claim 1 , wherein the processor is further configured, when executing the computer code instructions, to select or adjust a calibration parameter value of the fluid flow meter based on the identified pipe type.
11. A method of identifying a pipe type of a pipe associated with a fluid flow meter comprising:
transmitting, by a first ultrasonic transducer of an ultrasonic sensor of the fluid flow meter, a transmit signal to propagate, at least partially, through a fluid in the pipe;
receiving, by a second ultrasonic transducer of the ultrasonic sensor, a receive signal responsive to transmitting the transmit signal;
storing, by a memory, a plurality of pipe type signatures associated with a plurality of pipe types, each pipe type signature of a respective pipe type of the plurality of pipe types including one or more characteristics of receive signals associated with that pipe type;
computing, by the processor using the receive signal, a relative energy parameter indicative of a relative pipe signal energy of a pipe signal compared to a total energy of the receive signal, the pipe signal representing a first portion of the receive signal corresponding to a respective first portion of the transmit signal propagating substantially on, or within, a wall of the pipe between the first ultrasonic transducer and the second ultrasonic transducer, and the fluid signal representing a second portion of the receive signal corresponding to a respective second portion of the transmit signal propagating substantially through the fluid in the pipe between the first ultrasonic transducer and the second ultrasonic transducer;
determining, by the processor using the receive signal, a first time instance at which a first specific fraction of an energy of the pipe signal is received at the second ultrasonic transducer;
determining, by a processor using the receive signal, a second time instance at which a second specific fraction of an energy of fluid signal is received at the second ultrasonic transducer; and
identifying, by the processor, using the relative energy parameter, the first time instance, the second time instance, and a classifier defined based on the plurality of pipe type signatures, a pipe type of the pipe.
12. The method of claim 11 , further comprises computing wavelet coefficients of the receive signal or a signal portion of the receive signal.
13. The method of claim 11 , further comprises selecting one or more samples of the receive signal.
14. The method of claim 11 , further comprises computing an envelope function of the receive signal or a portion of the receive signal.
15. The method of claim 11 , wherein the classifier includes a Gaussian mixture model (GMM) classifier.
16. The method of claim 15 , wherein identifying a pipe type of the pipe includes:
computing, for each Gaussian mixture distribution defined by the GMM classifier, a probability of the relative energy parameter, the first time instance, and the second time instance belonging to that Gaussian mixture distribution; and
selecting the pipe type corresponding to the Gaussian mixture distribution associated with the highest probability.
17. The method of claim 11 , wherein the plurality of pipe type signatures include values, ranges, or regions associated with the relative energy parameter, the first time instance, and the second time instance.
18. The method of claim 17 , wherein identifying a pipe type of the pipe includes comparing the relative energy parameter, the first time instance, and the second time instance to respective values, ranges or regions in the plurality of pipe type signatures.
19. The method of claim 11 , further comprises selecting or adjusting a calibration parameter value of the fluid flow meter based on the identified pipe type.
20. A non-transitory computer-readable medium with computer code instructions stored thereon, the computer code instructions, when executed by a processor, cause the processor to:
cause a sensor of a fluid flow meter to transmit a transmit signal for propagating, at least partially, through a fluid in a pipe, and receive a receive signal responsive to transmitting the transmit signal;
compute, using the receive signal, a relative energy parameter indicative of a relative pipe signal energy of a pipe signal compared to a total energy of the receive signal, the pipe signal representing a first portion of the receive signal corresponding to a respective first portion of the transmit signal propagating substantially on, or within, a wall of the pipe between the first ultrasonic transducer and the second ultrasonic transducer, and the fluid signal representing a second portion of the receive signal corresponding to a respective second portion of the transmit signal propagating substantially through the fluid in the pipe between the first ultrasonic transducer and the second ultrasonic transducer;
determine a first time instance at which a first specific fraction of an energy of the pipe signal is received at the second ultrasonic transducer;
determine a second time instance at which a second specific fraction of an energy of fluid signal is received at the second ultrasonic transducer; and
identify, using the relative energy parameter, the first time instance, the second time instance, and a classifier defined based on a plurality of pipe type signatures, a pipe type of the pipe.