A linear fitting-based tip timing arrival time extraction algorithm

Abstract

The application relates to the field of engine detection, in particular to a blade tip timing arrival time extraction algorithm based on linear fitting, which comprises the following steps: collecting blade tip sensor signal data to perform discrete processing and obtaining a blade tip pulse signal sampling point sequence; and obtaining a blade tip timing arrival time by using the blade tip pulse signal sampling point sequence; the arrival time extraction algorithm based on linear fitting obtains a pulse rising edge time and a pulse falling edge time by searching the positions of the most rapid rising and the most rapid falling of the timing pulse, and takes a weighted average value of the two as the pulse arrival time, so that the influence of timing pulse amplitude and waveform change on timing accuracy can be effectively overcome or reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of engine detection, in particular to a linear fitting-based blade tip timing arrival time extraction algorithm. BACKGROUND

[0002] Large rotating machines such as aero-engines and gas turbines are widely used in the power systems of aviation and navigation, and are also key equipment widely used in the industrial sectors of electric power and petrochemical industry. The blade, as a core component of the rotor of a large rotating machine, is affected by various operating parameter changes such as blade vibration and blade tip clearance, which influences the normal operation and working efficiency of the entire system. In recent years, blade vibration and blade tip clearance measurement based on the blade tip timing principle have been widely researched and applied. The accuracy of blade tip arrival time extraction directly determines the precision of blade vibration or blade tip clearance measurement, and thus plays a crucial role in the entire measurement system. Currently, the threshold cutting method is usually used for arrival time extraction, which is greatly affected by the amplitude of the timing pulse and has limited precision. SUMMARY

[0003] In view of the deficiencies of the prior art, the present application provides a linear fitting-based blade tip timing arrival time extraction algorithm, which calculates the blade tip timing arrival time by collecting blade tip pulse data, thereby improving the calculation precision.

[0004] To achieve the above object, the present application provides a linear fitting-based blade tip timing arrival time extraction algorithm, comprising:

[0005] Collecting blade tip sensor signal data for discrete processing to obtain a blade tip pulse signal sampling point sequence;

[0006] Obtaining the blade tip timing arrival time by using the blade tip pulse signal sampling point sequence.

[0007] Preferably, the collecting blade tip sensor signal data for discrete processing to obtain a blade tip pulse signal sampling point sequence comprises:

[0008] Collecting blade tip sensor signal data for discrete processing to obtain blade tip sensor signal discrete data;

[0009] Establishing a blade tip pulse signal sampling point sequence by using the blade tip sensor signal discrete data;

[0010] The blade tip sensor signal discrete data is the blade tip sensor signal sampling time and the blade tip sensor signal sampling amplitude.

[0011] Preferably, the obtaining the blade tip timing arrival time by using the blade tip pulse signal sampling point sequence comprises:

[0012] The number of discrete sampling points of the tip pulse rising section and the number of discrete sampling points of the tip pulse falling section corresponding to the sequence of sampling points of the tip pulse signal are obtained;

[0013] The tip timing arrival time is obtained by using the number of discrete sampling points of the tip pulse rising section and the number of discrete sampling points of the tip pulse falling section.

[0014] Further, obtaining the tip timing arrival time by using the number of discrete sampling points of the tip pulse rising section and the number of discrete sampling points of the tip pulse falling section comprises:

[0015] The tip pulse state is obtained according to the number of discrete sampling points of the tip pulse rising section and the number of discrete sampling points of the tip pulse falling section;

[0016] It is judged whether the tip pulse state is an asymmetric state. If yes, the tip timing pulse rising edge time and the tip timing pulse falling edge time are obtained by using the sequence of sampling points of the tip pulse signal to perform linear fitting processing based on the least square method. Otherwise, the symmetric state fitting point value is obtained according to the number of discrete sampling points of the tip pulse rising section and the number of discrete sampling points of the tip pulse falling section.

[0017] The first tip timing arrival time is calculated by using the tip timing pulse rising edge time and the tip timing pulse falling edge time.

[0018] The second tip timing arrival time is obtained by using the sequence of sampling points of the tip pulse signal based on the symmetric state fitting point value.

[0019] The tip timing arrival time is obtained by using the first tip timing arrival time and the second tip timing arrival time.

[0020] Further, the tip timing pulse rising edge time and the tip timing pulse falling edge time obtained by using the sequence of sampling points of the tip pulse signal to perform linear fitting processing based on the least square method comprise:

[0021] The first slope sequence and the first intercept sequence are obtained by using the sequence of sampling points of the tip pulse signal to perform m-point linear fitting processing based on the least square method.

[0022] The tip timing pulse rising edge time is obtained by using the maximum slope in the first slope sequence and the intercept corresponding to the maximum slope in the first intercept sequence.

[0023] The second slope sequence and the second intercept sequence are obtained by using the sequence of sampling points of the tip pulse signal to perform n-point linear fitting processing based on the least square method.

[0024] The tip timing pulse falling edge time is obtained by using the maximum slope in the second slope sequence and the intercept corresponding to the maximum slope in the second intercept sequence.

[0025] Wherein, m is an integer ranging from 3 to the number of discrete sampling points in the rising section of the blade tip pulse, and n is an integer ranging from 3 to the number of discrete sampling points in the falling section of the blade tip pulse.

[0026] Furthermore, the calculation formula for calculating the first arrival moment of the blade tip timing using the rising edge moment of the blade tip timing pulse and the falling edge moment of the blade tip timing pulse is as follows:

[0027] T=bT1+dT2

[0028] Wherein, T is the first arrival moment of blade tip timing, b and d are constants, T1 is the rising edge moment of blade tip timing pulse, and T2 is the falling edge moment of blade tip timing pulse.

[0029] Furthermore, obtaining the second blade tip arrival timing based on the symmetric state fitting point value using the blade tip pulse signal sampling point sequence includes:

[0030] The symmetric state fitting error function is calculated using the blade tip pulse signal sampling point sequence;

[0031] Obtaining the symmetric state slope and the symmetric state intercept corresponding to the minimum value of the symmetric state fitting error function;

[0032] The symmetric state slope sequence k is established by using the symmetric state slope i ;

[0033] The symmetric state transverse intercept sequence T is established by using the symmetric state transverse intercept i ;

[0034] Get the symmetric state slope sequence k i The maximum value of the symmetric state slope is in the symmetric state intercept sequence T i The corresponding moment of the maximum slope of the intercept;

[0035] Get the symmetric state slope sequence k i The minimum value of the symmetric state slope is in the symmetric state intercept sequence T i The moment corresponding to the minimum slope and intercept in ;

[0036] The blade tip timing arrival second moment is calculated using the moment corresponding to the maximum slope transverse intercept and the moment corresponding to the minimum slope transverse intercept.

[0037] Furthermore, the calculation formula for calculating the symmetric state fitting error function using the blade tip pulse signal sampling point sequence is as follows:

[0038]

[0039] Among them, S i is the symmetric state fitting error function, k, Tk N is a symmetrical state fitting point value, y p t is a blade tip sensor signal sampling amplitude, t p t is a blade tip sensor signal sampling time.

[0040] Further, the symmetrical state fitting point value ranges from 3 to 0.5a, and a is an integer;

[0041] Wherein, a is the sum of the number of discrete sampling points of the blade tip pulse rising section and the number of discrete sampling points of the blade tip pulse falling section.

[0042] Further, the calculation formula of the blade tip timing reaching the second time is calculated by using the slope maximum value intercept corresponding time and the slope minimum value intercept corresponding time as follows:

[0043]

[0044] Wherein, T j is the blade tip timing reaching the second time, T x is the slope maximum value intercept corresponding time, T h is the slope minimum value intercept corresponding time.

[0045] Compared with the closest prior art, the present application has the beneficial effects:

[0046] The arrival time extraction algorithm based on linear fitting can effectively overcome or reduce the influence of the timing pulse amplitude and waveform change on the timing accuracy by searching the most rapid rising and falling positions of the timing pulse to obtain the rising and falling edge time and taking the weighted average value of the two as the pulse arrival time. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 is a flow chart of a blade tip timing arrival time extraction algorithm based on linear fitting provided by the present application;

[0048] Figure 2 is an example diagram of a fiber blade tip timing sensor collecting blade pulses of the blade tip timing arrival time extraction algorithm based on linear fitting provided by the present application;

[0049] Figure 3 is a linear fitting schematic diagram of the blade tip timing arrival time extraction algorithm based on linear fitting provided by the present application. DETAILED DESCRIPTION

[0050] The specific embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

[0051] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0052] Embodiment 1

[0053] The present application provides a linear fitting-based blade tip timing arrival time extraction algorithm, as shown in Figure 1 , comprising:

[0054] S1, collecting blade tip sensor signal data for discrete processing to obtain a blade tip pulse signal sampling point sequence;

[0055] S2, obtaining a blade tip timing arrival time using the blade tip pulse signal sampling point sequence.

[0056] S1 specifically comprises:

[0057] S1-1, collecting blade tip sensor signal data for discrete processing to obtain blade tip sensor signal discrete data;

[0058] S1-2, establishing a blade tip pulse signal sampling point sequence using the blade tip sensor signal discrete data;

[0059] wherein the blade tip sensor signal discrete data is a blade tip sensor signal sampling time and a blade tip sensor signal sampling amplitude.

[0060] S2 specifically comprises:

[0061] S2-1, obtaining a blade tip pulse rising section discrete sampling point quantity and a blade tip pulse falling section discrete sampling point quantity corresponding to the blade tip pulse signal sampling point sequence;

[0062] S2-2, obtaining a blade tip timing arrival time using the blade tip pulse rising section discrete sampling point quantity and the blade tip pulse falling section discrete sampling point quantity.

[0063] S2-2 specifically comprises:

[0064] S2-2-1, obtaining a blade tip pulse state according to the blade tip pulse rising section discrete sampling point quantity and the blade tip pulse falling section discrete sampling point quantity;

[0065] S2-2-2, judging whether the tip pulse state is asymmetric, if yes, linear fitting processing is performed on the tip pulse signal sampling point sequence based on the least square method to obtain the tip timing pulse rising edge time and the tip timing pulse falling edge time, otherwise, symmetric state fitting point values are obtained according to the number of discrete sampling points of the tip pulse rising segment and the number of discrete sampling points of the tip pulse falling segment;

[0066] S2-2-3, calculating the tip timing to reach the first time point by using the tip timing pulse rising edge time and the tip timing pulse falling edge time;

[0067] S2-2-4, obtaining the tip timing to reach the second time point by using the tip pulse signal sampling point sequence based on the symmetric state fitting point values;

[0068] S2-2-5, using the tip timing to reach the first time point and the tip timing to reach the second time point as the tip timing to reach time.

[0069] S2-2-2 specifically includes:

[0070] S2-2-2-1, performing m-point linear fitting processing on the tip pulse signal sampling point sequence based on the least square method to obtain a first slope sequence and a first intercept sequence;

[0071] S2-2-2-2, obtaining the intercept corresponding to the maximum slope in the first intercept sequence at the time corresponding to the maximum slope in the first intercept sequence as the tip timing pulse rising edge time;

[0072] S2-2-2-3, performing n-point linear fitting processing on the tip pulse signal sampling point sequence based on the least square method to obtain a second slope sequence and a second intercept sequence;

[0073] S2-2-2-4, obtaining the intercept corresponding to the maximum slope in the second intercept sequence at the time corresponding to the maximum slope in the second intercept sequence as the tip timing pulse falling edge time;

[0074] Wherein, m is an integer in the range of 3 to the number of discrete sampling points of the tip pulse rising segment, and n is an integer in the range of 3 to the number of discrete sampling points of the tip pulse falling segment.

[0075] The calculation formula of S2-2-3 is as follows:

[0076] T=bT1+dT2

[0077] Wherein, T is the tip timing to reach the first time point, b and d are constants, T1 is the tip timing pulse rising edge time, and T2 is the tip timing pulse falling edge time.

[0078] S2-2-4 specifically includes:

[0079] S2-2-4-1. Calculate the symmetric state fitting error function using the blade tip pulse signal sampling point sequence;

[0080] S2-2-4-2. Obtain the symmetric state slope and the symmetric state intercept corresponding to the minimum value of the symmetric state fitting error function;

[0081] S2-2-4-3. Use the symmetric state slope to establish a symmetric state slope sequence k i ;

[0082] S2-2-4-4. Use the symmetric state transverse intercept to establish a symmetric state transverse intercept sequence T. i ;

[0083] S2-2-4-5. Obtain the symmetric state slope sequence k i The maximum value of the symmetric state slope is in the symmetric state intercept sequence T i The corresponding moment of the maximum slope of the intercept;

[0084] S2-2-4-6, obtain the symmetric state slope sequence k i The minimum value of the symmetric state slope is in the symmetric state intercept sequence T i The moment corresponding to the minimum slope and intercept in ;

[0085] S2-2-4-7. Calculate the second arrival moment of the blade tip timing using the moment corresponding to the maximum slope transverse intercept and the moment corresponding to the minimum slope transverse intercept.

[0086] In this embodiment, a blade tip timing arrival time extraction algorithm based on linear fitting is firstly performed with an N-point least square linear fitting on the blade tip pulse signal sampling point sequence, that is, the k and T of each element of the blade tip pulse signal sampling point sequence are solved so that the calculation formula S2-2-4-1 has the minimum value. k , and we get a symmetric slope sequence and a symmetric intercept sequence.

[0087] The calculation formula for S2-2-4-1 is as follows:

[0088]

[0089] Among them, S i is the symmetric state fitting value, k, T k is the function parameter, N is the symmetric state fitting point value, y p is the blade tip sensor signal sampling amplitude, t p is the sampling moment of the blade tip sensor signal; the range of the symmetric state fitting point value is an integer between 3 and 0.5a;

[0090] Wherein, a is the sum of the number of discrete sampling points of the blade tip pulse rising section and the number of discrete sampling points of the blade tip pulse falling section.

[0091] The calculation formula of S2-2-4-6 is as follows:

[0092]

[0093] Wherein, T j is the second time when the blade tip timing arrives, T x is the time corresponding to the maximum slope intercept, T h is the time corresponding to the minimum slope intercept.

[0094] Embodiment 2:

[0095] The application provides a linear fitting-based blade tip timing arrival time actual application extraction algorithm, comprising:

[0096] When the blade sweeps through the blade tip timing sensor, the sensor output signal amplitude will experience a process from small to large and then from large to small, forming a blade tip pulse signal, as shown in Figure 2 .

[0097] Suppose that the minimum sensor output amplitude from the arrival of the a-1th blade to the arrival of the ath blade is Then the sensor signal between the occurrence time and the occurrence time is defined as the blade tip pulse signal of the ath blade. Figure 2 There are four pulse signals in the formula. Suppose that the maximum value of the ath blade tip pulse is Then the time period in which the signal amplitude y satisfies the following formula is defined as the pulse rising section, and the time period in which the signal amplitude y satisfies the following formula is defined as the pulse falling section.

[0098]

[0099]

[0100] It is worth noting that 0.2 in the formula is a preferred value obtained according to experience, and other values can also be taken according to the actual situation of the user.

[0101] In order to extract the pulse arrival time, first, the sensor signal is discretized using a sampling device. Suppose that the discretized blade tip pulse signal forms a sampling point sequence D i (i is a positive integer), the abscissa of D i is the sampling time t i , and the ordinate is the signal amplitude y i . Ntr is the number of sampling points of the blade tip pulse rising section after discretization defined by the first formula, and Ntf is the number of sampling points of the blade tip pulse falling section after discretization defined by the second formula.

[0102] For sensors with more symmetric tip-pulse (i.e. Ntr is close to Ntf), first perform N-point (N is an integer in the range of 3~(Ntr+Ntf) / 2) least square linear fitting on sequence D i , i.e. on points D i ~D i+N-1 , to get slope sequence k i and intercept sequence T i .

[0103]

[0104] Next, search the maximum and minimum of slope sequence k i , and get their indexes imax and imin, then T imax and T imin are taken as the rising edge time and falling edge time of the timing pulse respectively, while the blade arrival time T is equal to the average of the two, i.e. T = (T imax + T imin ) / 2.

[0105]

[0106] For sensors with asymmetric tip-pulse, first determine the rising edge time of the pulse. Perform Nr-point least square linear fitting on sequence D i , to get slope sequence k i ' and intercept sequence T i '. Search the maximum of slope sequence k i ', and get the maximum index imax, then T imax ' is taken as the rising edge time of the timing pulse. The preferred value of Nr is an integer in the range of 3~Ntr. Next determine the falling edge time of the pulse. Perform Nf-point least square linear fitting on sequence D i , to get slope sequence k i " and intercept sequence T i ". Search the minimum of slope sequence k i ", and get the minimum index imin, then T imin " is taken as the falling edge time of the timing pulse. The preferred value of Nf is an integer in the range of 3~Ntf. Finally, the blade arrival time T is the weighted average of T imax ' and T imin ", i.e. T = bT imax ' + dT imin ".

[0107] T = bT imax ' + dT imin "

[0108] Where b and d are values that make T imax ' < T < T imin " hold.

[0109] The implementation process of the algorithm is described below by taking N=5 as an example. As shown in FIG. 3, it is a symmetrical tip pulse signal, and the black dots are the tip pulse sampling point sequence D Figure 3 i A 5-point least square linear fitting is performed on D 19 23 The straight line l r obtained by fitting has a maximum slope, and its intercept T 19 is the pulse rising edge time; D 27 31 The straight line l f obtained by fitting has a minimum slope, and its intercept T 27 is the pulse falling edge time. The arrival time of the final tip timing is T=(T 19 +T 27 ) / 2.

[0110] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the above examples, those skilled in the art should understand that the specific embodiments of the present application can still be modified or replaced equivalently without departing from the spirit and scope of the present application, and any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered within the protection scope of the claims of the present application.​​​

Claims

1. A linear fitting based tip timing arrival time extraction algorithm, characterized in that, include: The blade tip sensor signal data is collected and discretely processed to obtain a blade tip pulse signal sampling point sequence; Obtaining the blade tip timing arrival time using the blade tip pulse signal sampling point sequence includes: Obtain the number of discrete sampling points in the blade tip pulse rising section and the number of discrete sampling points in the blade tip pulse falling section corresponding to the blade tip pulse signal sampling point sequence; Obtaining the blade tip timing arrival time by using the number of discrete sampling points in the blade tip pulse rising segment and the number of discrete sampling points in the blade tip pulse falling segment includes: Obtaining a tip pulse state according to the number of discrete sampling points in the rising section of the tip pulse and the number of discrete sampling points in the falling section of the tip pulse; Determine whether the blade tip pulse state is an asymmetric state. If so, use the blade tip pulse signal sampling point sequence to perform linear fitting processing based on the least squares method to obtain the blade tip timing pulse rising edge time and the blade tip timing pulse falling edge time. Otherwise, obtain the symmetric state fitting point value according to the number of discrete sampling points in the blade tip pulse rising segment and the number of discrete sampling points in the blade tip pulse falling segment; The blade tip timing arrival first moment is calculated using the blade tip timing pulse rising edge moment and the blade tip timing pulse falling edge moment; The blade tip timing arrival second moment is obtained based on the symmetrical state fitting point value using the blade tip pulse signal sampling point sequence; The blade tip timing arrival first moment and the blade tip timing arrival second moment are used as the blade tip timing arrival moments.

2. A linear fitting based tip timing arrival instant extraction algorithm as claimed in claim 1, wherein, The collecting of blade tip sensor signal data and performing discrete processing to obtain a blade tip pulse signal sampling point sequence comprises: Collecting blade tip sensor signal data and performing discrete processing to obtain blade tip sensor signal discrete data; Using the discrete data of the blade tip sensor signal to establish a blade tip pulse signal sampling point sequence; The discrete data of the blade tip sensor signal are the blade tip sensor signal sampling time and the blade tip sensor signal sampling amplitude.

3. A linear fitting based tip timing arrival instant extraction algorithm as claimed in claim 1, wherein, The method of performing linear fitting processing based on the least square method using the blade tip pulse signal sampling point sequence to obtain the blade tip timing pulse rising edge moment and the blade tip timing pulse falling edge moment comprises: The tip pulse signal sampling point sequence is used to perform m-point linear fitting based on the least square method to obtain the first slope sequence and the first intercept sequence; Obtaining the time corresponding to the maximum slope value in the first slope sequence and the cross-intercept value in the first cross-intercept sequence as the rising edge time of the blade tip timing pulse; The blade tip pulse signal sampling point sequence is used to perform n-point linear fitting based on the least square method to obtain the second slope sequence and the second intercept sequence; Obtaining a time corresponding to the maximum slope value in the second slope sequence and a time corresponding to the maximum slope value in the second intercept sequence as a falling edge time of the blade tip timing pulse; Wherein, m is an integer ranging from 3 to the number of discrete sampling points in the rising section of the blade tip pulse, and n is an integer ranging from 3 to the number of discrete sampling points in the falling section of the blade tip pulse.

4. A linear fitting based tip timing arrival instant extraction algorithm as claimed in claim 1, wherein, The calculation formula for calculating the first arrival moment of the blade tip timing using the rising edge moment of the blade tip timing pulse and the falling edge moment of the blade tip timing pulse is as follows: Wherein, T is the first time of tip timing, b, d are constants, T1 is the rising edge time of tip timing pulse, T2 is the falling edge time of tip timing pulse.

5. A linear fitting based tip timing arrival instant extraction algorithm as claimed in claim 1, wherein, The step of obtaining the second blade tip arrival timing based on the symmetric state fitting point value using the blade tip pulse signal sampling point sequence comprises: The symmetric state fitting error function is calculated using the blade tip pulse signal sampling point sequence; obtaining a symmetric state slope and a symmetric state intercept corresponding to a minimum value of the symmetric state fitting error function; The symmetric state slope sequence k is established by using the symmetric state slope i ; establishing a sequence T of symmetry state cross-sections using the symmetry state cross-section i ; acquiring a sequence of symmetry state slopes k i a maximum value of the symmetry state slopes in the sequence of symmetry state intercepts T i corresponding to the maximum value of the slopes acquiring a sequence of symmetry state slopes k i a minimum value of the symmetry state slopes in a sequence of symmetry state intercepts T i corresponding to the minimum value of the slopes calculating the second time when the blade tip timing arrives by using the time corresponding to the maximum value of the slope intercept and the time corresponding to the minimum value of the slope intercept.

6. A linear fitting based tip timing arrival instant extraction algorithm as claimed in claim 5, characterized in that, The calculation formula for calculating the symmetric state fitting error function by using the sequence of blade tip pulse signal sampling points is as follows: where S i is the symmetric state fitting error function, k, T k are function parameters, N is the symmetric state fitting point value, y p is the blade tip sensor signal sampling amplitude, t p is the blade tip sensor signal sampling time.

7. A linear fitting based tip timing arrival instant extraction algorithm as claimed in claim 6, characterized in that, The range of the symmetric state fitting point value is an integer within 3 to 0.5a; Wherein, a is the sum of the number of discrete sampling points of the blade tip pulse rising section and the number of discrete sampling points of the blade tip pulse falling section.

8. A linear fitting based tip timing arrival instant extraction algorithm as claimed in claim 5, wherein, The calculation formula for calculating the second time when the blade tip timing arrives by using the time corresponding to the maximum value of the slope intercept and the time corresponding to the minimum value of the slope intercept is as follows: Among them, T j The tip timing arrives at the second moment, T x is the time when the slope is maximum and the intercept is equal to the maximum value, T h is the moment when the slope intercept is minimum.

Images