Deposition early warning method and system for sonic nozzle gas flow standard device

Abstract

The invention relates to the technical field of gas flow measurement, and discloses a deposition early warning method and system for a sonic nozzle gas flow standard device, and the method comprises the steps: collecting a dynamic pressure signal of upstream gas of a sonic nozzle, and recognizing and segmenting a transient risk window; transmitting an ultrasonic detection signal to the airflow in the sonic nozzle, receiving a penetration signal of the airflow, and extracting a coherent acoustic characteristic quantity used for representing the state of a boundary layer of the throat wall surface of the nozzle; and performing comparative analysis on the coherent acoustic characteristic quantity and a pre-established reference characteristic spectrum to generate early warning information. According to the method, a progressive accurate diagnosis chain from macroscopic fluid dynamic capture to microscopic surface effect detection to metering risk quantitative evaluation is constructed, so that the fundamental technical difficulty that silence misalignment caused by deposition of a sonic nozzle gas flow standard device is difficult to perceive and quantitatively evaluate is solved; and the reliability of the detection result of the device is improved.

Description

Technical Field

[0001] This invention relates to the field of gas flow metering technology, and more specifically, to a deposition early warning method and system for a gas flow standard device for a sonic nozzle. Background Technology

[0002] The sonic nozzle method for gas flow rate standard devices is widely used as a reference device for transmitting gas flow rate values ​​due to its high accuracy and good repeatability. Its measurement accuracy depends fundamentally on the long-term stability of the nozzle throat geometry and surface aerodynamic characteristics.

[0003] In practical calibration, dynamic adjustment of the upstream pressure is required to establish the critical flow state. The industry has long focused on steady-state metrological performance and ensured it through periodic offline verification. However, in practice, it has been found that even with timely verification, some devices still exhibit inexplicable calibration data dispersion or systematic deviations during continuous use. The reason for this lies in a long-overlooked hidden risk: the pressure regulation process itself.

[0004] Traditional understanding views deposition as a slow, static process. However, in-depth research reveals that the real risk window is not in steady state, but rather exists at the instant of rapid valve activation. At this moment, the gas flow undergoes intense adiabatic expansion, causing a millisecond-level overshoot drop in local temperature at the nozzle throat. If the gas contains trace amounts of heavy hydrocarbons or water vapor, this transient low temperature is highly likely to trigger a condensation phase transition, forming an initial liquid film. More importantly, the initial deposition formed by phase transition or direct adsorption often begins at a thickness of only one molecule or a few molecular layers.

[0005] This nanoscale change in surface state can subtly alter the nozzle's discharge coefficient by affecting the turbulent structure of the wall boundary layer. Meanwhile, the macroscopic thermodynamic parameters, such as pressure and temperature, have already stabilized after the transient process ends, rendering traditional steady-state monitoring methods completely ineffective. This makes it impossible to provide early warnings during the deposition initiation stage, let alone assess its potential impact on the benchmark metrology performance, thus creating a hidden danger of silent inaccuracy in the reliability of the sonic nozzle as a metrology benchmark. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention discloses a deposition early warning method for a sonic nozzle gas flow rate standard device, the method comprising:

[0008] The dynamic pressure signal of the gas upstream of the sonic nozzle is collected. Based on the transient dynamic characteristics of the dynamic pressure signal, the transient risk window with the highest risk of phase change deposition is identified and segmented from the pressure regulation process.

[0009] Within the transient risk window, an ultrasonic detection signal is emitted into the airflow inside the sonic nozzle, and its penetration signal is received. By performing coherent demodulation analysis on the detection signal and the penetration signal, coherent acoustic features used to characterize the boundary layer state of the nozzle throat wall are extracted.

[0010] The coherent acoustic features are compared and analyzed with the pre-established baseline feature spectrum. Based on the analysis results, early warning information including the deposition risk level and its impact on the outflow characteristics of the sonic nozzle is generated.

[0011] Furthermore, the coherent demodulation analysis specifically involves complex acoustic impedance deviation spectrum analysis, including:

[0012] By performing cross-correlation and phase calculation on the detection signal and the penetration signal, a complex acoustic impedance spectrum reflecting the changes in complex acoustic impedance along the sound wave propagation path is obtained.

[0013] The deviation between the complex acoustic impedance spectrum within the transient risk window and the complex acoustic impedance spectrum under the steady-state reference is calculated to obtain the complex acoustic impedance deviation spectrum.

[0014] The scalar eigenvalues ​​of the complex acoustic impedance deviation spectrum at the dominant frequency are extracted as coherent acoustic eigenvalues.

[0015] Furthermore, methods for generating an assessment of the impact on nozzle outflow characteristics include:

[0016] The scalar eigenvalues ​​are input into a pre-calibrated quantization mapping model, which describes the mapping relationship between the scalar eigenvalues ​​and the correction amount of the theoretical outflow coefficient of the sonic nozzle.

[0017] Output the potential impact estimate of the current deposition state on the outflow coefficient. The potential impact estimate is reflected in the early warning information in the form of a correction percentage or uncertainty increment.

[0018] Furthermore, the acceleration of pressure change in the dynamic pressure signal is calculated;

[0019] The moment when the absolute value of the pressure change acceleration first exceeds the dynamic threshold is determined as the starting point of the transient risk window, where the dynamic threshold is determined based on the characteristics of historical pressure signals from the gas flow standard device during the stable operation phase.

[0020] After the starting point, analyze the spectral energy proportion of the dynamic pressure signal within the predetermined frequency band;

[0021] The point at which the proportion of spectral energy falls back from its peak to a predetermined proportion is defined as the end point of the transient risk window.

[0022] Furthermore, the method also includes a dynamic self-updating step for the reference feature spectrum:

[0023] After the calibration cycle is completed and no deposition risk is determined, the coherent acoustic features obtained this time are integrated into the historical benchmark feature spectrum with a preset weighting factor to achieve slow adaptive updating of the benchmark features.

[0024] Furthermore, before transmitting the ultrasonic detection signal, it also includes:

[0025] The encoding format of the ultrasonic detection signal is dynamically selected based on the estimated duration of the transient risk window.

[0026] When the window duration is less than the preset duration threshold, a unidirectional linear frequency modulated pulse is transmitted.

[0027] When the window duration is greater than or equal to the preset duration threshold, a pseudo-random binary encoded pulse is emitted.

[0028] Furthermore, when generating early warning information, sedimentation type identification is performed:

[0029] Analyze the response modes of different frequency components in coherent acoustic features;

[0030] If the low-frequency characteristic quantity changes significantly, the warning message indicates a tendency towards the risk of heavy hydrocarbon condensation; if the high-frequency characteristic quantity changes significantly, the warning message indicates a tendency towards the risk of particulate matter adsorption.

[0031] Furthermore, the method also includes a trigger condition optimization step:

[0032] By analyzing historical warning records, if a transient risk window is identified and a warning is triggered at the same relative time point during a predetermined number of consecutive pressure regulation processes, the system will automatically adjust the control parameters of the pressure regulating valve to smooth the pressure change rate at that time point and reduce the risk from the source.

[0033] Furthermore, the pre-established baseline characteristic spectrum is obtained through an in-situ calibration excitation step, specifically including:

[0034] After system initialization or maintenance, multiple pressure adjustment processes are performed under the clean state of the sonic nozzle. Coherent acoustic features are collected and their typical values ​​and fluctuation ranges are statistically calculated to construct a reference feature spectrum that accurately corresponds to the zero-deposition state.

[0035] Secondly, the present invention discloses a deposition early warning system for a sonic nozzle gas flow standard device, which is used to implement the deposition early warning method for the sonic nozzle gas flow standard device. The deposition early warning system for the sonic nozzle gas flow standard device includes: a high-frequency dynamic pressure sensing unit, a coherent ultrasonic processing unit, and an intelligent analysis and early warning unit.

[0036] A high-frequency dynamic pressure sensing unit is used to acquire dynamic pressure signals upstream of the sonic nozzle in real time.

[0037] The coherent ultrasonic processing unit is used to generate and transmit ultrasonic detection signals within the transient risk window, simultaneously receive penetration signals, and perform coherent demodulation analysis on the two signals to extract coherent acoustic features.

[0038] The intelligent analysis and early warning unit is connected to the coherent ultrasonic processing unit and has an embedded reference feature spectrum database.

[0039] The intelligent analysis and early warning unit is configured to: receive dynamic pressure signals and perform transient risk window segmentation, receive coherent acoustic feature quantities and compare and analyze them with the benchmark feature spectrum, and finally generate and output early warning information.

[0040] Compared with related technologies, the present invention has the following beneficial effects:

[0041] This invention precisely pinpoints the millisecond-level time window during pressure regulation—the time when the flow is most unstable and most likely to undergo adiabatic expansion and phase change—by calculating the transient dynamic characteristics of dynamic pressure signals. This allows the entire monitoring resource to focus on the physical moment when deposition actually occurs. Subsequently, within the locked transient window, coherent demodulation technology is used to process the ultrasonic signal penetrating the airflow, shifting the diagnostic dimension from bulk flow field parameters to the physical state of the wall boundary layer. Utilizing the high sensitivity of ultrasonic signals to viscoelastic changes on the wall caused by nanoscale adsorption or initial condensation, in-situ direct detection of surface initiation effects that would otherwise render traditional sensors ineffective is achieved. Finally, by intelligently comparing this microscopic acoustic characteristic with a statistically based benchmark spectrum, the abstract phase change is converted into a quantitative estimate of its impact on the core metrological parameter of the sonic nozzle, namely the outflow coefficient. This enables the warning output to include decision support information such as risk level and specific error prediction, solving the fundamental technical dilemma of the undetectable and unquantifiable silent inaccuracies caused by deposition in sonic nozzle gas flow standard devices, and improving the reliability of the device's detection results. Attached Figure Description

[0042] Figure 1 A schematic flowchart of the deposition early warning method for a sonic nozzle gas flow standard device provided by the present invention;

[0043] Figure 2 This invention provides a data processing flowchart for a deposition early warning system for a sonic nozzle gas flow standard device. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] Example 1

[0046] Please see Figure 1 As shown, this embodiment provides a deposition early warning method for a sonic nozzle gas flow standard device. The deposition early warning method includes steps one to three, which are described in detail below:

[0047] Step 1: Collect dynamic pressure signals of the gas upstream of the sonic nozzle. Based on the transient dynamic characteristics of the dynamic pressure signals, identify and segment the transient risk window with the highest risk of phase change deposition during the pressure regulation process.

[0048] In this embodiment, the transient risk window refers to a time interval during the pressure regulation (boosting or depressurizing) process upstream of the sonic nozzle, during which the valve action causes a drastic change in the airflow state, which is most likely to trigger gas adiabatic expansion and cooling and generate phase change deposition risk. This window usually lasts from tens of milliseconds to hundreds of milliseconds, which is much shorter than the duration of the entire pressure regulation process. It is the only time reference for this method to implement all subsequent diagnostic and early warning actions.

[0049] Transient dynamic characteristics refer to the physical features extracted from dynamic pressure signals that characterize the instantaneous rate of change of flow state and changes in energy distribution. In this embodiment, they specifically refer to two core features: pressure change acceleration and spectral energy migration in a specific frequency band (changes in the energy intensity of specific frequency components).

[0050] Pressure change acceleration describes the rate of change of pressure itself. For example, when pressure rises rapidly, its rate of change increases sharply from zero; this increase is acceleration. Numerically, it is the result of taking the second derivative of pressure with respect to time. When a valve operates rapidly, pressure change acceleration will show obvious positive or negative peaks, marking the beginning of a sudden change in the flow state.

[0051] Spectral energy migration refers to the phenomenon that the energy distribution of different frequency components in a pressure fluctuation signal changes significantly within a short period of time. For example, at the moment a valve is activated, the energy of certain specific frequencies may suddenly increase and then decrease.

[0052] An exemplary method for identifying and segmenting transient risk windows includes: calculating the pressure change acceleration of a dynamic pressure signal; determining the moment when the absolute value of the pressure change acceleration first exceeds a dynamic threshold as the starting point of the transient risk window, wherein the dynamic threshold is determined based on the historical pressure signal characteristics of a gas flow standard device during a stable operation phase; analyzing the spectral energy proportion of the dynamic pressure signal within a predetermined frequency band after the starting point; and determining the moment when the spectral energy proportion falls back from its peak to a predetermined proportion as the ending point of the transient risk window.

[0053] Specifically, the hardware configuration for data acquisition will be introduced first. The installation of the sensor is crucial to ensure that the measurement results of the dynamic pressure signal accurately reflect the core state of the flow field.

[0054] In this embodiment, a high-frequency dynamic pressure sensor can be used for measurement and installed through a standard pressure measurement port perpendicular to the pipe axis. This pressure measurement port can be located upstream of the sonic nozzle inlet, at a distance of at least 10 times the pipe diameter. This location is intended to ensure that the measurement point is in a stable flow region where the airflow is fully developed and large-scale eddies have decayed, thereby avoiding interference from local flow inhomogeneities on the pressure signal. During installation, it must be ensured that the sensor's pressure-sensing surface is flush with the inner wall of the pipe, without any protrusions or depressions, to prevent the generation of new flow disturbances.

[0055] In terms of signal acquisition, the sensor can be connected to a high-speed data acquisition card located in the control cabinet via a shielded cable. The acquisition card converts the continuous analog pressure signal into a digital signal at a sampling rate of no less than 20,000 times per second and transmits it to the processing unit in real time. This processing unit is generally integrated into the system, that is, the control system of the entire gas flow standard device.

[0056] For the real-time identification process of transient risk windows, under normal operating conditions, the system can continuously receive control signals (usually 4-20mA analog signals or bus digital commands) from the upstream regulating valve and simultaneously monitor the real-time pressure data stream collected by the aforementioned high-frequency pressure sensor.

[0057] Please see Figure 2 As shown, when the system clearly receives a pressure regulation command, or when real-time calculations detect that the rate of pressure change (i.e., the first derivative) continuously exceeds a very low set threshold, a pressure regulation process is determined to have started, and the transient risk window identification program is triggered.

[0058] The system can smooth and filter the acquired digital pressure sequence to suppress high-frequency electrical noise. Then, by calculating the ratio of the pressure difference between adjacent sampling points to the sampling time interval, the velocity of pressure change is obtained in real time. Next, using the same method, the rate of change of this velocity value itself is calculated to obtain the acceleration value of the pressure change, which directly quantifies the severity of the change in flow state.

[0059] Then, by continuously monitoring the absolute value of the pressure change acceleration calculated above, when this absolute value first exceeds a dynamically set threshold, the system immediately marks this moment as the start of the transient risk window. This threshold is not fixed but is based on learning from the equipment's historical normal operating data. For example, the system will statistically analyze the typical fluctuation range of pressure change acceleration during the steady flow phase and set a multiple (usually 3 times) of the upper limit of this range as the initial threshold. This setting method can automatically adapt to system characteristics under different pipe diameters, different media, and different initial pressures.

[0060] Next, the frequency component energy changes of the pressure signal are analyzed. Starting from the initial moment, the system performs short-time Fourier transform time-frequency analysis on the pressure data segment. This method is similar to a spectrometer, which can show the strength of each frequency component in the pressure fluctuation within each very short time slice. According to a large amount of industrial experimental data, in the transient process caused by valve action, the fluctuation energy in the frequency band of 1kHz to 10kHz usually concentrates and changes significantly. Therefore, this frequency band needs to be focused on in this embodiment.

[0061] Finally, based on the end time of the energy decay determination window, the system can calculate in real time the proportion of the fluctuation energy in the specific frequency band to the total fluctuation energy. Within the risk window, this proportion will first rise rapidly to a peak, and then gradually decrease. It can be determined that when the proportion drops more than half from the peak of this process (for example, to 50% of the peak), it means that the main transient energy accumulation process has basically ended, and the flow will tend to stabilize again. Therefore, this time point can be marked as the end time of the transient risk window.

[0062] Through the series of internal operations described above, the system can finally output a time interval [T0, T1] that is clearly defined by the start time T0 and the end time T1. This is the transient risk window that is precisely located for this specific voltage regulation action.

[0063] In the implementation of step one, by segmenting the transient risk window, a unified and accurate time coordinate system is provided for the entire early warning step. All subsequent analyses and judgments are based on the data within this window, making the diagnostic results closely match the physical process.

[0064] Step 2: Within the transient risk window, an ultrasonic detection signal is emitted into the airflow inside the sonic nozzle, and its penetration signal is received. By performing coherent demodulation analysis on the detection signal and the penetration signal, coherent acoustic features used to characterize the boundary layer state of the nozzle throat wall are extracted.

[0065] Coherent demodulation analysis is a signal processing technique that precisely extracts the amplitude and phase changes of a signal during propagation by comparing the received signal with a reference signal that is strictly synchronized with the transmitted signal. In this embodiment, the aim is to separate the subtle changes in the acoustic signal caused by minute changes in the wall state from strong background flow noise.

[0066] For example, the coherent demodulation analysis specifically includes complex acoustic impedance deviation spectrum analysis, which includes: performing cross-correlation and phase calculation on the probe signal and the penetration signal to obtain a complex acoustic impedance spectrum that reflects the changes in complex acoustic impedance along the sound wave propagation path; calculating the deviation between the complex acoustic impedance spectrum within the transient risk window and the complex acoustic impedance spectrum under the steady-state reference to obtain the complex acoustic impedance deviation spectrum; and extracting the scalar eigenvalues ​​of the complex acoustic impedance deviation spectrum at the dominant frequency as coherent acoustic features.

[0067] Among them, the complex acoustic impedance deviation spectrum is a key intermediate physical quantity obtained through coherent demodulation. Acoustic impedance describes the ease with which sound waves propagate in a medium and is a complex number containing a real part (resistance) and an imaginary part (resistance). The complex acoustic impedance spectrum is the distribution of acoustic impedance as a function of frequency. The deviation spectrum specifically refers to the difference between the spectrum measured within the transient risk window and the reference spectrum under steady-state clean conditions. This deviation directly reflects the additional impedance caused by deposits.

[0068] Coherent acoustic features refer to one or more specific parameters extracted from the above analysis for quantitative evaluation. In this embodiment, the scalar feature value of the complex acoustic impedance deviation spectrum is determined as the core feature because it is extremely sensitive to changes in wall viscoelasticity (such as the viscosity of the adsorbed liquid film).

[0069] Specifically, an ultrasonic transmitting transducer and an ultrasonic receiving transducer are installed at specific locations upstream and downstream of the sonic nozzle, respectively. During installation, it must be ensured that their acoustic faces are flush with the inner wall of the pipe, and that the central axes of the two transducers are strictly aligned to guarantee that the ultrasonic beam can effectively penetrate the airflow. The transmitting transducer is typically located in the upstream stable section (such as near the high-frequency dynamic pressure sensor mentioned in step one), and the receiving transducer is located sufficiently downstream of the nozzle (usually more than 5 times the pipe diameter) to ensure that the received signal has penetrated the entire flow field of interest (including the nozzle throat).

[0070] The system can provide a unified time base for the high-speed data acquisition card and the signal generation card through a synchronous clock generator. After the transient risk window is determined in the first step, the central processing unit immediately commands the signal generation card to generate and drive the transmitting transducer to emit ultrasonic detection signals only within the [T0, T1] window through precise time delay control.

[0071] It is worth mentioning that, in order to balance penetration capability, resolution and time requirements, in this embodiment, a linear frequency modulated pulse can be used as the detection signal. Its frequency is linearly scanned from a certain lower value to a certain higher value (e.g., from 0.8MHz to 1.2MHz) within the pulse duration. The advantage of this signal is that it has a large time-bandwidth product. Even if the transmission energy is low, a high signal-to-noise ratio and distance resolution can be obtained through subsequent pulse compression processing, thereby locating the acoustic interaction to the nozzle throat region.

[0072] For example, after successfully segmenting the transient risk window [T0, T1], the window duration ΔT (i.e., the time difference between T0 and T1) is calculated. Before performing coherent ultrasound detection, the system dynamically optimizes the encoding format of the detection signal based on this duration to ensure that the best quality diagnostic data is obtained within a limited time window.

[0073] The pulse width can be dynamically set according to the duration of the risk window. If the estimated window duration is very short, such as less than 5 milliseconds, a shorter linear frequency modulated pulse can be emitted. If the window is long, a longer pulse can be emitted or pseudo-random coding can be used to improve the average energy.

[0074] Specifically, a time threshold Th can be preset within the system (based on a comprehensive consideration of the ultrasonic transducer response characteristics, the minimum data length required for signal processing, and statistics of typical industrial voltage regulation processes. This ensures that the system has sufficient time to complete the complete transmission and reception of the pseudo-random coded signal; this value can be calibrated and fine-tuned during the system debugging phase according to the actual hardware performance). Before transmitting the signal, the system compares ΔT with Th:

[0075] If ΔT < Th, it is determined to be a short window. At this time, the time available for detection is extremely limited, and the system preferentially selects a unidirectional linear frequency modulated pulse as the transmission signal.

[0076] If ΔT ≥ Th, it is determined to be a long window, in which case there is relatively ample detection time, and the system selects a pseudo-random binary coded pulse as the transmitted signal. The core purpose of this dynamic selection mechanism is to always ensure that the signal-to-noise ratio and time resolution of the detected signal reach the optimal balance under the most stringent time constraints.

[0077] Next, for the reception of the ultrasonic detection signal, the ultrasonic receiving transducer converts the received sound wave signal into an electrical signal, which is then synchronously acquired by a high-speed acquisition card. The system reconstructs an identical ideal reference signal in the digital domain based on known transmitted signal parameters (such as FM start and end frequencies and duration). Then, a cross-correlation operation is performed between the received actual signal and this ideal reference signal. The essence of the cross-correlation operation is to slide the reference signal along the time axis to find the position that best matches the received signal. This process effectively suppresses random noise unrelated to the transmitted signal and accurately measures the total propagation time change of the ultrasonic wave from transmission to reception, as well as the attenuation of the signal amplitude. The output of the cross-correlation operation is a complex sequence containing amplitude and phase information.

[0078] According to acoustic principles, changes in sound wave propagation time and amplitude can be converted into changes in average acoustic impedance along the propagation path. Therefore, by processing the complex sequence obtained through cross-correlation operations, a complex acoustic impedance spectrum can be calculated, which reflects the acoustic characteristics of the fluid at different frequency components.

[0079] The key operation involves the system synchronously retrieving a pre-measured and stored baseline complex acoustic impedance spectrum (CAES) under conditions of no deposition risk (typically after initial system commissioning or thorough cleaning). Then, the difference between the CAES spectrum obtained within the current transient window and the baseline spectrum is calculated, yielding the CAES deviation spectrum. This deviation spectrum is stripped of the influence of background factors such as pipe geometry and fixed medium properties, purely characterizing the acoustic property changes that may be related to deposition caused by the current transient process.

[0080] The complex acoustic impedance deviation spectrum is a complex dataset containing a real part (in-phase component) and an imaginary part (quadrature component). Analysis revealed that the real part is mainly affected by bulk parameters such as fluid density and temperature. However, in this transient scenario, changes in these bulk parameters can be caused by a variety of factors, lacking strong specificity. The imaginary part, particularly the scalar eigenvalue calculated from the ratio of the imaginary to the real part, is extremely sensitive to the viscous dissipation effect of the boundary layer. Initial molecular adsorption or extremely thin liquid films significantly alter the acoustic vibration damping characteristics near the wall, clearly reflected in changes in the phase angle. Therefore, in this embodiment, the scalar eigenvalue of the complex acoustic impedance deviation spectrum at the dominant frequency is identified as the most representative coherent acoustic characteristic for subsequent risk assessment.

[0081] To support the extraction of acoustic features sensitive to nanoscale deposition from ultrasonic signals, this embodiment provides a complete and coherent signal processing flow as follows:

[0082] First, based on synchronous cross-correlation signal alignment and preprocessing, the system will process the received ultrasound penetration signal. With known transmission reference signal Cross-correlation operations are performed to achieve precise time alignment and initially improve the signal-to-noise ratio. Hilbert transforms are then applied to the aligned signal pairs to construct their respective analytic signals. and This makes it easier to directly obtain the instantaneous characteristics of the signal.

[0083] Secondly, the instantaneous phase difference is extracted as the core intermediate feature quantity, which is the instantaneous phase difference between the received signal and the transmitted signal. It can be obtained by calculating the difference in argument between the two analytical signals. The core relationship is: the instantaneous phase difference between the received signal and the transmitted signal. It is given by the following formula:

[0084] ;

[0085] in, This is the operator for taking the complex phase angle. It continuously and directly reflects the phase lag of sound waves along the propagation path caused by changes in medium properties (such as changes in the viscoelasticity of the wall), and is extremely sensitive to perturbations. Subsequently, the system monitors the transient risk window [T0, T1]. Feature extraction is performed on the sequence, such as calculating its root mean square value or a specific statistic, to obtain a scalar feature value that characterizes the overall phase shift within the current window. (This scalar eigenvalue) This refers to the core coherent acoustic features used for statistical comparison and early warning in the subsequent third step.

[0086] Next, the complex acoustic impedance spectrum is calculated for frequency domain analysis. Acoustic impedance is a core physical quantity characterizing the acoustic properties of a medium; it is a complex number that comprehensively reflects the medium's resistance to sound waves and its inertia. Its physical meaning is: in acoustics, for a plane wave model, the complex acoustic impedance at a certain point... Defined as the sound pressure at this point With particle vibration velocity The ratio, i.e. It is a frequency. The function, denoted as .

[0087] In this embodiment, estimation can be performed through frequency domain analysis, by conducting spectral analysis on the sound pressure signal at the receiving point. For the velocity of a particle The spectrum can be derived from sound pressure gradient measurements, or more practically, from the acoustic transfer function identified by the system, combined with the known spectrum of the transmitted signal. It is a complex number, which can be represented as:

[0088] ;

[0089] in, It is an imaginary unit (commonly used in electrical engineering). Replacement in mathematics To avoid confusion with the symbol for current, and to satisfy ); The real part is called acoustic impedance, which mainly characterizes the dissipation (absorption) of sound wave energy by the medium. It is the imaginary part, called acoustic impedance, which mainly characterizes the reaction of sound waves by the inertial or elastic effects of the medium (and its boundaries).

[0090] Finally, the complex acoustic impedance deviation spectrum is obtained. In order to eliminate the background influence of the invariant system characteristics such as pipes and fixed components, and to highlight the small changes caused by deposition, the deviation spectrum needs to be calculated.

[0091] The complex acoustic impedance spectrum calculated within the current detection period (transient risk window) can be denoted as: Retrieve the pre-calibrated and stored reference complex acoustic impedance spectrum of the nozzle under clean and healthy conditions from the system database. .

[0092] Spectral subtraction: For each frequency point Calculate the difference between the two:

[0093] ;

[0094] This is the complex acoustic impedance deviation spectrum, which specifically reveals the additional acoustic impedance that may be introduced by this transient process, and serves as a direct basis for subsequent evaluation. The aforementioned scalar eigenvalues... and deviation spectrum The changes in the imaginary part at the dominant frequency are strongly correlated and together constitute a set of coherent acoustic features for risk assessment.

[0095] In the implementation of step two, coherent demodulation combined with deviation spectral analysis enables the detection of initial adsorption at the molecular level, preceding any changes in macroscopic performance, thus achieving early warning. Simultaneously, ultrasonic path design and signal analysis ensure that the extracted characteristic quantities maximally reflect the wall state of the nozzle throat, a critical region directly related to metrological performance.

[0096] Moreover, step two relies on the triggering of step one to ensure that all expensive and sophisticated acoustic detection and analysis are precisely applied within the risk window, ensuring high system efficiency and avoiding the contamination of the baseline model by invalid data.

[0097] Step 3: Compare and analyze the coherent acoustic features with the pre-established benchmark feature spectrum. Based on the analysis results, generate early warning information that includes the deposition risk level and its impact on the sonic nozzle outflow characteristics.

[0098] Among them, coherent acoustic features specifically refer to the scalar feature values ​​extracted from step two and used for quantitative evaluation. (i.e., instantaneous phase difference) Statistical results within the transient risk window [T0, T1], such as root mean square value.

[0099] The baseline characteristic spectrum refers to the scalar characteristic values ​​collected and statistically analyzed by a sonic nozzle under known clean and healthy conditions within a typical pressure regulation transient risk window. A standard reference data set, which also includes average values. With normal fluctuation range Statistical characteristic models (such as standard deviation) are benchmarks for judging whether the current state is abnormal.

[0100] Furthermore, the method for generating an assessment of the impact on nozzle outflow characteristics includes: inputting scalar feature values ​​into a pre-calibrated quantization mapping model, which describes the mapping relationship between the scalar feature values ​​and the correction amount of the theoretical outflow coefficient of the sonic nozzle; and outputting an estimate of the potential impact of the current deposition state on the outflow coefficient, which is reflected in the early warning information in the form of a correction percentage or an uncertainty increment.

[0101] Among them, the quantization mapping model refers to a mathematical model established through experiments or theoretical analysis that links the degree of change of coherent acoustic characteristic quantities with the potential change of the sonic nozzle outflow coefficient. This model enables the system to translate abstract acoustic signal differences into a quantitative estimate of their impact on metrological performance.

[0102] The outflow characteristic refers to the core metering performance of the sonic nozzle, which is mainly reflected in the outflow coefficient. The outflow coefficient is a correction coefficient that converts the theoretical flow rate into the actual flow rate, and its stability directly determines the accuracy of the flow measurement.

[0103] In this embodiment, the pre-established reference feature spectrum is obtained through an in-situ calibration excitation step, which specifically includes: after system initialization or maintenance, performing multiple pressure adjustment processes in a clean state of the sonic nozzle, collecting coherent acoustic feature quantities and statistically calculating their typical values ​​and fluctuation ranges, thereby constructing a reference feature spectrum that accurately corresponds to the zero deposition state.

[0104] For example, the reliability of the reference characteristic spectrum is the cornerstone of the entire early warning system, and its establishment is accomplished through an in-situ calibration excitation step.

[0105] The initial setup begins with the nozzle undergoing professional cleaning and cleanliness verification. Then, the system automatically executes several standard pressure adjustment cycles at different pressure setpoints. For each cycle, the system records the transient risk window determined in the first step and performs coherent ultrasonic detection within that window. Finally, scalar characteristic values ​​extracted from all batches of data are statistically analyzed to calculate their average value. Sum of standard deviation (normal fluctuation range) This forms the initial baseline feature spectrum. , .

[0106] Step 3 occurs after a new voltage regulation process is completed, when the system obtains the scalar characteristic value of the current transient window and then executes a complete analysis and decision-making process.

[0107] Specifically, the system can retrieve pre-stored reference feature spectra from the storage unit. , and the current measurement obtained Compare this to the average μ and calculate its standard score. This value visually represents the degree to which the current measurement deviates from the typical healthy value, how many times it is within the normal fluctuation range, and the standard score. The larger the value, the higher the probability of an abnormal state and the greater the risk.

[0108] The system can calculate the standard score as described above. It compares the risk level with multiple preset thresholds to automatically classify the risk level and generate corresponding early warning information and maintenance suggestions.

[0109] Level 1 (Monitoring Level) Early Warning: When When the value is below the first threshold N1, it can be judged as "slight deviation," which means that although the current feature value fluctuates, it still falls within the range of... and Within the defined normal state statistics range of the nozzle itself. At this time, the system will record this event in the internal log and provide a gentle prompt on the human-machine interface (such as the status indicator light turning yellow), reminding the operator to pay attention to the subsequent trend of the nozzle, but no immediate maintenance action is required.

[0110] Level 2 (Action-Level) Warning: When S reaches or exceeds the first threshold N1 but falls below the second threshold N2, the system determines it to be a significant anomaly. This indicates that the nozzle throat condition has clearly deviated from the healthy baseline, with a high probability of deposition. Therefore, the system will generate a clear warning message, with the core recommendation being to significantly shorten the subsequent periodic calibration interval for this sonic nozzle. For example, if the original plan was to send it for inspection every six months, it is recommended to adjust it to monthly or quarterly online verification or rapid inspection, closely tracking the evolution of its metrological performance through more frequent monitoring.

[0111] Level 3 (Emergency) Warning: When When the second threshold N2 is reached or exceeded, the system determines it to be a serious anomaly, strongly indicating that deposition has likely occurred and has substantially affected the nozzle surface condition. The system then triggers the highest level audible and visual alarm and generates a warning message with clear instructions: it is recommended to immediately cease using the sonic nozzle for any standard transfer or precision measurement work, and to remove it from the device for professional offline cleaning and a comprehensive performance review.

[0112] The initial values ​​for thresholds N1 and N2 are primarily derived from statistical analysis of the large amount of historical health data used to establish the baseline feature spectrum. Specifically, by statistically analyzing the distribution of deviation multiples in the health dataset, N1 is set as a boundary value that can cover the vast majority (e.g., 95%) of normal fluctuations, while N2 is set as an extreme boundary value that is extremely rare in a healthy state (e.g., beyond 99.7%).

[0113] To directly link acoustic warnings with the most critical measurement accuracy, a quantization mapping model can be pre-set within the system. The core of this model is a proportionality coefficient K obtained through experimental calibration, the calibration process of which is as follows:

[0114] In the laboratory, using the same sonic nozzle as the target device, different known types (such as light hydrocarbon condensation, water film, dust adhesion) and different thicknesses (from sub-nanometer monolayers to micrometers) of deposition layers were simulated at its throat through controlled methods (such as micro-injection of specific component gases and temperature control), and each deposition state was denoted as m.

[0115] For each m, coherent acoustic characteristic quantity measurement is first performed: during the simulated voltage regulation transient process, steps one and two of this early warning method are executed to obtain the coherent acoustic characteristic quantity under this state, and its change value is recorded as... (i.e., scalar feature values ​​extracted from the current state) Compared with the average (the difference).

[0116] Next, the true value of the discharge coefficient is measured: immediately under steady-state conditions, the nozzle is precisely calibrated using a gas flow standard device with a higher accuracy class (such as a first-class standard), and the actual change in its discharge coefficient is measured. .

[0117] Finally, model fitting is performed: multiple sets of data are collected ( For data pairs of ΔCd, the proportionality coefficient can be obtained through linear regression analysis. , making This holds true in a statistical sense. For scenarios with complex relationships, piecewise linear regression or a lookup table can be used as a mapping model. Once established, this model serves as a knowledge base and is integrated into the intelligent analysis and early warning unit of the online early warning system. This enables the system to possess self-learning and preventative capabilities, continuously improving the reliability of the device.

[0118] In practical applications, the system will use the currently detected change value of the feature quantity. Input this model and calculate This allows for a quantitative estimate of the impact on the outflow coefficient, such as "the detected anomaly is expected to cause a shift of approximately -0.15% in the nozzle outflow coefficient." This value is output as an important component of the warning information. This enables users to not only receive qualitative alerts about potential risks but also to intuitively understand the severity of the risk, i.e., how much measurement error might be introduced, thus providing crucial quantitative evidence for decision-making.

[0119] Furthermore, when generating early warning information, the specific process of sediment type identification is performed: analyzing the response patterns of different frequency components in the coherent acoustic characteristic quantities; if the low-frequency characteristic quantity changes significantly, the early warning information indicates a tendency towards the risk of heavy hydrocarbon condensation; if the high-frequency characteristic quantity changes significantly, the early warning information indicates a tendency towards the risk of particulate matter adsorption.

[0120] In this embodiment, the system can also perform more refined analysis. For example, by observing the variation patterns of coherent acoustic characteristic quantities at different frequency components (e.g., whether the low-frequency band or the high-frequency band changes more), and comparing them with historical case libraries or physical models, a preliminary judgment on the sediment type can be added to the warning information, such as indicating "tendency towards heavy hydrocarbon condensation" or "tendency towards particulate matter adsorption". This can provide maintenance personnel with additional reference for selecting cleaning methods.

[0121] To enable the system to have self-learning and preventative capabilities and continuously improve device reliability, this early warning method also includes an optimization component. For example, a dynamic self-updating step for the reference feature spectrum: after a calibration cycle that determines there is no deposition risk has ended, the coherent acoustic features acquired this time are integrated into the historical reference feature spectrum with a preset weighting factor to achieve a slow, adaptive update of the reference features.

[0122] For example, to accommodate the extremely slow aging that may occur in the nozzle material, the system can introduce an adaptive mechanism. In each subsequent normal calibration cycle, if the warning system does not trigger any alarm (i.e., it is judged to be in a healthy state), the coherent acoustic characteristic Y measured in that cycle is incorporated into the historical reference spectrum with a small weighting factor, and μ and σ are slowly updated.

[0123] Because the aging of nozzle materials and transducer performance is an extremely slow process, typically measured in months or years, if the weighting factor is too large, random noise and atypical operating condition fluctuations contained in a single or a few measurements will be learned into the baseline too quickly and excessively, causing unexpected and rapid drift in the baseline μ and σ. This can cause the system to lose stability and may even misjudge aging signals as deposition risks. Therefore, in this embodiment, the weighting factor can be on the order of 0.001 or smaller, meaning that thousands of health measurements are required for a significant adjustment in the baseline. In this way, the basis for threshold determination (μ and σ) can gently adapt to the natural and slow performance baseline drift caused by long-term use of the equipment, thereby ensuring that the warning logic remains sensitive to sudden deposition throughout the entire equipment lifecycle while effectively avoiding false alarms caused by natural aging of the equipment.

[0124] Furthermore, the optimization section also includes a trigger condition optimization step: statistical analysis of historical warning records. If a transient risk window is identified and a warning is triggered at the same relative time point during a predetermined number of consecutive pressure regulation processes, the system automatically adjusts the control parameters of the pressure regulating valve to smooth the pressure change rate at that time point. This optimization step is an advanced extension function of this warning method. The system can execute the trigger condition optimization step, aiming to reduce repetitive risks from the source.

[0125] For example, the system can analyze historical warning records to identify a systematic pattern in which warnings are repeatedly triggered at the same time point during multiple consecutive pressure adjustments. Once such a pattern is confirmed, the system can automatically fine-tune the control parameters of the pressure regulating valve (e.g., optimize the valve's operating speed at that specific time point) to make the pressure change process smoother, thereby reducing or eliminating the transient risk at that fixed time point. This optimization process is usually implemented during maintenance or after simulation verification as an auxiliary function to continuously improve the operating status of the equipment.

[0126] It is worth mentioning that, in order to clarify how this method completes all monitoring and diagnostic steps in a short transient process, a supplementary explanation of the system timing design and feasibility is provided.

[0127] The core of this method lies in the fact that it does not require all signal acquisition, processing and analysis steps to complete all calculations and output the final results in real time and synchronously within the transient risk window. Instead, it adopts an architecture of real-time triggering capture and asynchronous fine processing. The timing allocation and estimated time consumption of key steps are shown in Table 1.

[0128] Table 1. Time allocation and estimated time consumption of key steps

[0129] Timing Stage Core tasks and operations Key technical support and performance indicators Estimated time T0: Window Recognition and Triggering 1. Calculate the acceleration due to pressure change. 2. Determine the threshold and lock the window starting point T0. 3. Send a synchronization trigger command to the ultrasonic unit. Hardware-level parallel computing is achieved using field-programmable gate arrays (FPGAs), with an algorithm latency of <1ms. ≤2ms T0→T1: Signal Synchronization Acquisition 1. Transmit a preset ultrasonic detection pulse (such as a linear frequency modulated signal). 2. Synchronously acquire and store the complete penetration signal waveform. 1. The signal generator and the acquisition card are controlled by the same clock source, with a synchronization error of <1µs. 2. The high-speed ADC stores data on the disk at a sampling rate of ≥10MS / s. Same as window duration (usually 20-200ms) After T1: Asynchronous Data Analysis 1. Perform in-depth processing on the stored acoustic data, including coherent demodulation and complex acoustic impedance spectrum calculation. 2. Extract feature quantities, compare them with benchmarks, and evaluate them using a quantization mapping model. 3. Generate the final early warning information. Complex algorithms are executed by a high-performance digital signal processor or a general-purpose CPU. This process occurs after data acquisition is complete and does not occupy the window time. Approximately 50-100ms

[0130] As shown in the table above, the core tasks of the system within the transient risk window (T0 to T1) are only "pressure event identification" and "acoustic waveform raw data capture." These two tasks are completed by high-speed dedicated hardware, with extremely short execution times, and are entirely within the window period. All computationally intensive data analysis tasks are executed asynchronously after the window ends, thus ensuring reliable capture and diagnosis of transient physical processes. The entire early warning process can be completed within a few hundred milliseconds after a pressure regulation operation, meeting the timeliness requirements of online monitoring.

[0131] In summary, by calculating the transient dynamic characteristics of the dynamic pressure signal, the millisecond-level time window during pressure regulation, where the flow is most unstable and most likely to undergo adiabatic expansion and phase change, is precisely identified. This allows the entire monitoring resource to focus on the physical moment when deposition actually occurs. Subsequently, within the identified transient window, coherent demodulation technology is used to process the ultrasonic signal penetrating the airflow, shifting the diagnostic dimension from bulk flow field parameters to the physical state of the wall boundary layer. Utilizing the high sensitivity of ultrasonic signals to viscoelastic changes on the wall caused by nanoscale adsorption or initial condensation, in-situ direct detection of surface initiation effects that would otherwise render traditional sensors ineffective is achieved. Finally, by intelligently comparing this microscopic acoustic characteristic with a statistically based benchmark spectrum, the abstract phase change is converted into a quantitative estimate of its impact on the core metrological parameter of the sonic nozzle, namely the outflow coefficient. This enables the warning output to include decision support information such as risk level and specific error prediction, solving the fundamental technical dilemma of the undetectable and unquantifiable silent inaccuracies caused by deposition in sonic nozzle gas flow standard devices, and improving the reliability of the device's detection results.

[0132] Example 2

[0133] This embodiment provides a deposition early warning system for a sonic nozzle gas flow standard device, which is used to implement the deposition early warning method for a sonic nozzle gas flow standard device disclosed in Embodiment 1. The system can be loaded as a subsystem onto the control system of the gas flow standard device. The deposition early warning system for the sonic nozzle gas flow standard device includes: a high-frequency dynamic pressure sensing unit, a coherent ultrasonic processing unit, and an intelligent analysis and early warning unit.

[0134] A high-frequency dynamic pressure sensing unit is used to acquire dynamic pressure signals upstream of the sonic nozzle in real time.

[0135] The coherent ultrasonic processing unit is used to generate and transmit ultrasonic detection signals within the transient risk window, simultaneously receive penetration signals, and perform coherent demodulation analysis on the two signals to extract coherent acoustic features.

[0136] The intelligent analysis and early warning unit is connected to the coherent ultrasonic processing unit and has an embedded reference feature spectrum database.

[0137] The intelligent analysis and early warning unit is configured to: receive dynamic pressure signals and perform transient risk window segmentation, receive coherent acoustic feature quantities and compare and analyze them with the benchmark feature spectrum, and finally generate and output early warning information.

[0138] Since this system uses the deposition warning method for the sonic nozzle gas flow standard device in Example 1, it has the same effect, which will not be repeated here.

[0139] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0140] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A deposition early warning method for a sonic nozzle gas flow rate standard device, characterized in that, The methods include: The dynamic pressure signal of the gas upstream of the sonic nozzle is collected. Based on the transient dynamic characteristics of the dynamic pressure signal, the transient risk window with the highest risk of phase change deposition is identified and segmented from the pressure regulation process. Within the transient risk window, an ultrasonic detection signal is emitted into the airflow inside the sonic nozzle, and its penetration signal is received. By performing coherent demodulation analysis on the detection signal and the penetration signal, coherent acoustic features used to characterize the boundary layer state of the nozzle throat wall are extracted. The coherent acoustic features are compared and analyzed with the pre-established baseline feature spectrum. Based on the analysis results, early warning information including the deposition risk level and its impact on the outflow characteristics of the sonic nozzle is generated.

2. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 1, characterized in that, Coherent demodulation analysis specifically involves complex acoustic impedance deviation spectrum analysis, including: By performing cross-correlation and phase calculation on the detection signal and the penetration signal, a complex acoustic impedance spectrum reflecting the changes in complex acoustic impedance along the sound wave propagation path is obtained. The deviation between the complex acoustic impedance spectrum within the transient risk window and the complex acoustic impedance spectrum under the steady-state reference is calculated to obtain the complex acoustic impedance deviation spectrum. The scalar eigenvalues ​​of the complex acoustic impedance deviation spectrum at the dominant frequency are extracted as coherent acoustic eigenvalues.

3. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 2, characterized in that, Methods for generating assessments of the impact on nozzle outflow characteristics include: The scalar eigenvalues ​​are input into a pre-calibrated quantization mapping model, which describes the mapping relationship between the scalar eigenvalues ​​and the correction amount of the theoretical outflow coefficient of the sonic nozzle. Output the potential impact estimate of the current deposition state on the outflow coefficient. The potential impact estimate is reflected in the early warning information in the form of a correction percentage or uncertainty increment.

4. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 1, characterized in that, Methods for identifying and segmenting transient risk windows include: Calculate the acceleration of pressure change in a dynamic pressure signal; The moment when the absolute value of the pressure change acceleration first exceeds the dynamic threshold is determined as the starting point of the transient risk window, where the dynamic threshold is determined based on the characteristics of historical pressure signals from the gas flow standard device during the stable operation phase. After the starting point, analyze the spectral energy proportion of the dynamic pressure signal within the predetermined frequency band; The point at which the proportion of spectral energy falls back from its peak to a predetermined proportion is defined as the end point of the transient risk window.

5. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 1, characterized in that, The method also includes a dynamic self-updating step for the reference feature spectrum: After the calibration cycle is completed and no deposition risk is determined, the coherent acoustic features obtained this time are integrated into the historical benchmark feature spectrum with a preset weighting factor to achieve slow adaptive updating of the benchmark features.

6. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 1, characterized in that, Before transmitting the ultrasonic detection signal, it also includes: The encoding format of the ultrasonic detection signal is dynamically selected based on the estimated duration of the transient risk window. When the window duration is less than the preset duration threshold, a unidirectional linear frequency modulated pulse is transmitted. When the window duration is greater than or equal to the preset duration threshold, a pseudo-random binary encoded pulse is emitted.

7. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 1, characterized in that, When generating early warning information, perform sediment type identification: Analyze the response modes of different frequency components in coherent acoustic features; If the low-frequency characteristic quantity changes significantly, the warning message indicates a tendency towards the risk of heavy hydrocarbon condensation; if the high-frequency characteristic quantity changes significantly, the warning message indicates a tendency towards the risk of particulate matter adsorption.

8. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 1, characterized in that, The method also includes a trigger condition optimization step: If, during a predetermined number of consecutive pressure regulation cycles, a transient risk window is identified and a warning is triggered at the same relative time point, the system will automatically adjust the control parameters of the pressure regulating valve to smooth the pressure change rate at that time point.

9. The deposition early warning method for a sonic nozzle gas flow standard device according to claim 1, characterized in that, The pre-established baseline characteristic spectrum is obtained through an in-situ calibration excitation step, specifically including: After system initialization or maintenance, multiple pressure adjustment processes are performed under the clean state of the sonic nozzle. Coherent acoustic features are collected and their typical values ​​and fluctuation ranges are statistically calculated to construct a reference feature spectrum that accurately corresponds to the zero-deposition state.

10. A deposition early warning system for a sonic nozzle gas flow rate standard device, used to implement the deposition early warning method for a sonic nozzle gas flow rate standard device according to any one of claims 1-9, characterized in that, The system includes: A high-frequency dynamic pressure sensing unit is used to acquire dynamic pressure signals upstream of the sonic nozzle in real time. The coherent ultrasonic processing unit is used to generate and transmit ultrasonic detection signals within the transient risk window, simultaneously receive penetration signals, and perform coherent demodulation analysis on the two signals to extract coherent acoustic features. The intelligent analysis and early warning unit is connected to the coherent ultrasonic processing unit and has an embedded reference feature spectrum database. The intelligent analysis and early warning unit is configured to: receive dynamic pressure signals and perform transient risk window segmentation, receive coherent acoustic feature quantities and compare and analyze them with the benchmark feature spectrum, and finally generate and output early warning information.

Images