A method and system for analyzing the pumpability of geopolymer concrete

Abstract

The application provides a geopolymer concrete pumpability analysis method and system, and relates to the technical field of geopolymer concrete pumpability analysis; the application sets a measuring pipe section on a pumping pipeline, and real-time collection of ultrasonic sound velocity and main frequency attenuation coefficient constitutes a two-dimensional joint characteristic value sequence; the application generates a calibrated health voiceprint baseline by weighting and fusing the real-time data on site with preset prior knowledge, and compensates by using the real-time temperature of the slurry; in the pumping process, the application continuously calculates an instability evolution index reflecting the deviation of the slurry from the stable state, and compares the index with two-stage early warning thresholds to determine the pumpability. The application adopts a dynamic baseline calibration mode of weighting and fusing the real-time waveform on site with preset waveform templates, effectively solves the problem that the fixed baseline is easy to fail in the traditional method, and further fuses the multi-dimensional acoustic characteristics with the dynamic calibration baseline, so that the early instability evolution of the internal structure of the slurry can be sensitively captured.

Description

Technical Field

[0001] This invention relates to the field of geopolymer concrete pumpability analysis technology, specifically to a method and system for analyzing the pumpability of geopolymer concrete. Background Technology

[0002] Geopolymer concrete is a novel inorganic cementitious material formed through depolymerization-condensation reactions using aluminosilicate industrial wastes such as fly ash and slag as raw materials under the action of an alkali activator. Due to its low-carbon and environmentally friendly properties, resistance to chemical corrosion, and excellent mechanical properties, it is increasingly being adopted in engineering scenarios with stringent requirements for concrete pumpability, such as long-distance high-rise building pumping construction, continuous tunnel lining casting, and integral molding of large-volume foundation slabs. However, unlike mature cement-based concrete, the rheological properties of geopolymer concrete are extremely sensitive to time and ambient temperature. During pumping, especially when the supply is interrupted or construction procedures change, the geological polymerization reaction within the slurry continues, and the cementitious products gradually build a network structure, causing the yield stress and plastic viscosity of the slurry to increase rapidly. Once this structure exceeds the pressure compensation capacity of the pumping system, localized blockages will form in the pipeline. This can result in significant time and costly pipeline cleaning, or even the complete failure of the pumping pipeline, or even a pipe burst. Therefore, how to perceive the dynamic evolution of the internal structure of the polymer concrete slurry in the pipeline in real time and accurately at the construction site, and to determine the current pumpability status in real time, is an urgent problem to be solved to achieve safe and efficient pumping.

[0003] Currently, the main methods used in the industry to monitor the pumpability of concrete include pipeline pressure monitoring and offline rheological testing methods such as pipe runners. The former involves installing pressure sensors at the pump outlet or key pipeline nodes to monitor real-time changes in pressure values; the latter involves sampling the slurry before pumping and using a rheometer to measure its rheological parameters over time. However, existing methods have significant shortcomings in real-time assessment of pumpability: pipeline pressure monitoring is a passive detection method, reflecting the macroscopic mechanical response triggered after the slurry structure has been established to a certain extent. At this point, a high pressure gradient has often formed in the pipeline, and the risk of blockage has already occurred. The intervention window for operators is extremely short, making it a post-event response and unable to provide critical instability identification signals during the instability evolution of the internal structure of the slurry. Offline methods such as tube slides can obtain rheological parameters, but their test environment differs greatly from the actual pressure, shear, and temperature field coupling conditions inside the pipeline. Furthermore, they cannot continuously track the instantaneous establishment rate of the slurry structure during the intermittent period. For geopolymer concrete, which is highly sensitive to temperature and has rapid reaction kinetics, the data from such open-loop tests cannot be correlated with the real-time state of the slurry inside the pipe, thus losing its guiding significance for real-time assessment of the current pumpability state.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for analyzing the pumpability of geopolymer concrete, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for analyzing the pumpability of geopolymer concrete, comprising the following steps: Step 1: Install a measuring pipe section in the pumping pipeline. The measuring pipe section is equipped with an acoustic window. Through an ultrasonic transducer installed on the outside of the acoustic window, emit instantaneous pulse ultrasonic waves into the polymer concrete slurry inside the pipe using the pulse transmission method, and receive the transmitted signal after passing through the slurry. At the same time, collect the real-time temperature of the slurry in the measuring pipe section. Step 2: Under pumping conditions, extract the ultrasonic velocity and the dominant frequency attenuation coefficient from the transmitted signal to form a two-dimensional joint feature value sequence that varies with time; in the initial stage of pumping, extract a segment of the two-dimensional joint feature value sequence as a healthy baseline sample under normal pumping conditions, calculate the statistical characteristics of the healthy baseline sample, and construct a healthy acoustic waveform baseline that characterizes the stable flow state of the slurry. Step 3: During continuous pumping, continuously extract the two-dimensional joint feature value sequence and calculate the real-time fluctuation characteristics; use the real-time temperature of the slurry to perform temperature compensation on the real-time fluctuation characteristics and the healthy acoustic waveform baseline, and continuously compare the temperature-compensated real-time fluctuation characteristics with the temperature-compensated healthy acoustic waveform baseline to obtain the instability evolution index that reflects the degree of deviation of the current internal structure of the slurry from the stable state. Step 4: Compare the instability evolution index with the preset two-level warning thresholds to determine the current pumpability status of the pumping pipeline.

[0007] Furthermore, the measuring pipe section is a detachable structure, connected to the straight section of the pumping pipeline via a flange, and an alkali-resistant sound-transmitting window flush with the inner wall of the measuring pipe section is embedded in the inner wall of the measuring pipe section. The ultrasonic transducer is a transceiver integrated longitudinal wave transducer with a center frequency in the range of [50kHz, 500kHz]. It is fixed to the outer surface of the acoustic window through a coupling agent layer to emit instantaneous pulsed ultrasonic waves perpendicular to the slurry flow direction and receive transmitted signals.

[0008] Furthermore, the ultrasonic velocity is calculated by the transit time of the transmitted signal and the inner diameter of the measured pipe section, and the dominant frequency attenuation coefficient is calculated by the logarithmic attenuation rate of the dominant frequency component of the transmitted signal amplitude spectrum relative to the dominant frequency component of the transmitted signal amplitude spectrum; the ultrasonic velocity and the dominant frequency attenuation coefficient are synchronously acquired at equal time intervals to form a two-dimensional joint feature value sequence that varies with time.

[0009] Furthermore, the healthy voiceprint baseline is constructed as follows: In the same proportion of geopolymer concrete samples as the construction site, pre-pumping tests were conducted at different temperatures. During the period when the pumping pressure was within the preset steady-state fluctuation range and the sample entropy of the two-dimensional joint feature value sequence was lower than the preset stability threshold, the two-dimensional joint feature value sequence was extracted as the pre-test baseline sample. The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the dominant frequency attenuation coefficient sequence in the pre-test baseline sample are calculated respectively to form a preset multidimensional statistical feature vector; at the same time, the original curve shape of the two-dimensional joint feature value of the pre-test baseline sample changing with time is recorded as a preset waveform template; the preset waveform template is the time-domain waveform shape representation of the two-dimensional joint feature value sequence; the preset multidimensional statistical feature vector at the corresponding temperature and the preset waveform template are associated and stored to construct a preset healthy voiceprint baseline library at different temperatures; During actual pumping, when pumping begins and the online calibration triggering condition is met, a two-dimensional joint feature value sequence with a duration of the first preset duration is extracted as an on-site calibration sample; the mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the main frequency attenuation coefficient sequence in the on-site calibration sample are calculated respectively to form an on-site multidimensional statistical feature vector; at the same time, the original curve shape of the two-dimensional joint feature value of the on-site calibration sample over time is recorded as an on-site waveform template; A weighted average method is used to fuse the on-site multidimensional statistical feature vector with the preset multidimensional statistical feature vector at the corresponding temperature to generate a calibrated multidimensional statistical feature vector; the on-site waveform template is weighted and fused with the preset waveform template at the corresponding temperature to generate a calibrated waveform template. Each feature component in the calibrated multidimensional statistical feature vector is mapped to its equivalent value at a preset reference temperature according to a preset temperature normalization function to obtain a temperature-normalized health baseline multidimensional statistical feature vector; the temperature-normalized health baseline multidimensional statistical feature vector and the calibrated waveform template are stored together as a health voiceprint baseline. The online calibration triggering conditions are: the fluctuation amplitude of the pumping pressure within a continuous first preset time period does not exceed the preset pressure fluctuation threshold, and the sample entropy of the on-site multidimensional statistical feature vector does not exceed the preset multiple threshold of the sample entropy of the preset multidimensional statistical feature vector at the corresponding temperature retrieved from the preset healthy voiceprint baseline library.

[0010] Furthermore, the calculation method for the real-time fluctuation characteristics is as follows: Using the current sampling time as the endpoint, a two-dimensional joint feature value sequence with a duration of the second preset duration is extracted to form a real-time observation window; The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the dominant frequency attenuation coefficient sequence within the real-time observation window are calculated to form a real-time multidimensional statistical feature vector, which serves as the real-time fluctuation feature. The second preset duration is no greater than the first preset duration.

[0011] Furthermore, the instability evolution index is calculated as follows: Calculate the deviation distance between the real-time multidimensional statistical feature vector and the temperature-normalized multidimensional statistical feature vector of the healthy baseline in the healthy voiceprint baseline, and map the deviation distance to the [0,1] interval; Simultaneously, the dynamic time warping distance between the two-dimensional joint feature value sequence within the real-time observation window and the calibrated waveform template in the healthy voiceprint baseline is calculated; the dynamic time warping distance and the deviation distance are weighted and fused to obtain the instability evolution index.

[0012] Furthermore, the temperature compensation specifically includes: Before calculating the deviation distance, each feature component in the real-time multidimensional statistical feature vector corresponding to the real-time observation window is mapped to its equivalent value at a preset reference temperature using the same preset temperature normalization function to obtain a temperature-normalized real-time multidimensional statistical feature vector. The deviation distance is calculated based on the temperature-normalized real-time multidimensional statistical feature vector and the temperature-normalized health baseline multidimensional statistical feature vector.

[0013] Furthermore, the two-level warning thresholds include a first warning threshold and a second warning threshold, and the second warning threshold is greater than the first warning threshold; both the first warning threshold and the second warning threshold are represented by an instability evolution index; The pumpability status includes normal, critical instability, and impending blockage. The logic for determining the pumpability status of the pumping pipeline is as follows: If the instability evolution index is less than the first warning threshold, the current pumpability is determined to be normal; if the instability evolution index is not less than the first warning threshold and less than the second warning threshold, the current pumpability is determined to be critically unstable, triggering a warning signal; if the instability evolution index is not less than the second warning threshold, the current pumpability is determined to be about to be blocked, triggering a forced intervention signal.

[0014] Furthermore, the specific values ​​of the first warning threshold and the second warning threshold are determined in the following manner: In a geopolymer concrete slurry sample with the same mix ratio as the construction site, pre-pumping tests were conducted at different temperatures. The two-dimensional joint feature value sequence from normal pumping to the entire process of pipe blockage was recorded as the pre-pumping test feature data. The pre-pumping test feature data was labeled with the corresponding pumpability status label frame by frame. Based on the characteristic data of the pre-pumping test, the instability evolution index is calculated frame by frame to form an instability evolution index sample set; Using the pumpability state label as the classification ground truth and the unstable evolution index sample set as the input variable of the classifier, the receiver operating characteristic curve analysis method is adopted to determine the first and second warning thresholds with the optimization objective of maximizing the early warning time of blockage and minimizing the false alarm rate.

[0015] The present invention also provides a system for analyzing the pumpability of geopolymer concrete, which is used to implement the above-mentioned method for analyzing the pumpability of geopolymer concrete, comprising: The multi-parameter in-situ acquisition module is used to install a measuring pipe section in the pumping pipeline. The measuring pipe section is equipped with an acoustic window. Through an ultrasonic transducer installed on the outside of the acoustic window, instantaneous pulse ultrasonic waves are emitted into the polymer concrete slurry inside the pipe using the pulse transmission method, and the transmitted signal after passing through the slurry is received. At the same time, the real-time temperature of the slurry in the measuring pipe section is acquired. The slurry structure quantification module is used to extract the ultrasonic velocity and the dominant frequency attenuation coefficient from the transmitted signal under pumping conditions, forming a two-dimensional joint feature value sequence that varies with time. In the initial stage of pumping, a segment of the two-dimensional joint feature value sequence is extracted as a healthy baseline sample under normal pumping conditions. The statistical characteristics of this healthy baseline sample are calculated to construct a healthy acoustic waveform baseline that characterizes the stable flow state of the slurry. The pressure increase prediction module is used to continuously extract the two-dimensional joint feature value sequence during continuous pumping and calculate the real-time fluctuation characteristics. It uses the real-time temperature of the slurry to perform temperature compensation on the real-time fluctuation characteristics and the healthy acoustic waveform baseline, and continuously compares the temperature-compensated real-time fluctuation characteristics with the temperature-compensated healthy acoustic waveform baseline to obtain the instability evolution index that reflects the degree of deviation of the current internal structure of the slurry from the stable state. The pumpability analysis and determination module is used to compare the instability evolution index with the preset two-level warning thresholds to determine the current pumpability status of the pumping pipeline.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention employs pulse transmission ultrasound technology on in-situ measurement pipe sections of pumping pipelines to continuously and in real-time capture ultrasonic velocity and dominant frequency attenuation coefficient without interfering with slurry flow and chemical reactions, constructing a two-dimensional joint feature value sequence that varies over time. In the initial stage of pumping, a healthy baseline sample under normal pumping conditions is extracted, and its statistical characteristics are calculated to construct a healthy acoustic waveform baseline characterizing the stable flow state of the slurry. During continuous pumping, the two-dimensional joint feature value sequence is continuously extracted, and real-time fluctuation characteristics are calculated. After temperature compensation of the fluctuation characteristics and the healthy acoustic waveform baseline using the real-time temperature of the slurry, the two are continuously compared to obtain an instability evolution index reflecting the degree to which the current internal structure of the slurry deviates from a stable state. This invention normalizes the influence of temperature on ultrasonic characteristics, making data collected under different temperature conditions comparable and significantly improving the accuracy of state determination. This invention also innovatively introduces a weighted fusion mechanism based on the deviation distance of the healthy acoustic waveform baseline and the dynamic time-regulated distance between it and the waveform template to generate an instability evolution index, effectively improving the detection sensitivity of subtle changes in the early stages of slurry instability. By comparing the instability evolution index with the preset two-level warning threshold, the pumpability state of geopolymer concrete can be divided into three levels: normal, critical instability, and impending blockage. Thus, in the early stages of the intermittent period before the signs of blockage are obvious, graded warnings or mandatory intervention signals can be issued. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a schematic diagram illustrating the trend of the two-dimensional joint feature value sequence over time in this invention; Figure 3 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] Example: Please see Figures 1 to 2 The present invention provides a technical solution: A method for analyzing the pumpability of geopolymer concrete, comprising the following steps: Step 1: Install a measuring pipe section in the pumping pipeline. The measuring pipe section is equipped with an acoustic window. Through an ultrasonic transducer installed on the outside of the acoustic window, emit instantaneous pulse ultrasonic waves into the polymer concrete slurry inside the pipe using the pulse transmission method, and receive the transmitted signal after passing through the slurry. At the same time, collect the real-time temperature of the slurry in the measuring pipe section.

[0021] In this embodiment, the measuring pipe section is a detachable structure, connected to the straight section of the pumping pipeline via a flange. The use of a detachable flange-connected measuring pipe section facilitates cleaning or replacement of the acoustic window surface during pumping intervals, thereby eliminating the attenuation of acoustic coupling efficiency caused by the deposition of early hydration products of the geopolymer slurry on the window surface. It also allows for the replacement of measuring pipe sections of corresponding sizes according to pumping pipelines of different diameters, improving the field adaptability of the system.

[0022] An alkali-resistant acoustic window, flush with the inner wall of the measuring pipe section, is embedded in its inner surface. The preferred material for this window is alumina ceramic or silicon carbide ceramic, and its thickness is determined based on the center frequency of the selected ultrasonic transducer, typically taking an integer multiple of half the longitudinal wave wavelength at that frequency to achieve maximum sound transmittance. The inner surface of the acoustic window is flush with the inner wall of the measuring pipe section to avoid forming steps or depressions in the slurry flow path, preventing the generation of local turbulent vortices from causing additional scattering interference to the acoustic signal. The joint between the acoustic window and the inner wall of the pipe section is filled with alkali-resistant sealant to ensure that the overall sealing of the pipeline is not lower than the rated operating pressure of the pumping system.

[0023] The ultrasonic transducer is a transceiver integrated longitudinal wave transducer with a center frequency in the range of [50kHz, 500kHz]. It is fixed to the outer surface of the acoustic window through a coupling agent layer to emit instantaneous pulsed ultrasonic waves perpendicular to the slurry flow direction and receive transmitted signals.

[0024] The specific value of the center frequency of the ultrasonic transducer should be determined by comprehensively considering three factors: the maximum aggregate particle size of the geopolymer concrete slurry, the characteristic acoustic attenuation coefficient of the slurry, and the inner diameter of the measuring pipe section. The specific determination method is as follows: The lower frequency limit is determined based on the following: the ultrasonic wave length should be no less than 3 to 5 times the maximum aggregate particle size in the slurry to avoid aggregate particles entering the Rayleigh scattering region of the sound wave wavelength, causing severe scattering attenuation and waveform distortion of the transmitted signal. Let the maximum aggregate particle size in the geopolymer concrete slurry be... The ultrasonic velocity in the slurry is The transducer center frequency Should meet Typically, for fine aggregate geopolymer slurries with a maximum aggregate size of 10 mm, if the predicted sound velocity... For speeds of approximately 1800 m / s to 2200 m / s, the upper frequency limit should not exceed approximately 36 kHz to 44 kHz; for paste or mortar systems with a maximum aggregate size of 5 mm, this upper frequency limit is approximately 72 kHz to 88 kHz. Therefore, when measuring geopolymer mortar or paste without coarse aggregate, a higher operating frequency can be selected within the range of [50 kHz, 500 kHz] to obtain better time resolution; when measuring objects containing coarse aggregate, frequencies near the lower limit of this range should be preferred.

[0025] The upper frequency limit is determined by the attenuation coefficient of sound waves in the slurry. With frequency Positive correlation can be approximated as In this context, the exponent n typically takes the value 1 to 2 in heterogeneous suspension systems. It is a proportionality coefficient related to the properties of the medium itself. It represents the attenuation value at a specific reference frequency (usually 1 MHz), and its value depends on the composition of the geopolymer concrete paste (such as aggregate type, paste concentration, viscosity, etc.). Parameter The values ​​of n and are usually determined through preliminary experiments, measuring the acoustic attenuation of geopolymer concrete slurry with known mix proportions at different frequencies, and then determining their specific values ​​through curve fitting (such as the least squares method), rather than a fixed universal constant.

[0026] If the transducer frequency is too high, the signal-to-noise ratio (SNR) of the transmitted signal will be severely degraded, and the amplitude of the transmitted signal may even be submerged below the noise floor of the receiving circuit. Therefore, the upper limit of the transducer's center frequency should be determined by the constraint that a stable transmitted signal with an SNR of not less than 20 dB can still be obtained under the condition of the maximum measuring pipe diameter. In actual selection, for geopolymer slurry with a specific mix ratio, a penetration test can be conducted in the laboratory using a simulated pipe section with the same inner diameter as the measuring pipe section, and the frequency can be gradually reduced until the above SNR requirement is met, thereby determining the upper limit of the frequency.

[0027] The couplant layer is used to fill the microscopic air gap between the ultrasonic transducer's radiating surface and the outer surface of the acoustic window, achieving acoustic impedance matching. The couplant should be an industrial ultrasonic couplant with a moderate viscosity, which does not cure or flow within the pumping ambient temperature range; epoxy resin-based high-viscosity couplants are preferred. The thickness of the couplant layer should be controlled to be less than one-quarter of the longitudinal wave wavelength at its operating frequency, and air bubbles should be evenly squeezed out during application to ensure consistent acoustic coupling.

[0028] For geopolymer mortar systems containing fine aggregate (maximum particle size not exceeding 5 mm), a transceiver with a center frequency of 100 kHz to 200 kHz is preferred; for geopolymer concrete systems containing coarse aggregate (maximum particle size not exceeding 25 mm), a transceiver with a center frequency of 50 kHz to 80 kHz is preferred. The rationale for this selection range lies in balancing sufficient penetration capability with sensitivity to detecting changes in the microstructure of the mortar.

[0029] Mechanical slots or positioning marks are provided on the outer surface of the acoustic window to assist in the repeated installation and positioning of the ultrasonic transducer, ensuring the consistency and reproducibility of the sound beam path after each installation.

[0030] A thermocouple temperature sensor is also embedded in the wall of the measuring pipe section to collect the real-time temperature of the slurry inside the measuring pipe section when receiving the transmitted signal.

[0031] The installation method and response characteristics of thermocouple temperature sensors must meet the following requirements: Installation Location: The temperature sensing end of the thermocouple temperature sensor should be embedded inside the measuring tube section wall, near the inner wall surface, and located circumferentially within a range of 5mm to 20mm upstream of the acoustic window. This ensures that the measured temperature accurately reflects the actual temperature of the slurry flowing through the sound beam propagation path. The radial distance between the temperature sensing end and the inner wall surface of the measuring tube section is preferably 1mm to 3mm. The thermal conductivity of the tube wall material itself ensures timely temperature response while preventing the sensor from being directly exposed to the scouring and alkaline corrosion of the flowing slurry.

[0032] Response time: The thermal response time constant of the selected thermocouple temperature sensor should not exceed 5 seconds to ensure that the temperature acquisition can reflect the average temperature trend of the slurry near the sound beam propagation path when the slurry temperature changes rapidly.

[0033] The sampling rate of the thermocouple temperature sensor should be consistent with the acquisition frequency of the ultrasonic transmission signal. That is, each time an ultrasonic pulse is emitted and a transmission signal acquisition is completed, the current temperature value is recorded synchronously, forming a strictly one-to-one time-stamped data pair. The synchronously recorded current temperature measurement value is the real-time reading output by the thermocouple at that sampling moment, used to characterize the temperature state of the slurry near the sound beam propagation path at that moment. Given that the thermocouple itself has a thermal response hysteresis of no more than 5 seconds, under extreme conditions of rapid temperature changes, the temperature reading can be corrected in post-processing using a moving average of the temperature sequence or a first-order hysteresis compensation algorithm to further improve the temperature compensation accuracy.

[0034] In addition, the thermocouple temperature sensor should be calibrated and verified before the first use of the measuring pipe section and after each cumulative pumping volume (e.g., 500 cubic meters). The calibration can be performed using a two-point method with an ice-water mixture (0°C) and boiling water (100°C, which should be corrected according to the local atmospheric pressure) to ensure that the static temperature measurement error does not exceed ±0.5°C.

[0035] Step 2: Under pumping conditions, extract the ultrasonic velocity and dominant frequency attenuation coefficient from the transmitted signal to form a two-dimensional joint feature value sequence that varies with time; in the initial stage of pumping, extract a segment of the two-dimensional joint feature value sequence as a healthy baseline sample under normal pumping conditions, calculate the statistical characteristics of the healthy baseline sample, and construct a healthy acoustic baseline that characterizes the stable flow state of the slurry.

[0036] In this embodiment, the ultrasonic velocity The value is calculated by measuring the transit time of the transmitted signal and the inner diameter of the pipe section. The specific calculation formula is as follows: Where D is the inner diameter of the measuring pipe section (in meters). The transit time (in seconds) of the ultrasonic transmission signal.

[0037] The transit time is determined as follows: the start time of the transmitted pulse is taken as the zero point, and the arrival time is taken as the moment of the first positive zero-crossing in the transmitted signal waveform. The difference between the two is the transit time. The first positive zero-crossing refers to the moment when the signal waveform crosses from a negative value to a positive value after the amplitude of the transmitted signal first exceeds a preset amplitude threshold (usually 10% of the maximum amplitude). Zero-crossing detection, rather than peak detection, is used as the basis for determining the arrival time because the zero-crossing position is not sensitive to signal amplitude fluctuations. Even when changes in the acoustic attenuation of the slurry cause fluctuations in the received signal amplitude, it can still provide a stable and accurate estimate of the arrival time, thus ensuring the robustness of the sound velocity calculation.

[0038] To eliminate the fixed acoustic delay introduced by the wall of the measuring pipe section and the acoustic window, the system needs to be calibrated beforehand. The calibration method is as follows: Fill a pipe section of the same specifications as the measuring pipe section with a standard medium of known sound velocity (e.g., pure water at room temperature, whose sound velocity can be obtained from a table based on the measured temperature using classical formulas). Measure the apparent transit time of the ultrasonic pulse as it passes through this standard medium. Subtract the ideal transit time calculated based on the sound velocity of the standard medium from this value; the difference is the system's fixed delay. In actual pumping measurements, the apparent transit time obtained using the zero-crossing method described above is subtracted from... Then, the actual transit time of the slurry is determined. .

[0039] The main frequency attenuation coefficient The logarithmic attenuation rate of the dominant frequency component of the transmitted signal amplitude spectrum relative to the dominant frequency component of the transmitted signal amplitude spectrum is calculated using the following formula: Where L is the propagation distance of the sound wave in the slurry, which in this embodiment is equal to the inner diameter D of the measuring pipe section; The amplitude spectrum of the transmitted signal at the dominant frequency The amplitude at that point (unit: V or dimensionless relative value); For the amplitude spectrum of the transmitted signal at the main frequency The amplitude at a given point (unit: V or dimensionless relative value); ln(·) represents the natural logarithm.

[0040] clock speed The frequency component with the largest amplitude in the amplitude spectrum of the emitted ultrasonic pulse is defined as the actual operating frequency of the transmitting transducer. In actual measurements, due to manufacturing tolerances and installation variations in the transducer, the actual frequency of the transmitted signal may deviate slightly from its nominal center frequency. Therefore, the frequency should be accurately determined through spectral analysis of the actual transmitted signal. .

[0041] The unit of the main frequency attenuation coefficient is dB / m or Np / m. In this embodiment, Np / m is used as the unit of measurement to facilitate subsequent numerical calculations and distance calculations.

[0042] The ultrasonic velocity and dominant frequency attenuation coefficient are collected synchronously at equal time intervals, forming a two-dimensional joint feature value sequence that varies over time. The selection of the equal time interval should balance the real-time capture of feature changes and the rationality of data storage; a range of 0.1 to 2 seconds is recommended. For conventional pumping conditions of geopolymer concrete, a collection time interval of 0.5 seconds is preferred. This value can reflect the evolution of the internal structure of the slurry in a timely manner on a time scale of several seconds to tens of seconds without generating excessive data redundancy.

[0043] In this embodiment, the healthy voiceprint baseline is constructed as follows: Pre-pumping tests were conducted at different temperatures on geopolymer concrete samples with the same mix proportions as those used at the construction site. During the period when the pumping pressure was within the preset steady-state fluctuation range and the sample entropy of the two-dimensional joint feature value sequence was lower than the preset stability threshold, the two-dimensional joint feature value sequence was extracted as the baseline sample for the pre-test.

[0044] The pre-pumping test is conducted before formal construction, using geopolymer concrete slurry with the exact same mix proportions as the construction site. The setup for the pre-pumping test should be consistent with the actual pumping pipeline system, and should at least include pumping pipes of the same specifications, pump type, and measuring pipe sections as those on site.

[0045] The selected test temperature range should cover the combined variations of ambient temperature and slurry outlet temperature that may occur at the construction site. For example, if the construction season spans from spring to autumn, the ambient temperature may be between 5℃ and 40℃, and the slurry outlet temperature may be between 15℃ and 35℃. In this case, the pre-pumping test should be conducted within the temperature range of 5℃ to 40℃, with at least 8 test temperature points set in temperature steps of no more than 5℃. At each test temperature point, the slurry temperature can be controlled by a constant temperature water bath or the test temperature conditions can be naturally formed by utilizing the ambient temperature at different times.

[0046] At each test temperature point, the pumping system was started and gradually adjusted to the normal operating flow rate, and the pumping pressure and the two-dimensional joint characteristic value sequence were continuously monitored. The pumping pressure being within the preset steady-state fluctuation range means that, for at least 60 consecutive seconds, the difference between the maximum and minimum measured pumping pressure values ​​does not exceed ±10% of the average pumping pressure during that period. The ±10% fluctuation threshold was selected based on the following: Under normal pumping conditions, due to the inherent pulsating characteristics of piston pumps, the pumping pressure exhibits periodic fluctuations around the average value, the amplitude of which typically does not exceed 10% of the average pressure; if this range is exceeded, it indicates possible local blockage or uneven material supply, and the samples collected in this case cannot represent a truly stable flow state.

[0047] The sample entropy of the two-dimensional joint eigenvalue sequence being lower than a preset stability threshold means that, within a time period satisfying the aforementioned pressure steady-state conditions, several consecutive segments of the two-dimensional joint eigenvalue sequence are extracted using a sliding window method, and the sample entropy of each segment is calculated. When the sample entropy of a certain segment is lower than the preset stability threshold, it is determined that the flow state corresponding to that segment has reached sufficient homogeneity. The preset stability threshold is preferably set to 0.2.

[0048] The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the dominant frequency attenuation coefficient sequence in the pre-test baseline sample are calculated respectively to form a pre-set multidimensional statistical feature vector. The mean reflects the central trend of the sound velocity and attenuation coefficient under normal pumping conditions; the standard deviation reflects the natural fluctuation amplitude of the sound velocity and attenuation coefficient under normal pumping conditions; and the sample entropy reflects the internal structural complexity of the sound velocity sequence and attenuation coefficient sequence under normal pumping conditions.

[0049] Simultaneously, the original curve shape of the two-dimensional joint eigenvalues ​​of the pre-test baseline samples changing over time is recorded as a preset waveform template; the preset waveform template is a complete record of the time-domain shape of the original curve of the two-dimensional joint eigenvalues ​​in the pre-test baseline samples. Specifically, this template saves the time-domain shape of the curve within the intercepted time period. The sequence is a numerical sequence arranged by time index to characterize the microscopic dynamic trajectory of the evolution of ultrasonic velocity and dominant frequency attenuation coefficient over time under normal pumping conditions; among which, This represents the ultrasonic velocity value calculated at the t-th sampling time. Represents the sampling time t, for the main frequency The calculated dominant frequency attenuation coefficient value is given, where t is the sampling time index. The complete time-series information of the waveform morphology is preserved, rather than just the statistics; this is because some early slurry instability processes may not immediately manifest as significant shifts in the mean and variance, but rather as changes in subtle fluctuation patterns in the time-series waveform, such as the appearance of periodic oscillations or changes in local trends. By comparing the waveform with a template (as described below with dynamic time warping distance), these early anomalies can be captured more sensitively.

[0050] The pre-set multidimensional statistical feature vectors corresponding to each temperature point T A pre-set healthy voiceprint baseline library is constructed by associating and storing the baselines with preset waveform templates. This baseline library can be implemented using a lookup table, with temperature as the index key and the storage addresses of the statistical feature vector and waveform template as the index values. For operating temperatures between two adjacent test temperatures, the baseline library can generate a reference baseline for the corresponding temperature using linear interpolation during actual retrieval, avoiding temperature coverage gaps caused by insufficiently fine test temperature steps. The interpolation objects are the components in the preset multidimensional statistical feature vector, and the interpolation formula is: ,in and These are the two adjacent test temperatures closest to T. This represents the preset multidimensional statistical feature vector corresponding to the (k+1)th temperature point. This represents the preset multidimensional statistical feature vector corresponding to the k-th temperature point, where k represents the temperature index.

[0051] During actual pumping, when pumping begins and the online calibration triggering conditions are met, a two-dimensional joint feature value sequence with a duration of the first preset duration is extracted as an on-site calibration sample. The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the main frequency attenuation coefficient sequence in the on-site calibration sample are calculated respectively to form a multi-dimensional statistical feature vector on-site. At the same time, the original curve shape of the two-dimensional joint feature value of the on-site calibration sample over time is recorded as an on-site waveform template.

[0052] The online calibration triggering conditions must simultaneously meet the following two conditions: Condition 1: Within a continuous first preset time period, the fluctuation range of the pumping pressure does not exceed a preset pressure fluctuation threshold. The fluctuation range is calculated as follows: Where P is the pumping pressure sampling sequence within the first preset time period, The mean of the pumping pressure sampling sequence. This is a preset relative threshold for pressure fluctuation. The preferred value is 10%, and the pressure fluctuation caused by normal pump pulsation generally does not exceed 10% of the average value; exceeding this value means that the system has not yet entered a steady state or there is an anomaly.

[0053] Condition 2: The sample entropy of the on-site calibration sample does not exceed a preset multiple threshold value of the sample entropy of the preset multidimensional statistical feature vector at the corresponding temperature retrieved from the preset healthy acoustic baseline library. The preset multiple threshold value is between 1.5 and 2, preferably 1.8. The preset healthy baseline is a sample entropy level characterizing ideal stable flow obtained under strictly controlled test conditions. However, on-site working conditions inevitably have slight fluctuations in mix proportions and environmental noise interference. Therefore, the on-site sample entropy is allowed to be appropriately higher than the preset value; however, if it exceeds 1.8 times, it indicates that there is a significant difference in structural complexity between the on-site slurry flow state and the ideal state. If calibration is forced at this time, some early instability characteristics will be included in the baseline, reducing the sensitivity of subsequent detection. The value of 1.8 times was determined after comparative analysis of the sample entropy of multiple sets of normal pumping and critical instability pumping data of geopolymer concrete. At this multiple, the boundary between normal fluctuations and early instability can be better distinguished.

[0054] The value of the first preset duration needs to balance the representativeness of the calibration sample and the timeliness of calibration triggering: if it is too short, the sample statistic estimation will be unrobust; if it is too long, the calibration opportunity may be missed (i.e., calibration should be completed within the time window when the pumping state still meets the calibration conditions). Based on the statistical analysis of the stable duration of the pumping state during a typical polymer concrete pumping cycle, the preferred value range for the first preset duration is 30 seconds to 180 seconds.

[0055] A weighted average method is used to fuse the field multidimensional statistical feature vector with the preset multidimensional statistical feature vector at the corresponding temperature to generate a calibrated multidimensional statistical feature vector; the specific calculation formula is as follows: in, This refers to the calibrated multidimensional statistical feature vector generated by weighted averaging and fusing the on-site multidimensional statistical feature vector with the preset multidimensional statistical feature vector at the corresponding temperature when the real-time temperature of the slurry is T. This represents the field multidimensional statistical feature vector calculated from the extracted field calibration samples at a real-time slurry temperature of T. represents the weighting coefficient of the field data, with a value range of (0,1); represents the preset multidimensional statistical feature vector at the corresponding temperature T, obtained by retrieving or interpolating from the preset health voiceprint baseline library; the operator + represents the addition of each corresponding component of the vector.

[0056] Calibrated fluctuation template The generation method is as follows: in, This represents the original curve shape of the two-dimensional joint eigenvalues ​​recorded from the field calibration sample over time at a real-time slurry temperature of T, i.e., the field waveform template. This refers to a preset waveform template corresponding to the current real-time temperature T of the slurry, which is retrieved or interpolated from a preset healthy acoustic baseline library; the weighted average of the waveform template refers to the point-by-point numerical average of two time series at corresponding time sampling points.

[0057] The weighted average is achieved by averaging the values ​​of two time series of equal duration at corresponding time sampling points, with the field data weighting coefficients... The preferred value range is 0.3 to 0.7. The value of λ directly affects the balance between the speed at which the calibrated baseline adapts to the actual working conditions on site and the degree to which it retains the pre-set prior knowledge: the larger the value of λ, the more the baseline tends to trust the data collected on site, and the calibrated baseline can adapt to the particularities of the actual working conditions more quickly, but the risk of being affected by accidental interference on site also increases. The smaller the value, the more the baseline tends to retain prior knowledge that has been fully verified in the pre-experiment, resulting in greater robustness, but a reduced ability to adapt to the specificities of the field.

[0058] Considering that the pre-pumping test used slurry with the same mix proportions as the construction site and underwent thorough testing at multiple temperature points, the pre-established baseline already possesses high reference value. Simultaneously, minor mix proportion drift and air bubble content differences that may occur during the mixing and transportation of the slurry on-site also need to be corrected through on-site calibration. In this embodiment... The preferred value is 0.4, which assigns 40% weight to the field data and 60% weight to the preset data. This ensures that the calibrated baseline can quickly reflect the actual field conditions while maintaining sufficient robustness to avoid excessive baseline shift due to accidental interference.

[0059] The ultrasonic velocity and dominant frequency attenuation coefficient of geopolymer concrete slurry are both temperature-dependent; the bulk modulus and viscosity of water in the slurry decrease with increasing temperature, leading to a decrease in sound velocity; simultaneously, changes in the hydration reaction rate within the slurry also affect the microstructure, thus influencing the acoustic attenuation characteristics. To achieve a unified representation of the healthy baseline under different temperature conditions, all characteristic components need to be normalized to the same preset reference temperature.

[0060] Each feature component in the calibrated multidimensional statistical feature vector is mapped to its equivalent value at a preset reference temperature according to a preset temperature normalization function to obtain a temperature-normalized health baseline multidimensional statistical feature vector; the temperature-normalized health baseline multidimensional statistical feature vector and the calibrated waveform template are jointly stored as a health voiceprint baseline.

[0061] The preset temperature normalization function is established as follows: In a geopolymer concrete sample with the same mix proportions as the construction site, the ultrasonic velocity and dominant frequency attenuation coefficient of the slurry under stable flow conditions were measured as a function of temperature in a temperature-controlled laboratory environment. Specifically, normal, stable-flowing geopolymer concrete slurry with the same mix proportions was placed in a temperature-controlled testing device. The temperature was changed in increments of 2°C within a temperature range of 5°C to 45°C, and after stabilizing at each temperature point, the ultrasonic velocity was measured and recorded. and the main frequency attenuation coefficient To obtain discrete data pairs and ,in This indicates a specific temperature value (e.g., 25℃).

[0062] Polynomial fitting (preferably second- or third-order polynomials) is performed on the above data points to establish the following temperature mapping model: Where T is the real-time temperature of the slurry. These are the sound speed temperature fitting coefficients. This represents the attenuation coefficient minus the temperature fitting coefficient. Represents discrete data points The established function model, namely the expected ultrasonic velocity at temperature T, Represents discrete data points The established function model is the expected frequency attenuation coefficient at temperature T.

[0063] Based on this model, characteristic values ​​measured at any temperature T can be mapped to equivalent values ​​at a preset reference temperature; the preset temperature normalization function is defined as: in, These are the ultrasonic velocity and the dominant frequency attenuation coefficient, respectively, measured at temperature T. These are the temperature-normalized ultrasonic velocity and the dominant frequency attenuation coefficient mapped to the reference temperature, respectively. This indicates that the preset reference temperature value will be used. Substitute into the sound speed-temperature fitting function In the calculation, the desired speed of sound at the reference temperature is obtained; This indicates that the preset reference temperature value will be used. Substitute the attenuation coefficient into the temperature fitting function The expected attenuation coefficient at the reference temperature is calculated.

[0064] Preset reference temperature The principle for selecting the value is: choose the median value of the pre-pumping test temperature range or the temperature value with the highest expected frequency at the construction site. In this embodiment, the reference temperature... Set the temperature to 20℃.

[0065] Once the temperature normalization function and its fitting coefficients are established in step 2, they can be directly reused in the real-time temperature compensation in the subsequent step 3 without repeated calibration. For the standard deviation and sample entropy components, the same normalization function as the mean component is used for temperature compensation, and its fitting coefficients are also determined by the calibration data mentioned above.

[0066] The temperature-normalized healthy baseline multidimensional statistical feature vector is stored together with the calibrated waveform template to form a complete healthy acoustic waveform baseline. This healthy acoustic waveform baseline will serve as a reference benchmark in subsequent step 3 to assess the degree of deviation of the real-time flow state of the slurry from the normal range.

[0067] Step 3: During continuous pumping, continuously extract the two-dimensional joint feature value sequence and calculate the real-time fluctuation characteristics; use the real-time temperature of the slurry to perform temperature compensation on the real-time fluctuation characteristics and the healthy acoustic baseline, and continuously compare the temperature-compensated real-time fluctuation characteristics with the temperature-compensated healthy acoustic baseline to obtain the instability evolution index that reflects the degree of deviation of the current internal structure of the slurry from the stable state.

[0068] In this embodiment, the real-time fluctuation characteristics are calculated as follows: Using the current sampling time as the endpoint, a two-dimensional joint feature value sequence with a duration of the second preset duration is extracted to form a real-time observation window.

[0069] The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the dominant frequency attenuation coefficient sequence within the real-time observation window are calculated to form a real-time multidimensional statistical feature vector, which serves as the real-time fluctuation feature.

[0070] The second preset duration should not exceed the first preset duration. Its specific value should be determined through preliminary experiments, taking into account both the response speed to changes in pumping status and statistical stability. If the second preset duration is too long, it will reduce the sensitivity to detecting instability time; if the second preset duration is too short, the sample size within the window will be insufficient to obtain statistically stable mean and standard deviation. Typically, the preferred range for the second preset duration is 10 to 60 seconds. For example, a first preset duration of 120 seconds and a second preset duration of 60 seconds; such a value ensures a minute-level response to status changes while also ensuring sufficient data points within the window (e.g., if the sampling frequency is 10Hz, the window contains 300 sampling points), making the calculation results of statistical characteristics stable and reliable.

[0071] To eliminate the influence of temperature changes on the dominant frequency attenuation coefficient of ultrasonic velocity and ensure the comparability of instability evolution indices calculated under different temperature conditions, this embodiment performs temperature normalization processing on ultrasonic velocity and dominant frequency attenuation coefficient within the real-time observation window.

[0072] The preset temperature normalization function and its fitting coefficients used in the temperature normalization process are exactly the same as those described in step 2, and are directly reused here. Specifically, the sound speed-temperature fitting function established in step 2 is used. and attenuation coefficient-temperature fitting function and preset reference temperature The ultrasonic velocity at each sampling moment within the real-time observation window and the main frequency attenuation coefficient Mapping is performed using the following formula: By using the above additive mapping method, the fluctuation characteristics of the original measurements (such as variance and sample entropy) can be kept basically unchanged, and only the mean is shifted, so that the health baseline and real-time characteristics are statistically comparable.

[0073] In this embodiment, the instability evolution index is calculated as follows: Calculate the deviation distance between the real-time multidimensional statistical feature vector and the temperature-normalized multidimensional statistical feature vector of the healthy baseline in the healthy voiceprint baseline. This deviation distance is either Euclidean distance or Mahalanobis distance. Taking Mahalanobis distance as an example, the calculation formula is as follows: in, express and Mahalanobis distance between them This represents the real-time six-dimensional statistical feature vector after temperature normalization. A six-dimensional statistical eigenvector representing the health baseline. It is the inverse matrix of the six-dimensional feature vector covariance matrix of the healthy baseline sample. This covariance matrix S is pre-calculated and stored using a large amount of data from the healthy baseline sample when constructing the healthy voiceprint baseline in step 2. It reflects the degree of variation and interrelation of each feature component under normal conditions.

[0074] Mahalanobis distance transforms anisotropic and scale-asymmetric feature spaces into standard Euclidean space through inverse matrix whitening of the covariance matrix. For example, if the mean velocity of sound fluctuates very little under normal conditions (small variance), even a small deviation will significantly contribute to the Mahalanobis distance, thereby increasing the weight of this sensitive feature in the instability index.

[0075] Mapping the deviation distance to the [0,1] interval yields the statistical instability component. The mapping function is a sigmoid function, as shown in the following equation: in, The sensitivity adjustment coefficient is a positive real number. The specific value of this parameter should be determined through preliminary experiments. Specifically, in the pre-pumping test of a geopolymer concrete slurry sample with the same mix proportion as the construction site, different degrees of instability are artificially created (e.g., by changing the water-cement ratio, introducing air bubbles, etc.), and the instability evolution index of the transition stage from normal to slight instability is recorded. The target is: when the slurry is just determined by experienced on-site operators to have begun to exhibit abnormal fluidity through visual inspection or auxiliary means, the expected statistical instability component at this point is... To reach a preset value, such as 0.3. This can be achieved by adjusting... The value of is such that the calculated value under this critical state is The above formula approximates 0.3 after mapping. For example, if the mean Mahalanobis distance in this critical state is 2.5, then by calculating... It can be determined The rationale for using an exponential mapping is that it provides a smooth, continuous transformation with an output range strictly between [0,1], which can well reflect the nonlinear cumulative process of instability from quantitative to qualitative change. Here, 0 indicates that the slurry is in a stable flow state completely consistent with the healthy baseline, and 1 indicates that the slurry has experienced the most severe degree of instability.

[0076] To capture early, subtle changes in the morphology of slurry flow patterns over time, this embodiment simultaneously calculates the dynamic time warping distance between the two-dimensional joint eigenvalue sequence within the real-time observation window and the calibrated waveform template in the healthy acoustic baseline.

[0077] Real-time observation window The normalized internal temperature ultrasonic velocity sequence and the dominant frequency decay sequence constitute a two-dimensional time series. , This represents the normalized speed of sound at point 1 of the real-time sequence. This represents the normalized speed of sound at point n in the real-time sequence. This represents the normalized decay coefficient at point 1 of the real-time sequence. This represents the normalized attenuation coefficient at point n of the real-time sequence; the calibrated waveform template stored in the healthy voiceprint baseline is a two-dimensional sequence with the same structure. , These represent the ultrasonic velocity values ​​at the 1st and mth sampling points, respectively, within the calibrated waveform template stored in the healthy voiceprint baseline. These represent the main frequency attenuation coefficient values ​​of the 1st and mth sampling points in the calibrated waveform template stored in the healthy voiceprint baseline, respectively; where n is the total number of sampling points contained in the real-time observation window, and m is the total number of sampling points in the template.

[0078] Calculation using a multidimensional dynamic time warping algorithm and Minimum cumulative distance between This algorithm allows for local stretching or compression of sequences along the time axis, thereby achieving optimal alignment.

[0079] The dynamic time warping algorithm finds an optimal warping path. ,in , , This minimizes the cumulative distance along this path; They represent, Let r and k represent the r-th and k-th elements on the normalized path P, respectively, where r represents the element index and k represents the number of steps or grid points contained in the optimal normalized path P. These represent the first and second digits of the regularized path. Step, corresponding real-time sequence Index and baseline template sequence The index.

[0080] The cumulative distance matrix D is calculated iteratively using dynamic programming: For i>1 or j>1: in, The initial value for the cumulative distance matrix is ​​set directly at the starting point. Because the only path to the starting point is itself, there is no historical cost. This represents the local Euclidean distance between the starting points of two sequences; Starting from the starting point (1,1), find the minimum cumulative distance to grid point (i,j). For the first real-time sequence The element and the baseline template sequence in the element Euclidean distance between elements , Let represent the minimum cumulative distances from the starting point (1,1) to grid points (i-1,j), (i,j-1), and (i-1,j-1), respectively. Let 'i' be used to index and traverse the two-dimensional time series within the real-time observation window. The data point j is used to index and traverse the calibrated waveform template sequence stored in the healthy voiceprint baseline. Data points in the data.

[0081] Final dynamic time-warped distance The last element of the cumulative distance matrix Divide by the normalized path length K to normalize the path length, as shown in the following formula: Where K represents the number of grid points traversed by the optimal warping path P. The reason for using dynamic time warping distance is that pressure fluctuations and slight changes in flow rate during pumping can cause subtle distortions or shifts in the signal along the time axis. Directly comparing points using traditional Euclidean distance might misinterpret this temporal misalignment as a difference in signal morphology, leading to false alarms. DTW distance, through flexible matching, can more accurately identify the actual waveform changes caused by alterations in the internal structure of the slurry.

[0082] Finally, the dynamic time-normalized distance is... It is also mapped to the [0,1] interval to obtain the morphological instability component. The mapping also uses a sigmoid function, as shown in the following equation: in, This is the morphological sensitivity adjustment coefficient, and its determination method is similar to that in the statistical instability component. Similarly, pre-test calibration is also needed to ensure that when observable flow patterns begin to become turbulent, To achieve the desired value. For example, it can be taken as... The same calibration target value of 0.3 is used for inverse calculation.

[0083] For the statistically unstable components and morphological instability components Weighted fusion is performed to obtain the final instability evolution index. As shown in the following formula: in, and These are the statistical instability weights and the morphological instability weights, respectively. .

[0084] The specific values ​​of the weights should be determined using receiver operating characteristic (ROC) curve analysis. In the pre-pumping trial, for the same batch of labeled data, only... and only use They were tested as early warning indicators, and their sensitivity and false alarm rate in detecting early instability were recorded.

[0085] The objective was to minimize false alarms while maintaining high sensitivity. ROC curves for the two metrics were plotted at different thresholds, and their area under the curve and performance at specific operating points (e.g., detection rate when a false alarm rate of <5% was required) were compared.

[0086] If the test results show that, based on statistical distribution It can more stably reflect the continuous deterioration of the overall state, while based on waveform morphology It is more sensitive to sudden, short-lived anomalies, but may have a higher false alarm rate. Therefore, a more stable weighting scheme can be set, for example... , If practice proves that the contributions of the two are equal, then the arithmetic mean should be used, i.e. .

[0087] Step 4: Compare the instability evolution index with the preset two-level warning thresholds to determine the current pumpability status of the pumping pipeline.

[0088] In this embodiment, the two-level warning threshold includes a first warning threshold. Second warning threshold ,and The constraint relationship; both the first and second warning thresholds are represented by the instability evolution index; the first warning threshold is used to distinguish between the normal state and the critical instability state, and the second warning threshold is used to distinguish between the critical instability state and the impending pipe blockage state.

[0089] Based on the instability evolution index continuously calculated in step 3, this embodiment divides the pumpability state of the pumping pipeline into three levels: normal, critical instability, and impending blockage; the specific determination logic is as follows: If the instability evolution index is less than the first warning threshold, the current pumpability status is determined to be normal. In this state, the internal structure of the geopolymer concrete slurry remains stable, the real-time fluctuation characteristics of ultrasonic velocity and dominant frequency attenuation coefficient are highly consistent with the healthy acoustic waveform baseline, the pumping process is smooth, no intervention measures are required, and the system continues to maintain online monitoring.

[0090] The typical instability evolution index value range for the normal state is approximately 0-0.3. At this time, the dynamic time warping distance between the two-dimensional joint feature value sequence in the real-time observation window and the calibrated waveform template in the healthy acoustic baseline is generally no more than 0.15, and the deviation distance (Euclidean distance or Mahalanobis distance) is generally no more than 0.2.

[0091] If the instability evolution index is not less than the first warning threshold and less than the second warning threshold, the current pumpability state is determined to be critically unstable, triggering a warning signal. In this state, detectable signs of instability have begun to appear in the internal structure of the slurry, such as localized agglomeration of solid particles, increased gas content, or decreased slurry homogeneity, but it has not yet developed to the point of clogging the pipeline. The warning signal can be a visual warning signal (such as a yellow warning label on the monitoring interface), an auditory warning signal (such as an intermittent buzzer), or a combination of both, to remind on-site operators to closely monitor the pumping status and to take preventative control measures.

[0092] The preventive control measures include, but are not limited to: appropriately reducing the pumping flow rate to reduce the shear rate in the pipeline, checking and adjusting the addition rate of admixtures (such as water-reducing agents and retarders), and supplementing the stirring of the slurry in the hopper to restore homogeneity; the typical instability evolution index value corresponding to the critical instability state is approximately 0.35-0.6.

[0093] In practical engineering applications, there is usually a considerable transition time window from the first detection of a critical instability state to the development of an impending pipe blockage state. This time window is generally no less than 3 minutes, which is sufficient for on-site operators to implement the aforementioned preventive control measures.

[0094] If the instability evolution index is not less than the second warning threshold, the current pumpability state is determined to be impending pipe blockage, triggering a mandatory intervention signal. In this state, the internal structure of the slurry has become severely unstable, and the flow resistance of the slurry within the pipe increases sharply. If mandatory intervention measures are not taken immediately, a complete pipe blockage will occur within a very short time. The mandatory intervention signal can be an audible alarm signal (such as a continuous alarm sound), a visual alarm signal (such as a flashing red indicator on the monitoring interface), or an electrical signal that directly outputs an automatic intervention command to the pumping control system.

[0095] The mandatory intervention measures include, but are not limited to: immediately initiating reverse pumping operation of the pumping pipeline to loosen the blockage precursor, automatically increasing the pumping pressure to overcome the local resistance peak, and urgently stopping pumping to await manual investigation and handling; that is, the typical instability evolution index value corresponding to the blockage state is approximately 0.65-1.

[0096] In this embodiment, the specific values ​​of the first and second warning thresholds are not arbitrarily selected, but are systematically determined by conducting pre-pumping tests on geopolymer concrete slurry samples with the same mix proportions as those at the construction site, combined with receiver operating characteristic (ROC) curve analysis. The specific implementation steps of this determination method are as follows: Before formal construction, a pre-pumping test was conducted using a geopolymer concrete slurry sample with the exact same mix ratio as the actual construction site, on a pumping pipeline system identical or equivalent to the actual construction system. The pre-pumping test began with normal pumping and was continuously monitored until the slurry completely blocked the pipeline. Throughout the pre-pumping test, all operations from steps 1 to 3 were performed simultaneously, continuously recording a two-dimensional joint eigenvalue sequence (including ultrasonic velocity sequence and dominant frequency attenuation coefficient sequence) and the real-time temperature of the slurry, forming pre-pumping test characteristic data. The time span of the pre-pumping test characteristic data covers the complete evolution process from stable pumping, initial instability, significant instability to complete pipe blockage.

[0097] The pre-pumping test characteristic data was labeled frame by frame, and a corresponding pumpability state label was assigned to the two-dimensional joint feature value of each frame (i.e., each sampling time). The pumpability state labels included at least three categories: label 0 indicated the normal state, label 1 indicated the critical instability state, and label 2 indicated the impending pipe blockage state. The labeling was based on the state judgment records made by professional technicians during the pre-pumping test according to the actual pumping performance (such as pumping pressure change trends, pipeline vibration, and slurry morphology at the discharge port), as well as the accurate time of pipe blockage occurrence. The labeling interval for the impending pipe blockage state was defined as the time period extending backward from the moment of complete pipe blockage to the moment when the pumping pressure began to show an irreversible and rapid upward trend.

[0098] Based on the recorded pre-pumping test characteristic data, the Instability Evolution Index (IEI) corresponding to each sampling time is calculated frame by frame according to the method described in step 3, forming an Instability Evolution Index sample set. Each IEI value in this sample set corresponds to a known pumpability state label.

[0099] Using the pumpability state label as the classification truth and the instability evolution index sample set as the continuous input variable of the classifier, receiver operating characteristic (ROC) curve analysis was employed to determine... and The optimal value. Specifically: Determine the first warning threshold Label 0 (normal state) is defined as the negative class, and labels 1 (critical instability state) and 2 (imminent blockage state) are combined and defined as the positive class. As a variable binary classification threshold, i.e., when IEI ≥ When the IEI is less than 1, it is considered positive (unstable). If the result is negative (normal), it is determined by traversing the range [0,1] with a preset step size (e.g., 0.01). Calculate each of the following possible values: Plot the ROC curve for the true positive rate (TPR, i.e., the proportion of correct warnings of instability) and false positive rate (FPR, i.e., the proportion of normal cases mistakenly reported as instability) under the given values.

[0100] The formula for calculating the true positive rate (TPR) is: The formula for calculating the false positive rate is: Among them, TP (true positive) is a case where the actual state is unstable (label 1 or 2) and is correctly identified as positive (IEI ≥ ). The sample tree; FN (false negative) is a case where the actual state is unstable but is incorrectly judged as negative (IEI < 0.05). The number of samples; FP (false positive) refers to samples that are actually normal (label 0) but are incorrectly identified as positive (IEI ≥ 0). The sample size; TN (true negative) is the number of samples that are actually normal but correctly identified as negative (IEI < 0.05). The number of samples.

[0101] Determine the second warning threshold Label 0 (normal state) and Label 1 (critical instability state) are combined and defined as the negative class, while Label 2 (imminent blockage state) is defined as the positive class. As a variable binary classification threshold, it traverses the range [0,1] with a preset step size. Calculate each of the following possible values: Plot the ROC curve for the true positive rate and false positive rate under the given values.

[0102] for and In this embodiment, the selection of parameters aims to maximize the early warning time for pipe blockage events while minimizing the false alarm rate. Here, the early warning time refers to the time interval from the moment the system first issues a warning signal of the corresponding level (critical instability indication or impending pipe blockage alarm) to the moment when complete pipe blockage actually occurs; the false alarm rate is quantitatively measured by the false positive rate (FPR).

[0103] The specific threshold selection criteria are as follows: For On the ROC curve, select the IEI value that maximizes TPR under the constraint that FPR ≤ preset upper limit of allowable false alarm rate (e.g., 0.05, i.e., 5%). The optimal value; for On the ROC curve, select the IEI value that maximizes TPR under the constraint that FPR ≤ preset upper limit of allowable false alarm rate (e.g., 0.02, i.e., 2%). The optimal value. The selection basis for the above-mentioned preset allowable upper limit of false alarm rate is: for the first warning threshold Because it triggers a warning signal rather than mandatory intervention, a relatively high false alarm rate is allowed to avoid missing early signs of instability; for the second warning threshold Since it triggers a forced intervention signal, which may have a direct impact on the construction process, a lower false alarm rate is required to ensure the high reliability of the alarm.

[0104] It should be noted that in the above label merging method, the first warning threshold is determined. The critical instability state (label 1) and the impending pipe blockage state (label 2) are merged into a positive category. The purpose of this is to make... It can sensitively detect various abnormal states, including early instability; and determine the second warning threshold. At that time, the critical instability state (label 1) and the normal state (label 0) are merged into a negative class. The purpose of this is to ensure... Forced intervention signals are only triggered when the slurry has entered a high-risk stage where pipe blockage is imminent, to avoid accidental activation during critical instability. This could lead to unnecessary forced shutdowns. If, during actual verification, accidental activation is found during critical instability... The ratio exceeds the acceptable range (e.g., in critically unstable samples) (If the proportion of samples misjudged as being about to become blocked exceeds 5%), further adjustments can be made to the label 1 sample from... Excluded from ROC analysis, and re-determined only using label 0 as the negative class and label 2 as the positive class. The optimal value.

[0105] The results determined by the above ROC analysis and The optimal value should be applied in at least one independent verification pre-pumping test to verify its actual early warning performance. If, during the verification test, the early warning time from the system's first issuance of the impending pipe blockage alarm to the actual occurrence of complete pipe blockage is less than the preset safety margin (e.g., not less than 120 seconds), or the duration of the critical instability state is insufficient for operators to perform preventative adjustments (e.g., less than 180 seconds), then the value should be appropriately reduced. and / or The value of is determined in advance to provide early warnings until the engineering safety requirements are met.

[0106] In this embodiment, in a simulated pumping test with the same mix ratio as the construction site, ultrasonic velocity and dominant frequency attenuation coefficient were continuously collected at equal time intervals (sampling interval of 5 seconds) from the start of normal pumping, through the critical instability stage, until the moment of near-complete pipe blockage; some representative time-series data are shown in the table below: Table 1. Example of two-dimensional joint eigenvalue sequence for the pumping process of geopolymer concrete: According to Table 1 above and Figure 2 It can be observed that the first stage (0–55 seconds) is the normal pumping stabilization period. Within this range, the ultrasonic velocity fluctuates randomly within a narrow band of 1849.8 m / s to 1852.3 m / s, and the dominant frequency attenuation coefficient exhibits similar stable random fluctuations between 12.44 Np / m and 12.56 Np / m. These fluctuations are inherent natural fluctuations caused by the random rearrangement of solid particles within the slurry and slight changes in local shear rates under normal pumping conditions. Their fluctuation amplitude and sequence complexity are both low, satisfying the healthy baseline construction conditions of sample entropy below the preset stabilization threshold and pumping pressure within the preset steady-state fluctuation range.

[0107] The second stage (60–105 seconds): the critical instability evolution period. From 60 seconds onwards, the ultrasonic velocity begins a monotonically decreasing trend from 1849 m / s, dropping to 1822.8 m / s by 105 seconds; simultaneously, the dominant frequency attenuation coefficient monotonically increases from 12.6 Np / m to 14.68 Np / m. During this stage, the rates (slopes) of change for both are relatively gentle but directional, forming a divergent characteristic of decreasing sound velocity and increasing attenuation. This indicates that the alkali-activated reaction within the geopolymer concrete has led to the over-construction of the cementitious network structure, a gradual increase in the slurry yield stress and plastic viscosity, and a continuous rise in flow resistance. Since this stage is still in the early stages of instability, although the macroscopic fluidity of the slurry has undergone detectable changes, it has not yet reached the level that would trigger a sharp fluctuation in pipeline pressure. Therefore, conventional pressure monitoring methods cannot provide effective early warning signals during this stage. The successful capture at this stage is the core advantage of this embodiment compared to the pipeline pressure monitoring method described in the background art. That is, this embodiment can output an early identification signal of critical instability during the evolution of the internal structure instability of the slurry, rather than after the risk of pipe blockage has been formed.

[0108] The third stage (110–145 seconds): the impending blockage period. From 110 seconds onwards, the rates of change of ultrasonic velocity and the dominant frequency attenuation coefficient undergo significant abrupt changes: the rate of decrease in sound velocity and the rate of increase in the attenuation coefficient both accelerate simultaneously. By 145 seconds, the ultrasonic velocity has dropped to 1794 m / s, deviating by 57 m / s from the healthy baseline average (approximately 1851 m / s); the dominant frequency attenuation coefficient rises to 18.8 Np / m, deviating by 6.3 Np / m from the healthy baseline average (approximately 12.5 Np / m). During this stage, the internal cementitious network of the slurry is highly cross-linked, and the slurry is close to losing its flowability. The pressure inside the pipe will rise sharply. If no immediate forced intervention is taken, complete blockage will occur within a very short time.

[0109] The data in Table 1 above and Figure 2 The combined trend of the dual-feature reverse divergence fully verifies that the two-dimensional joint feature value sequence constructed in this embodiment can continuously, sensitively, and stably track the dynamic process of geopolymer concrete slurry from stable flow to structural instability.

[0110] Please see Figure 3 The present invention also provides a system for analyzing the pumpability of geopolymer concrete, which is used to implement the above-mentioned method for analyzing the pumpability of geopolymer concrete, comprising: The multi-parameter in-situ acquisition module is used to install a measuring pipe section in the pumping pipeline. The measuring pipe section is equipped with an acoustic window. Through an ultrasonic transducer installed on the outside of the acoustic window, instantaneous pulse ultrasonic waves are emitted into the polymer concrete slurry inside the pipe using the pulse transmission method, and the transmitted signal after passing through the slurry is received. At the same time, the real-time temperature of the slurry in the measuring pipe section is acquired. The slurry structure quantification module is used to extract the ultrasonic velocity and the dominant frequency attenuation coefficient from the transmitted signal under pumping conditions, forming a two-dimensional joint feature value sequence that varies with time. In the initial stage of pumping, a segment of the two-dimensional joint feature value sequence is extracted as a healthy baseline sample under normal pumping conditions. The statistical characteristics of this healthy baseline sample are calculated to construct a healthy acoustic waveform baseline that characterizes the stable flow state of the slurry. The pressure increase prediction module is used to continuously extract the two-dimensional joint feature value sequence during continuous pumping and calculate the real-time fluctuation characteristics. It uses the real-time temperature of the slurry to perform temperature compensation on the real-time fluctuation characteristics and the healthy acoustic waveform baseline, and continuously compares the temperature-compensated real-time fluctuation characteristics with the temperature-compensated healthy acoustic waveform baseline to obtain the instability evolution index that reflects the degree of deviation of the current internal structure of the slurry from the stable state. The pumpability analysis and determination module is used to compare the instability evolution index with the preset two-level warning thresholds to determine the current pumpability status of the pumping pipeline.

[0111] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0112] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0113] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0114] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for analyzing the pumpability of geopolymer concrete, characterized in that, The specific steps include: Step 1: Install a measuring pipe section in the pumping pipeline. The measuring pipe section is equipped with an acoustic window. Through an ultrasonic transducer installed on the outside of the acoustic window, emit instantaneous pulse ultrasonic waves into the polymer concrete slurry inside the pipe using the pulse transmission method, and receive the transmitted signal after passing through the slurry. At the same time, collect the real-time temperature of the slurry in the measuring pipe section. Step 2: Under pumping conditions, extract the ultrasonic velocity and the dominant frequency attenuation coefficient from the transmitted signal to form a two-dimensional joint feature value sequence that varies with time; in the initial stage of pumping, extract a segment of the two-dimensional joint feature value sequence as a healthy baseline sample under normal pumping conditions, calculate the statistical characteristics of the healthy baseline sample, and construct a healthy acoustic waveform baseline that characterizes the stable flow state of the slurry. Step 3: During continuous pumping, continuously extract the two-dimensional joint feature value sequence and calculate the real-time fluctuation characteristics; use the real-time temperature of the slurry to perform temperature compensation on the real-time fluctuation characteristics and the healthy acoustic waveform baseline, and continuously compare the temperature-compensated real-time fluctuation characteristics with the temperature-compensated healthy acoustic waveform baseline to obtain the instability evolution index that reflects the degree of deviation of the current internal structure of the slurry from the stable state. Step 4: Compare the instability evolution index with the preset two-level warning thresholds to determine the current pumpability status of the pumping pipeline.

2. The method for analyzing the pumpability of geopolymer concrete according to claim 1, characterized in that, The measuring pipe section is a detachable structure, connected to the straight section of the pumping pipeline via a flange. An alkali-resistant sound-transmitting window flush with the inner wall of the pipe is embedded in the inner wall of the measuring pipe section. The ultrasonic transducer is a transceiver integrated longitudinal wave transducer with a center frequency in the range of [50kHz, 500kHz]. It is fixed to the outer surface of the acoustic window through a coupling agent layer to emit instantaneous pulsed ultrasonic waves perpendicular to the slurry flow direction and receive transmitted signals.

3. The method for analyzing the pumpability of geopolymer concrete according to claim 1, characterized in that, The ultrasonic velocity is calculated by the transit time of the transmitted signal and the inner diameter of the measured pipe section. The dominant frequency attenuation coefficient is calculated by the logarithmic attenuation rate of the dominant frequency component of the transmitted signal amplitude spectrum relative to the dominant frequency component of the transmitted signal amplitude spectrum. The ultrasonic velocity and the dominant frequency attenuation coefficient are collected synchronously at equal time intervals to form a two-dimensional joint feature value sequence that varies with time.

4. The method for analyzing the pumpability of geopolymer concrete according to claim 1, characterized in that, The healthy voiceprint baseline is constructed as follows: In the same proportion of geopolymer concrete samples as the construction site, pre-pumping tests were conducted at different temperatures. During the period when the pumping pressure was within the preset steady-state fluctuation range and the sample entropy of the two-dimensional joint feature value sequence was lower than the preset stability threshold, the two-dimensional joint feature value sequence was extracted as the pre-test baseline sample. The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the main frequency attenuation coefficient sequence in the pre-test baseline sample are calculated respectively to form a preset multidimensional statistical feature vector; at the same time, the original curve shape of the two-dimensional joint feature value of the pre-test baseline sample changing with time is recorded as a preset waveform template. The preset waveform template is a time-domain waveform representation of a two-dimensional joint feature value sequence; the preset multidimensional statistical feature vector at the corresponding temperature and the preset waveform template are associated and stored to construct a preset health voiceprint baseline library at different temperatures; In the actual pumping process, when pumping starts and the online calibration triggering condition is met, a two-dimensional joint feature value sequence with a duration of the first preset duration is extracted as an on-site calibration sample. The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the main frequency attenuation coefficient sequence in the field calibration sample are calculated respectively to form a field multidimensional statistical feature vector; at the same time, the original curve shape of the two-dimensional joint feature value of the field calibration sample over time is recorded as a field waveform template. A weighted average method is used to fuse the field multidimensional statistical feature vector with the preset multidimensional statistical feature vector at the corresponding temperature to generate a calibrated multidimensional statistical feature vector. The on-site waveform template is weighted and fused with the preset waveform template at the corresponding temperature to generate a calibrated waveform template; Each feature component in the calibrated multidimensional statistical feature vector is mapped to its equivalent value at a preset reference temperature according to a preset temperature normalization function to obtain the temperature-normalized health baseline multidimensional statistical feature vector. The temperature-normalized health baseline multidimensional statistical feature vector and the calibrated waveform template are jointly stored as the health voiceprint baseline; The online calibration triggering conditions are: the fluctuation amplitude of the pumping pressure within a continuous first preset time period does not exceed the preset pressure fluctuation threshold, and the sample entropy of the on-site multidimensional statistical feature vector does not exceed the preset multiple threshold of the sample entropy of the preset multidimensional statistical feature vector at the corresponding temperature retrieved from the preset healthy voiceprint baseline library.

5. The method for analyzing the pumpability of geopolymer concrete according to claim 4, characterized in that, The calculation method for the real-time fluctuation characteristics is as follows: Using the current sampling time as the endpoint, a two-dimensional joint feature value sequence with a duration of the second preset duration is extracted to form a real-time observation window; The mean, standard deviation, and sample entropy of the ultrasonic velocity sequence and the dominant frequency attenuation coefficient sequence within the real-time observation window are calculated to form a real-time multidimensional statistical feature vector, which serves as the real-time fluctuation feature. The second preset duration is no greater than the first preset duration.

6. The method for analyzing the pumpability of geopolymer concrete according to claim 5, characterized in that, The instability evolution index is calculated as follows: Calculate the deviation distance between the real-time multidimensional statistical feature vector and the temperature-normalized multidimensional statistical feature vector of the healthy baseline in the healthy voiceprint baseline, and map the deviation distance to the [0,1] interval; Simultaneously, the dynamic time-warped distance between the two-dimensional joint eigenvalue sequence within the real-time observation window and the calibrated waveform template in the healthy voiceprint baseline is calculated; The instability evolution index is obtained by weighting and fusing the dynamic time regularization distance and deviation distance.

7. The method for analyzing the pumpability of geopolymer concrete according to claim 6, characterized in that, The temperature compensation specifically includes: Before calculating the deviation distance, each feature component in the real-time multidimensional statistical feature vector corresponding to the real-time observation window is mapped to its equivalent value at a preset reference temperature using the same preset temperature normalization function to obtain a temperature-normalized real-time multidimensional statistical feature vector. The deviation distance is calculated based on the temperature-normalized real-time multidimensional statistical feature vector and the temperature-normalized health baseline multidimensional statistical feature vector.

8. The method for analyzing the pumpability of geopolymer concrete according to claim 1, characterized in that, The two-level warning thresholds include a first warning threshold and a second warning threshold, and the second warning threshold is greater than the first warning threshold; both the first warning threshold and the second warning threshold are represented by an instability evolution index; The pumpability status includes normal, critical instability, and impending blockage. The logic for determining the pumpability status of the pumping pipeline is as follows: If the instability evolution index is less than the first warning threshold, the current pumpability state is determined to be normal; if the instability evolution index is not less than the first warning threshold and less than the second warning threshold, the current pumpability state is determined to be critical instability, triggering a warning signal. If the instability evolution index is not less than the second warning threshold, the current pumpability state is determined to be impending blockage, triggering a mandatory intervention signal.

9. The method for analyzing the pumpability of geopolymer concrete according to claim 8, characterized in that, The specific values ​​of the first warning threshold and the second warning threshold are determined in the following way: In a geopolymer concrete slurry sample with the same mix ratio as the construction site, pre-pumping tests were conducted at different temperatures. The two-dimensional joint feature value sequence from normal pumping to the entire process of pipe blockage was recorded as the pre-pumping test feature data. The pre-pumping test feature data was labeled with the corresponding pumpability status label frame by frame. Based on the characteristic data of the pre-pumping test, the instability evolution index is calculated frame by frame to form an instability evolution index sample set; Using the pumpability state label as the classification ground truth and the unstable evolution index sample set as the input variable of the classifier, the receiver operating characteristic curve analysis method is adopted to determine the first and second warning thresholds with the optimization objective of maximizing the early warning time of blockage and minimizing the false alarm rate.

10. A system for analyzing the pumpability of geopolymer concrete, characterized in that, The aforementioned geopolymer concrete pumpability analysis system is used to implement the geopolymer concrete pumpability analysis method according to any one of claims 1-9, comprising: The multi-parameter in-situ acquisition module is used to install a measuring pipe section in the pumping pipeline. The measuring pipe section is equipped with an acoustic window. Through an ultrasonic transducer installed on the outside of the acoustic window, instantaneous pulse ultrasonic waves are emitted into the polymer concrete slurry inside the pipe using the pulse transmission method, and the transmitted signal after passing through the slurry is received. At the same time, the real-time temperature of the slurry in the measuring pipe section is acquired. The slurry structure quantification module is used to extract the ultrasonic velocity and the dominant frequency attenuation coefficient from the transmitted signal under pumping conditions, forming a two-dimensional joint feature value sequence that varies with time. In the initial stage of pumping, a segment of the two-dimensional joint feature value sequence is extracted as a healthy baseline sample under normal pumping conditions. The statistical characteristics of this healthy baseline sample are calculated to construct a healthy acoustic waveform baseline that characterizes the stable flow state of the slurry. The pressure increase prediction module is used to continuously extract the two-dimensional joint feature value sequence during continuous pumping and calculate the real-time fluctuation characteristics. It uses the real-time temperature of the slurry to perform temperature compensation on the real-time fluctuation characteristics and the healthy acoustic waveform baseline, and continuously compares the temperature-compensated real-time fluctuation characteristics with the temperature-compensated healthy acoustic waveform baseline to obtain the instability evolution index that reflects the degree of deviation of the current internal structure of the slurry from the stable state. The pumpability analysis and determination module is used to compare the instability evolution index with the preset two-level warning thresholds to determine the current pumpability status of the pumping pipeline.

Images