Method and apparatus for online identification of safe operation and advance detection of unsafe operation of a system or process
First Claim
1. A method for an online identification of safe operation and advance detection of unsafe operation of a system or a process in presence of noise in sensor measurements or fluctuations in variables measured, said method consisting the steps of:
- (a) obtaining an online signal from one or more sensors monitoring/measuring process states or operation of the system at specified time intervals, wherein said online signal being in the form of a time series data relating to variation in one or more process variables;
(b) digitizing said online signal of step (a) to obtain a digitized data set;
(c) differentiating the digitized data set of step (b) to obtain a first derivative data set;
(d) taking wavelet transformation of the first derivative data set of step (c) to obtain scalogram in terms of wavelet coefficients;
(e) computing power distribution for individual wavelet scale of step (d) from said wavelet coefficients at all wavelet scales;
(f) computing from the power distribution of step (e) a critical wavelet scale at which finer scales are separable with respect to coarser scales, wherein said finer scales have wavelet coefficients attributable to noise and said coarser scales have wavelet coefficients attributable to true first derivative data (from step (d));
(g) setting the value of the wavelet scales as zero, when the wavelet coefficients corresponding to the finer scales are less than or equal to the critical wavelet scale of step (f);
or retaining the value of the wavelet scales, when the wavelet coefficients corresponding to the coarser scales are greater than the critical wavelet scale;
(h) computing inverse wavelet transform of the wavelet coefficients of step (g) at different scales to get de-noised first derivative data;
(i) repeating the steps (d) through (h) on de-noised first derivative data to obtain de-noised second derivative data;
(j) testing criteria for unsafe operation based on simultaneous positivity in time of the de-noised first derivative data and de-noised second derivative data, subject to the condition that the power in the coarser scales for the de-noised second derivative data is finite;
(k) detecting unsafe process operation and initiating necessary corrective measures when the criteria in step (j) are satisfied, and (l) identifying safe process operation and repeating steps (a) to (j) when the criteria in step (j) are not satisfied.
3 Assignments
0 Petitions
Accused Products
Abstract
A method and apparatus for online identification of safe operation and advance detection of unsafe operation of systems or processes from signals containing noise wherein the method and apparatus uses wavelet transformations to accurately infer from criteria about the changes in the monitored process variables and its acceleration as a basis for advance detection of unsafe operation, the said process of online identification of safe operation and advance detection of unsafe operation has applications in a wide variety of situations where digitized data could be made available.
-
Citations
28 Claims
-
1. A method for an online identification of safe operation and advance detection of unsafe operation of a system or a process in presence of noise in sensor measurements or fluctuations in variables measured, said method consisting the steps of:
-
(a) obtaining an online signal from one or more sensors monitoring/measuring process states or operation of the system at specified time intervals, wherein said online signal being in the form of a time series data relating to variation in one or more process variables;
(b) digitizing said online signal of step (a) to obtain a digitized data set;
(c) differentiating the digitized data set of step (b) to obtain a first derivative data set;
(d) taking wavelet transformation of the first derivative data set of step (c) to obtain scalogram in terms of wavelet coefficients;
(e) computing power distribution for individual wavelet scale of step (d) from said wavelet coefficients at all wavelet scales;
(f) computing from the power distribution of step (e) a critical wavelet scale at which finer scales are separable with respect to coarser scales, wherein said finer scales have wavelet coefficients attributable to noise and said coarser scales have wavelet coefficients attributable to true first derivative data (from step (d));
(g) setting the value of the wavelet scales as zero, when the wavelet coefficients corresponding to the finer scales are less than or equal to the critical wavelet scale of step (f);
or retaining the value of the wavelet scales, when the wavelet coefficients corresponding to the coarser scales are greater than the critical wavelet scale;
(h) computing inverse wavelet transform of the wavelet coefficients of step (g) at different scales to get de-noised first derivative data;
(i) repeating the steps (d) through (h) on de-noised first derivative data to obtain de-noised second derivative data;
(j) testing criteria for unsafe operation based on simultaneous positivity in time of the de-noised first derivative data and de-noised second derivative data, subject to the condition that the power in the coarser scales for the de-noised second derivative data is finite;
(k) detecting unsafe process operation and initiating necessary corrective measures when the criteria in step (j) are satisfied, and (l) identifying safe process operation and repeating steps (a) to (j) when the criteria in step (j) are not satisfied. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. An apparatus for online identification of safe operation and advance detection of unsafe operation of a system or a process in presence of noise in sensor measurements or fluctuations in variables measured, said apparatus consisting:
-
(a) one or more sensors monitoring/measuring process states or operation of the system at specified time intervals for providing an online signal, wherein said online signal being in the form of a time series data relating to variation in one or more process variables;
(b) a digitizing means connected to the sensors for digitizing said online signal of step (a);
(c) a differentiating means coupled to said digitizing means for differentiating the digitized data set of and obtain a first derivative data set;
(d) a first computing means programmed to receive the first derivative data set and compute wavelet transform of the first derivative data set to obtain scalogram in terms of wavelet coefficients;
(e) a second computing means programmed to compute power distribution for individual wavelet scale of step (d) from said wavelet coefficients at all wavelet scales;
(f) a third computing means configured to receive the power distribution from the second computing means and programmed to compute a critical wavelet scale at which finer scales are separable with respect to coarser scales, wherein said finer scales have wavelet coefficients attributable to noise and said coarser scales have wavelet coefficients attributable to true first derivative data;
(g) a means for setting the value of the wavelet scales as zero, when the wavelet coefficients corresponding to the finer scales are less than or equal to the critical wavelet scale of step (f);
or retaining the value of the wavelet scales, when the wavelet coefficients corresponding to the coarser scales are greater than the critical wavelet scale;
(h) a means for computing inverse wavelet transform of the wavelet coefficients of step-(g) at different scales to get de-noised first derivative data;
(i) a testing means for testing criteria for unsafe operation based on simultaneous positivity in time of the de-noised first derivative data and de-noised second derivative data, subject to the condition that the power in the coarser scales for the de-noised second derivative data is finite;
(j) an identifying means for identifying the operation as unsafe operation/safe operation, said identifying means being connected to the testing means at its input end and being connected to an alarm/error correcting system at its out put end. - View Dependent Claims (17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28)
-
Specification