Gas path particle concentration measuring method based on characteristic waveform recognition

Abstract

The invention relates to a gas path particle concentration measurement method based on characteristic waveform recognition, and an adopted measurement system comprises an electrostatic sensor, a signal preprocessing module, a signal acquisition module and a particle concentration calculation module. The gas path particle concentration measurement method comprises the following steps: acquiring an electrostatic sensor signal subjected to preliminary pretreatment; performing cross-correlation matching to extract an effective waveform position; extracting characteristic values: taking absolute values of peak values of waveform signals meeting similar conditions in the original acquisition signals, and summing the absolute values to serve as the characteristic values; establishing a characteristic value and particle concentration corresponding relation model; and particle concentration monitoring.

Description

Technical Field

[0001] The invention relates to the technical field of signal processing, and in particular to a real-time measurement method for gas path particle concentration based on characteristic waveform recognition. Background Art

[0002] In the gas system, the state of the gas inside the pipeline is a key concern in scientific research and engineering applications. In particular, the concentration of particulate matter in the gas in the pipeline is crucial to evaluate the operating efficiency of the gas system and the health of the equipment connected before and after it. For example, in the intake and exhaust ducts of aircraft engines, by monitoring the concentration of particulate matter, the external environmental conditions can be effectively evaluated, and measures can be taken in advance to deal with the harsh environmental factors that may pose a threat to the engine. In addition, this monitoring can also help detect problems such as engine blade damage caused by particulate wear to ensure that engine performance is not affected. In the engine design and development stage, precise control of the particle concentration in the experimental environment can significantly improve the R&D efficiency and engine performance. Similarly, in the coal powder transportation pipeline, the monitoring of particle concentration has important guiding value for optimizing the pipeline transportation capacity and improving the coal combustion efficiency. Therefore, the development and improvement of particle identification technology in gas pipelines has far-reaching practical significance for all fields involving the measurement of particle concentration in gas transportation systems.

[0003] The real-time measurement method of gas path particle concentration based on characteristic waveform recognition uses the characteristics of particles carrying charges due to friction with the gas path pipeline to achieve gas path particle concentration measurement through charge sensing and characteristic waveform recognition. It has the advantages of non-contact, online measurement, high sensitivity and low cost. At present, related research at home and abroad mainly focuses on the monitoring of exhaust particles at the tail nozzle of aircraft engines. The realized technology is mainly aimed at fault diagnosis of aircraft engines and other equipment according to the different charges carried by abnormal particles, and lacks a measurement method for particle concentration. In addition, most of the existing studies use the root mean square value of the extracted acquisition signal as a characteristic parameter to establish a connection with the particle concentration in the pipeline. The root mean square value is usually used to characterize the dispersion or fluctuation amplitude of the data set, and in the particle concentration measurement, it reflects the chaos level of the signal in the pipeline. The principle is that when charged particles pass through the sensor section, the waveform will cause a sudden change. However, the actual collected data show that in addition to the waveform changes caused by charged particles, there are also certain noise signals, and these noises are difficult to completely eliminate through conventional filtering methods. This results in the noise signal being included in the calculation of the RMS value, making the reflected signal chaos in the pipeline higher than the actual situation. Summary of the invention

[0004] The purpose of the present invention is to provide a method for measuring the concentration of particles in a gas pipeline based on characteristic waveform recognition, which can realize real-time monitoring. The method fully considers the characteristics of the sensor sensing signal being a Gaussian waveform when the charged gas pipeline particles in the gas pipeline pass through the sensor cross section, and uses the waveform cross-correlation method to accurately identify the characteristic waveform and establish a concentration correlation curve, thereby realizing the particle concentration monitoring in the gas pipeline.

[0005] To achieve the above object, the present invention is implemented by the following technical solutions:

[0006] A method for measuring gas path particle concentration based on characteristic waveform recognition, the measurement system used includes an electrostatic sensor, a signal preprocessing module, a signal acquisition module, and a particle concentration calculation module; the method for measuring gas path particle concentration includes the following steps:

[0007] S1: obtaining a preliminary preprocessed electrostatic sensor signal;

[0008] S2: Extract the effective waveform position by cross-correlation matching. The method is:

[0009] The electrostatic sensing signal after preliminary preprocessing is collected and transmitted to the particle concentration calculation module for processing;

[0010] The formula for determining the contrast waveform used for cross-correlation comparison is as follows:

[0011]

[0012] Among them, parameters α, μ, and σ are unknown coefficients, and the parameter α is selected in the range of 0.2-1.0; parameters μ and σ jointly determine the shape characteristics of the waveform, and parameter design needs to consider the sampling rate and the gas flow rate in the gas pipeline;

[0013] For a certain gas pipeline structure, the values ​​of parameters α, μ, and σ are determined through experiments, and the comparison waveforms are obtained;

[0014] The absolute value of the collected original electrostatic sensor waveform signal is taken and then cross-correlated with the comparison waveform to obtain the waveform signal that meets the similarity condition in the original electrostatic sensor waveform;

[0015] S3: Feature value extraction: take the absolute value of the peak value of the waveform signal that meets the similarity condition in the original acquisition signal and sum it as the feature value;

[0016] S4: Establishing a model for the relationship between characteristic values ​​and particle concentration;

[0017] S5: Particle concentration monitoring.

[0018] Furthermore, the electrostatic sensor is embedded inside the gas pipeline to obtain the electrostatic induction signal generated when the charged circuit particles pass through; the signal preprocessing module is used to receive the original sampled electrostatic sensor signal and implement preprocessing; the signal acquisition module is used to implement high-frequency sampling of the preprocessed electrostatic sensor signal; the particle concentration calculation module is used to complete the cross-correlation comparison between the high-frequency sampling signal and the characteristic waveform and the subsequent particle concentration calculation work.

[0019] Furthermore, before step S1, the definition and calculation method of the gas path particle concentration in the gas path pipeline are clarified:

[0020] In the gas pipeline, the gas path particle concentration in the pipeline is defined as the ratio of the mass of gas path particles passing through the sensor cross section to the volume of air passing through in a certain period of time. The volume of air passing through per unit time is the product of the gas flow rate in the pipeline and the time, that is,

[0021]

[0022] For an annular circulation gas pipeline, assuming that the flow velocity of the gas path particles in the pipeline is consistent with the wind speed in the pipeline and is evenly distributed in the pipeline, the time t for a single particle to circulate once is 0 is the total volume V in the pipe 0 The ratio of the fan flow rate Q is

[0023]

[0024] In a certain period of time t, the total weight of gas path particles passing through the sensor cross section is

[0025]

[0026] Where m is the mass of particles added to the gas circulation pipeline; the total volume of air passing through the sensor cross section within a certain time t is

[0027] V=Q·t

[0028] Within a certain time t, the concentration of gas path particles passing through the sensor cross section is

[0029]

[0030] Furthermore, the method of step S1 is: for the charge signal obtained by the electrostatic sensor, the signal preprocessing module is processed by charge amplification, voltage amplification, high-pass filtering and low-pass filtering circuits in sequence; the high-pass filtering and low-pass filtering modules are used to filter out power frequency interference and high-frequency noise respectively, so as to obtain an electrostatic sensing signal that has been preliminarily preprocessed.

[0031] Furthermore, in step S2, the method of cross-correlation processing is as follows: by cross-correlation calculation, the comparison waveform is scanned and compared with each position of the original waveform, the similarity between each position in the original waveform and the comparison waveform is quantified, and the waveform after cross-correlation processing is obtained; by setting a threshold value, the peak waveform signal that meets the similarity condition is extracted to obtain the waveform signal that meets the similarity condition in the original electrostatic sensor waveform.

[0032] Furthermore, S4: a method for establishing a corresponding relationship model between characteristic values ​​and particle concentrations is to change the particle concentration in the pipeline to conduct experiments, obtain corresponding characteristic value results under different particle concentrations, and use a linear fitting method to establish a corresponding relationship model between characteristic values ​​and particle concentrations, and complete system calibration.

[0033] Aiming at the requirement of real-time monitoring of particle concentration, the present invention proposes an online testing method for gas path particles based on gas path pipeline particle identification. This method has the following advantages:

[0034] (1) A method for calculating and analyzing the particle concentration in a circular circulation pipeline is proposed, which solves the problem of difficulty in calibrating and controlling the particle concentration during the experiment.

[0035] (2) The induced charge on the electrostatic sensor and the charge of the particles in the gas pipeline present a unique single-valued correspondence. On this basis, a method for preprocessing the electrostatic sensor signal is proposed to provide support for subsequent signal matching and characteristic value calibration.

[0036] (3) By cross-correlating the collected signal with the Gaussian-like signal to identify the particle characteristic waveform and further establish a characteristic relationship, it is more scientific and further improves the measurement sensitivity, better meeting the real-time monitoring performance of particle concentration required by the gas pipeline.

[0037] (4) The method of extracting the useful signal characteristics of the original waveform after gas pipeline waveform recognition and linearly calibrating it with the particle concentration in the pipeline can further improve the resolution and sensitivity of particle online monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a structural block diagram of the gas path particle online testing system based on gas path pipeline particle identification of the present invention.

[0039] Figure 2 The flow chart of the on-line testing method of gas path particles based on gas path pipeline particle identification of the present invention is as follows:

[0040] Figure 3 : is a section of collected signal graph, in which the dotted line is the signal curve when no particles are injected, and the solid line is the signal curve after the particles are injected.

[0041] Figure 4 A typical Gaussian-like curve for a single charge used in cross-correlation calculations.

[0042] Figure 5 For the general Figure 2 The result after taking the absolute value of the collected signal and performing cross-correlation algorithm processing.

[0043] Figure 6 This is the algorithm flow chart of the particle concentration calculation module. DETAILED DESCRIPTION

[0044] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0045] To achieve the above object, the present invention is implemented by the following technical solutions:

[0046] like Figure 1 As shown, the real-time measurement system for gas path particle concentration designed by the present invention includes an electrostatic sensor, a signal preprocessing module, a signal acquisition module, and a particle concentration calculation module.

[0047] The electrostatic sensor is embedded inside the gas pipeline to obtain the electrostatic induction signal generated when charged particles pass through it; the signal preprocessing module is used to receive the original sampled electrostatic sensor signal and realize the conversion from charge signal to voltage signal as well as amplification and noise reduction preprocessing; the signal acquisition module is used to realize high-frequency sampling of the electrostatic sensor signal after preprocessing; the particle concentration calculation module is used to complete the cross-correlation comparison between the collected signal and the characteristic waveform and the subsequent particle concentration calculation.

[0048] The real-time measurement method of gas path particle concentration mainly includes the following steps:

[0049] S1: Clarify the definition and calculation method of gas path particle concentration in gas path pipelines

[0050] In the gas pipeline, the gas path particle concentration in the pipeline is defined as the ratio of the mass of gas path particles passing through the sensor cross section to the volume of air passing through in a certain period of time. The volume of air passing through per unit time is the product of the gas flow rate in the pipeline and the time, that is,

[0051]

[0052] For annular circulating gas pipelines, the same gas path particle will pass through the sensor cross section multiple times per unit time, which is different from the single passing of gas path particles in ordinary gas pipelines, and the method of calculating concentration also changes accordingly. Assuming that the flow speed of gas path particles in the pipeline is consistent with the wind speed in the pipeline and is evenly distributed in the pipeline, the time t for a single particle to circulate once is 0 is the total volume V in the pipe 0The ratio of the fan flow rate Q is

[0053]

[0054] Then within a certain time t, the total weight of the gas path particles passing through the sensor cross section is

[0055]

[0056] Where m is the mass of particles added to the air circulation pipeline. The total volume of air passing through the sensor cross section within a certain time t is

[0057] V=Q·t

[0058] Therefore, within a certain time t, the concentration of gas path particles passing through the sensor cross section is

[0059]

[0060] S2: Preliminary preprocessing of electrostatic sensor signals

[0061] Electrostatic sensors are used to capture electrostatic induction signals generated by charged particles inside gas pipelines. Based on the principle of electrostatic induction, when gas pipeline particles are charged by friction or contact, these charged particles will generate a constantly changing electrostatic field around the electrostatic sensor, causing the induced charge on the sensor probe to fluctuate. According to the uniqueness theorem of the electrostatic field, it can be obtained that when the boundary conditions of the pipeline, insulator and sensor electrodes and the electrostatic spatial charge distribution of the gas pipeline system are determined, the induced charge on the sensor metal electrode and the charge carried by the particles present a unique corresponding relationship. This relationship provides a theoretical basis for quantifying the charge carried by particles in the gas pipeline by monitoring the output signal of the electrostatic sensor.

[0062] For the charge signal collected by the electrostatic sensor, the signal preprocessing module uses charge amplification, voltage amplification, high-pass filtering and low-pass filtering circuits arranged in sequence for processing. First, the original charge signal is converted and amplified into a voltage signal for subsequent processing by the charge amplification circuit. Subsequently, the voltage amplification module is used to further amplify the above voltage signal to enhance the signal strength. The high-pass filtering and low-pass filtering modules are used to filter out power frequency interference and high-frequency noise, respectively, so as to obtain the electrostatic sensor signal after preliminary preprocessing.

[0063] S3: Cross-correlation matching to extract valid waveform positions

[0064] The signal acquisition module is used to collect the electrostatic sensor signal processed by the signal preprocessing module and transmit it to the particle concentration calculation module to analyze the collected signal waveform. Figure 3 : is a section of collected signal graph, in which the dotted line is the signal curve when no particles are injected, and the solid line is the signal curve after the particles are injected.

[0065] The results show that compared with the signal without particle injection, the signal generated by charged particles presents a significant Gaussian distribution characteristic. Therefore, the useful signal generated by charged particles can be extracted from the signal through feature recognition method and then analyzed.

[0066] After analyzing the signal collection results of repeated experiments, combined with the waveform information such as the amplitude and Gaussian distribution characteristics of the induced signal generated by the charged particles, the waveform formula for cross-correlation comparison was finally determined as follows:

[0067]

[0068] Combined with the experimental results, in this experimental environment, the parameters are finally determined to be α = 0.3, μ = 0.4, and σ = 1.5. The parameter α determines the amplitude of the contrast waveform. After the system circuit is amplified, the parameter selection range is between 0.2-1.0 when applied to different working conditions. The parameters μ and σ jointly determine the shape characteristics of the waveform. The parameter design needs to refer to the sampling rate of the system and the gas flow rate in the gas pipeline, depending on the specific situation.

[0069] The waveform used for cross-correlation comparison in this experiment is as follows Figure 4 As shown in the figure, the collected signal is cross-correlated with the waveform curve, and the comparison waveform is scanned and compared with each position of the original waveform through cross-correlation calculation, and the similarity between each position in the original waveform and the waveform used for cross-correlation comparison is quantified to obtain the waveform after cross-correlation processing. Therefore, for the result after cross-correlation processing, each maximum value can reflect the similarity between each spike signal in the original data and the comparison waveform, and then by setting the threshold method, the spike waveform signal that meets the higher similarity condition can be extracted. Figure 5 For the general Figure 3 After taking the absolute value of the collected signal, Figure 4 The result after the cross-correlation algorithm is used to process the contrast signal. After setting the threshold to 0.2, the number of maximum values ​​that can be obtained is 6, which corresponds to Figure 4 The number of spike waveforms that meet similar conditions in the original acquired signal is 6.

[0070] S4: Feature extraction

[0071] After completing the cross-correlation matching and extracting effective waveforms in step S3, the waveform signal that meets the similarity condition in the original acquisition signal is obtained. At the same time, according to the theoretical analysis of the preprocessing part of the annular electrostatic sensor signal, the amplitude of the electrostatic induction signal waveform is proportional to the charge of the gas path particles passing through the pipeline, and can also be considered to be proportional to the particle concentration in the pipeline. Therefore, the peak value of the waveform signal that meets the similarity condition in the original acquisition signal is extracted, and the absolute value is summed as the characteristic value. For example, Figure 3After cross-correlation comparison, there are 6 peak signals that meet the similarity condition. The peak signals of these 6 peaks are extracted, and their absolute values ​​are taken and summed up as the characteristic value of this section of the signal. The characteristic value can be used as a basis for judging the total amount of charged particles in the air path passing through the cross-section of the electrostatic sensor within a period of time, and can also reflect the particle concentration passing through the electrostatic sensor during this period of time.

[0072] S5: Establishing a model for the relationship between eigenvalues ​​and particle concentration

[0073] On the basis of step S4, the particle concentration in the pipeline is continuously changed to obtain the corresponding characteristic value results under different particle concentrations, and a linear fitting method is used to establish the peak sum as a corresponding relationship model between the characteristic value and the particle concentration.

[0074] S6: Calibration result inspection

[0075] Based on the corresponding relationship model between the characteristic value and the particle concentration established in step S5, the calibration results are tested. Without changing the external environment, within the particle measurement range, the particle concentrations of another M values ​​(generally M ≥ 5) are detected by controlling the mass of the injected particles, and the measured values ​​and actual values ​​of the M particle concentrations are obtained. The error between the measured value and the actual value is compared. If the error meets the requirements, proceed to the next step. Otherwise, return to step S5 and re-calibrate.

[0076] S7: Complete system calibration and realize real-time monitoring of particle concentration.

[0077] The present invention is not limited to the embodiments described above. The above description of the specific embodiments is intended to describe and illustrate the technical solution of the present invention. The above specific embodiments are merely illustrative and not restrictive. Without departing from the scope of the present invention and the scope of protection of the claims, a person of ordinary skill in the art can also make many forms of specific changes under the enlightenment of the present invention.

Claims

1. A method for measuring gas path particle concentration based on characteristic waveform recognition, wherein the measurement system used comprises an electrostatic sensor, a signal preprocessing module, a signal acquisition module, and a particle concentration calculation module; the method for measuring gas path particle concentration comprises the following steps: S1: obtaining a preliminary preprocessed electrostatic sensor signal; S2: Extract the effective waveform position by cross-correlation matching. The method is: The electrostatic sensing signal after preliminary preprocessing is collected and transmitted to the particle concentration calculation module for processing; The formula for determining the contrast waveform used for cross-correlation comparison is as follows: Among them, parameters α, μ, and σ are unknown coefficients, and the parameter α is selected in the range of 0.2-1.0; parameters μ and σ jointly determine the shape characteristics of the waveform, and parameter design needs to consider the sampling rate and the gas flow rate in the gas pipeline; For a certain gas pipeline structure, the values ​​of parameters α, μ, and σ are determined through experiments, and the comparison waveforms are obtained; The absolute value of the collected original electrostatic sensor waveform signal is taken and then cross-correlated with the comparison waveform to obtain the waveform signal that meets the similarity condition in the original electrostatic sensor waveform; S3: Feature value extraction: take the absolute value of the peak value of the waveform signal that meets the similarity condition in the original acquisition signal and sum it as the feature value; S4: Establishing a model for the relationship between characteristic values ​​and particle concentration; S5: Particle concentration monitoring.

2. The method for measuring gas path particle concentration based on characteristic waveform recognition according to claim 1, characterized in that: The electrostatic sensor is embedded in the gas pipeline to obtain the electrostatic induction signal generated when charged particles pass through the gas pipeline. The signal preprocessing module is used to receive the original sampled electrostatic sensor signal and implement preprocessing; the signal acquisition module is used to implement high-frequency sampling of the preprocessed electrostatic sensor signal; The particle concentration calculation module is used to complete the cross-correlation comparison between the high-frequency sampling signal and the characteristic waveform and the subsequent particle concentration calculation.

3. The method for measuring gas path particle concentration based on characteristic waveform recognition according to claim 1, characterized in that: Before step S1, the definition and calculation method of the gas path particle concentration in the gas path pipeline are clarified: In the gas pipeline, the gas path particle concentration in the pipeline is defined as the ratio of the mass of gas path particles passing through the sensor cross section to the volume of air passing through in a certain period of time. The volume of air passing through per unit time is the product of the gas flow rate in the pipeline and the time, that is, For annular circulation gas pipelines, assuming that the flow velocity of the gas path particles in the pipeline is consistent with the wind speed in the pipeline and is evenly distributed in the pipeline, the time t0 for a single particle to circulate once is the ratio of the total volume V0 in the pipeline to the fan flow Q, that is, In a certain period of time t, the total weight of gas path particles passing through the sensor cross section is Where m is the mass of particles added to the gas circulation pipeline; the total volume of air passing through the sensor cross section within a certain time t is V=Q·t Within a certain time t, the concentration of gas path particles passing through the sensor cross section is 4. The method for measuring gas path particle concentration based on characteristic waveform recognition according to claim 1, characterized in that: The method of step S1 is: for the charge signal obtained by the electrostatic sensor, the signal preprocessing module sequentially processes it through charge amplification, voltage amplification, high-pass filtering and low-pass filtering circuits; The high-pass filter and low-pass filter modules are used to filter out power frequency interference and high-frequency noise respectively, so as to obtain electrostatic sensing signals that have undergone preliminary preprocessing.

5. The method for measuring gas path particle concentration based on characteristic waveform recognition according to claim 1, characterized in that: In step S2, the cross-correlation processing method is as follows: by cross-correlation calculation, the comparison waveform is scanned and compared with each position of the original waveform, and the similarity between each position in the original waveform and the comparison waveform is quantified to obtain a waveform after cross-correlation processing; By setting a threshold value, the peak waveform signal that meets the similar condition is extracted, and the waveform signal that meets the similar condition in the original electrostatic sensor waveform is obtained.

6. The method for measuring gas path particle concentration based on characteristic waveform recognition according to claim 4, characterized in that: The threshold is set to 0.

2.

7. The method for measuring gas path particle concentration based on characteristic waveform recognition according to claim 1, characterized in that: S4: The method for establishing a corresponding relationship model between characteristic values ​​and particle concentrations is: changing the particle concentration in the pipeline to conduct experiments, obtaining the corresponding characteristic value results under different particle concentrations, and using a linear fitting method to establish a corresponding relationship model between characteristic values ​​and particle concentrations, and complete the system calibration.

Images