| CPC G07C 5/0808 (2013.01) [G01M 17/013 (2013.01); G07C 5/0833 (2013.01)] | 18 Claims |
|
1. An abnormality detection apparatus for detecting an abnormality that occurs in a wheel assembly, the abnormality detection apparatus comprising:
a sensing system including:
rotational bodies each rotating along with a corresponding one of wheel assemblies; and
sensors each continuously detecting a physical amount changed by a corresponding one of the rotational bodies;
a computer including an I/O interface and a CPU, each sensor being connected to the CPU via a communication line and the I/O interface, the physical amounts detected by each sensor being transmitted to the computer in real time, the CPU of the computer being programmed to function as:
a signal obtaining unit configured to sequentially obtain signals from the sensing system indicating a rotation speed of the wheel assembly, as pulses having a rise;
a first index calculation unit configured to calculate a first index indicating a temporal variation of the rise for each of the pulses corresponding to one rotation of the wheel assembly;
a spectrum calculation unit configured to calculate a frequency spectrum of rotational orders from a first order to an mth order (m≥1) of the first index by performing frequency analysis on the first index calculated for each of the pulses;
a standardization unit configured to standardize a gain for each of the rotational orders of the frequency spectrum, using a mean value and a standard deviation of the gain for each rotational order when there is no abnormality; and
a second index calculation unit configured to calculate a second index for determining whether there is an abnormality, based on the standardized gain; and
a display device, configured to output an alarm in response to a determination that there is abnormality, based on the standardized gain, wherein the computer is configured to make the determination with increased sensitivity to changes in a frequency spectrum and using an alarm threshold automatically derived from a statistical distribution based on a pre-defined significance level, without setting empirical thresholds for each rotational order.
|
|
17. A non-transitory computer-readable medium recording an abnormality detection program for detecting an abnormality that occurs in a wheel assembly attached to an axle, the program causing a computer to execute:
controlling a sensing system, the sensing system including:
rotational bodies each rotating along with a corresponding one of wheel assemblies; and
sensors each continuously detecting a physical amount changed by a corresponding one of the rotational bodies, wherein the computer includes an I/O interface and a CPU, each sensor being connected to the CPU via a communication line and the I/O interface, the physical amounts detected by each sensor being transmitted to the computer in real time;
sequentially obtaining signals from the sensing system indicating a rotation speed of the wheel assembly, as pulses having a rise;
calculating a first index indicating a temporal variation of the rise for each of the pulses corresponding to one rotation of the wheel assembly;
calculating a frequency spectrum of rotational orders from a first order to an mth order (m≥1) of the first index by performing frequency analysis on the first index calculated for each of the pulses;
standardizing a gain for each of the rotational orders of the frequency spectrum, using a mean value and a standard deviation of the gain for each rotational order when there is no abnormality;
calculating a second index for determining whether there is an abnormality, based on the standardized gain; and
controlling a display device to output an alarm in response to a determination that there is abnormality, based on the standardized gain, wherein the computer is configured to make the determination with increased sensitivity to changes in a frequency spectrum and using an alarm threshold automatically derived from a statistical distribution based on a pre-defined significance level, without setting empirical thresholds for each rotational order.
|
|
18. An abnormality detection method, executed by a computer for detecting an abnormality that occurs in a wheel assembly attached to an axle, the method comprising:
operating a sensing system, the sensing system including:
rotational bodies each rotating along with a corresponding one of wheel assemblies; and
sensors each continuously detecting a physical amount changed by a corresponding one of the rotational bodies, wherein the computer includes an I/O interface and a CPU, each sensor being connected to the CPU via a communication line and the I/O interface, the physical amounts detected by each sensor being transmitted to the computer in real time;
sequentially obtaining signals from the sensing system indicating a rotation speed of the wheel assembly, as pulses having a rise;
calculating a first index indicating a temporal variation of the rise for each of the pulses corresponding to one rotation of the wheel assembly;
calculating a frequency spectrum of rotational orders from a first order to an mth order (m≥1) of the first index by performing frequency analysis on the first index calculated for each of the pulses;
standardizing a gain for each of the rotational orders of the frequency spectrum, using a mean value and a standard deviation of the gain for each rotational order when there is no abnormality;
calculating a second index for determining whether there is an abnormality, based on the standardized gain; and
controlling a display device to output an alarm in response to a determination that there is abnormality, based on the standardized gain, wherein the computer is configured to make the determination with increased sensitivity to changes in a frequency spectrum and using an alarm threshold automatically derived from a statistical distribution based on a pre-defined significance level, without setting empirical thresholds for each rotational order.
|