Intelligent test method and device for UHPC workability time loss

Abstract

The invention discloses an intelligent testing method and device for UHPC workability time loss, and relates to the technical field of concrete detection. According to the method, by constructing a rheology-workability mapping model containing physical constraints, the yield stress and the plastic viscosity of UHPC are converted into a slump or expansion degree index universal for engineering, and rapid quantitative analysis of time loss is achieved; the matched testing device comprises an environment simulation bin and a double-section stirring blade, the environment simulation bin can simulate the temperature and humidity environment in the transportation process, the double-section stirring blade adopts the design that the lower portion is in a fishbone shape and the upper portion is in a flat plate shape, and the double technical problems of plastic recovery and rheological testing are solved at the same time on the basis of a single blade. According to the method, continuous and automatic monitoring of UHPC time loss in a closed simulation environment is realized, the problems that a traditional method depends on manpower, steps are tedious and a real transportation environment cannot be simulated are solved, and the detection efficiency and the engineering applicability of data are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of UHPC slump / spreading loss over time detection technology, specifically relating to a UHPC over time loss testing method and apparatus. Background Technology

[0002] Ultra-high performance concrete (UHPC) is a new type of building material that combines high strength, high toughness and long service life, and has been widely used in bridges, building curtain walls and reinforcement projects.

[0003] UHPC, a novel material with a low water-cement ratio, high cementitious material content, and fiber reinforcement, exhibits extremely sensitive workability to the environment and significant loss over time. Uncontrolled loss of workability can lead to serious engineering problems such as pump blockage and fiber clumping. However, current testing systems still use the standard for ordinary concrete (GB / T50080), employing slump or spread loss over time as the metric. This not only ignores the unique rheological behavior of UHPC—its high viscosity and thixotropic nature—and fails to accurately quantify key rheological parameters such as yield stress and plastic viscosity, but also limits testing to constant temperature and humidity conditions, lacking the ability to simulate the complex temperature and humidity gradients encountered in actual construction. Existing evaluation systems cannot accurately reflect the performance changes of UHPC under actual construction conditions. Furthermore, the cumbersome operation, high dispersion, and low efficiency of traditional testing methods make it difficult for current technologies to accurately assess the dynamic construction performance of UHPC in real-world environments.

[0004] In the field of concrete rheological parameters and slump monitoring, existing patented technologies have significant limitations. For example, patent CN109387457A uses a double-lid cylindrical inversion method to achieve secondary mixing and testing of rheological parameters over time. However, this relies on manual inversion, leading to inconvenient equipment handling, low automation, and the need for additional mixing equipment, making the process cumbersome. Furthermore, its absolute rheological parameter monitoring only calculates the loss rate over time by comparing initial and post-effect rheological parameters, failing to correlate with industry standard slump / spread indicators. Additionally, the testing intervals are subject to interface effect errors caused by mixing interruptions. While patent CN118254282A attempts real-time monitoring during transportation, it primarily relies on the idealized Bingham rheological model, which struggles to accurately reflect the nonlinear behavior of complex concrete under shear stress. It also fails to adequately consider the impact of temperature changes on concrete rheological properties, resulting in decreased measurement accuracy in high or low temperature environments. Moreover, some technologies rely on manual image recognition, are highly susceptible to lighting and environmental factors, have high model training costs, and poor practicality. Traditional testing is often conducted in stages in the laboratory, making continuous monitoring of transportation performance changes impossible. Repeated transfers and secondary mixing further compromise accuracy.

[0005] Therefore, developing a convenient, accurate, and efficient testing method that can realistically simulate the plasticity retention performance of UHPC under transportation conditions is crucial for improving the application effect of UHPC in engineering practice. Summary of the Invention

[0006] To address the aforementioned technical problems, one objective of this invention is to provide a method for testing the time-dependent loss of a UHPC. This method constructs a rheological-operability dynamic mapping model that is fully compatible with industry standard requirements and solves the problem of data integration during engineering acceptance. Another objective of this invention is to provide a device for testing the time-dependent loss of a UHPC. This device enables continuous state recovery and data acquisition during a single loading process, eliminating interface effects.

[0007] A smart testing method for UHPC performance loss over time includes the following steps: S1. Benchmark Calibration: Obtain the mix proportion information of UHPC slurry, determine its physical workability index at the initial moment, and simultaneously determine its initial rheological characteristic parameters; utilize the initial rheological characteristic parameters, mix proportion information, and measured physical workability index; The parameters of the prefabricated rheology-workability mapping model are fine-tuned to complete the calibration of the rheology-workability mapping model. The parameter fine-tuning refers to adjusting the weight parameters of the model through optimization algorithms under the premise of fixing the main structure of the model, so that the output of the rheology-workability mapping model matches the initial measured data of the current slurry. S2. Environmental simulation and settling: Place the UHPC slurry from step S1 or the same type of freshly mixed UHPC slurry in a closed test environment. According to the preset time test plan, adjust the temperature and humidity in the test environment to allow the slurry to settle for a preset time. S3. Rheological recovery and secondary testing: After settling, the slurry was stirred using a rheological testing device. Torque and speed data during the stirring process were collected, and the time-dependent rheological characteristic parameters of the slurry were extracted after processing. S4. Intelligent prediction: Input the time-dependent rheological characteristic parameters obtained in step S3, the mix ratio information and environmental parameters into the rheology-workability mapping model calibrated in step S1, and output the predicted physical workability index of the slurry at that moment. S5. Calculation of time loss: Based on the measured physical workability index obtained in step S1 and the predicted physical workability index obtained in step S4, calculate the time loss rate of UHPC slurry.

[0008] Furthermore, in step S3, the specific method for extracting the time-dependent rheological characteristic parameters includes: controlling the stirring blades of the rheological testing device to shear the slurry at multiple speeds; using the Bingham model or the Herschel-Bulkley model to perform regression analysis on the collected torque and speed data, and calculating the yield stress (τ) and plastic viscosity (η) as the time-dependent rheological characteristic parameters; In the data processing, a wavelet threshold denoising algorithm is used to process the original torque signal. The algorithm selects db series or sym series wavelet basis functions for multi-level decomposition.

[0009] Furthermore, the rheological-workability mapping model in step S1 adopts the following construction strategy: The input layer vector includes: yield stress, plastic viscosity, mix proportion information, ambient temperature, ambient humidity, and settling time; the output layer vector includes: slump (S) or spread (L); the model architecture is selected from at least one of deep neural network (DNN), support vector regression (SVR), random forest (RF), or gradient boosting decision tree (GBDT); during model training, a physical constraint loss function is introduced to ensure that the output predicted physical workability index is negatively correlated with the yield stress; The mixing ratio information includes the types of raw materials and the mixing ratio parameters.

[0010] Furthermore, the types of raw materials include any combination of several of the following: cement, silica fume, quartz sand, water, fiber, fly ash, and mineral powder.

[0011] Furthermore, the mix proportion information includes the amount of raw materials, the water-cement ratio, and the paste-aggregate ratio.

[0012] Furthermore, the initial measured data in step S1 refers to the slump or spread in workability, and the viscosity and yield stress in rheology; In step S5, the time loss rate is the percentage of the difference between the initial workability index value and the workability index value after time, relative to the initial value; the workability index value refers to the spread or slump of the UHPC slurry.

[0013] Furthermore, in step S1, after completing the rheology-workability mapping model calibration, a historical mapping model corresponding to the UHPC slurry is generated; It also includes a model applicability determination step: when testing again, if the UHPC slurry in this test is of the same type as the UHPC slurry corresponding to the existing historical mapping model, the historical mapping model is directly called and step S2 is executed; otherwise, the rheology-workability mapping model is calibrated again. "Same type" means that the types of raw materials in the mix proportion information of UHPC slurry remain unchanged, and the fluctuation of the water-reducing agent dosage in the mix proportion parameters is ≤ the benchmark value ± 0.2 percentage points.

[0014] Furthermore, in step S2, the control of the test environment adopts a gradient change mode; in the initial stage of static setting, the temperature and humidity are controlled to reach the target set value in a controlled gradual manner; during the static holding period, the temperature fluctuation range is controlled to be ≤±2℃ and the humidity fluctuation range is controlled to be ≤±4%RH; the time test scheme includes setting multiple non-equal interval test time points within the initial condensation time range.

[0015] Furthermore, the stirring blades used in the rheological testing device in step S3 adopt a dual-segment asymmetric structure, including an upper blade and a lower blade. During the test, the stirring chamber is raised to immerse the lower blade in the slurry for low-speed pre-shearing. After the rheological equilibrium state is determined, the blade continues to descend until the upper blade is completely immersed in the slurry for full parameter acquisition.

[0016] A UHPC time-lapse test device includes a fixed support, a rheological test device, an environmental simulation chamber, and a control and computing unit; The rheological testing device includes a mixing chamber, a mixing motor, and a mixing shaft. The housing of the mixing motor is connected to a fixed bracket. The output shaft of the mixing motor is fixedly connected to the mixing shaft through a torque sensor. The mixing shaft is equipped with mixing blades, which are dual-segment mixing blades. The environmental simulation chamber includes an environmental chamber and a temperature control device; the environmental chamber has an open top structure and is connected to the mixing chamber via a lifting mechanism; the humidity and temperature inside the environmental chamber are regulated by the temperature control device. The stirring motor, torque sensor, and temperature control device are all electrically connected to the control and computing unit for controlling action execution, data acquisition, data preprocessing, rheological equilibrium determination, and running the rheological-workability mapping model.

[0017] Furthermore, the dual-section stirring blade includes an upper blade and a lower blade arranged coaxially; the lower blade has a fishbone-shaped cross structure, including a central axis and inclined branch protrusions radiating outwards; the upper blade has a flat cross structure; the ratio of the projected areas of the upper blade and the lower blade on the plane of rotation is 1:(0.36-0.81).

[0018] Furthermore, the top of the environmental chamber is equipped with an automatically opening and closing sealing cover. When rheological testing is performed, the sealing cover extends in; when in a static state, the sealing cover closes to maintain the microenvironment inside the environmental chamber; the inner wall of the stirring chamber is coated with a fluorocarbon hydrophobic coating.

[0019] Compared with existing technologies, the advantages of this invention are: This invention constructs a rheology-workability mapping model under physical constraints, which can calculate the slump or spread required for engineering acceptance based on the scientific parameters (yield stress, plastic viscosity) measured by the rheometer. This model retains the objectivity of rheological testing and is fully compatible with current engineering standards, thus resolving the contradiction between "accurate measurement" and "usability".

[0020] By using gradient temperature and humidity control in an environmental simulation chamber, this invention is no longer limited to constant temperature testing, but can simulate the real environmental process of slurry changes over time during transportation (such as gradual temperature increase, surface moisture evaporation, etc.), making the test results of time loss more consistent with engineering practice.

[0021] The innovative dual-section mixing blade design (fishbone-shaped lower section + flat upper section) breaks down thixotropic structures and prevents aggregate accumulation, while the upper blade ensures uniform shearing. Combined with wavelet threshold denoising algorithms, it effectively eliminates mechanical vibration and electromagnetic interference, ensuring a high signal-to-noise ratio in high-viscosity, fibrous slurry testing.

[0022] By introducing a physical constraint loss function to train the model, the erroneous predictions that may occur in pure data-driven models (such as the expansion degree increasing when the yield stress increases) that violate physical common sense are avoided; at the same time, the set mix ratio fluctuation threshold judgment mechanism does not require retraining when the raw materials fluctuate slightly, which greatly improves the detection efficiency. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall structure of the UHPC time-loss testing device of the present invention; Figure 2 This is a perspective view of the dual-section stirring blade of the present invention; Figure 3 This is a front view of the dual-section stirring blade of the present invention; Figure 4 This is a flowchart illustrating the UHPC time-lapse loss test method of the present invention; Wherein: 1 is a fixed support, 2 is a rheological testing device, 20 is a mixing tank, 21 is a mixing motor, 22 is a mixing shaft, 23 is a mixing blade, 230 is an upper blade, 231 is a lower blade, 3 is an environmental simulation chamber, 30 is an environmental chamber, 31 is a humidification device, 32 is a cooling and heating device, 33 is a lifting mechanism, 34 is a cover, and 4 is a calculation unit. Detailed Implementation

[0024] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0025] Example 1 like Figure 1-3 As shown, this embodiment provides a UHPC time-loss testing device to simulate the time-dependent changes in the flow properties of UHPC slurry under different temperature and humidity environments. The device mainly includes a fixed support 1, a rheological testing device 2, and an environmental simulation chamber 3.

[0026] 1. Overall structure and connection relationships The fixed support 1 serves as the supporting structure for the entire device. Both the rheological testing device 2 and the environmental simulation chamber 3 are mounted on the fixed support 1.

[0027] The rheological testing device 2 includes a mixing chamber 20, a stirring motor 21, and a stirring shaft 22. The housing of the stirring motor 21 is fixed to the fixed bracket 1, and its output shaft is fixedly connected to the stirring shaft 22 via a torque sensor. The stirring shaft 22 is equipped with specially designed dual-section stirring blades 23. The stirring shaft 22 can extend into the mixing chamber 20 to drive the stirring blades 23 to stir and perform rheological testing on the UHPC slurry inside the chamber.

[0028] The environmental simulation chamber 3 includes an environmental chamber 30, which is a box-like structure with an open top and fixed to a fixed bracket 1. A mixing chamber 20 is connected to the environmental chamber 30 via a lifting mechanism 33. By controlling the lifting mechanism 33, the mixing chamber 20 can be raised or lowered to fully enter the environmental chamber 30. The environmental chamber 30 has an openable and closable cover 34 at its upper opening, with sealing strips at the cover seams. After the mixing chamber 20 descends into the environmental chamber 30, the cover 34 can be closed, placing the mixing chamber 20 in a sealed environment with controlled temperature and humidity.

[0029] 2. Environmental Simulation and Control Module To accurately simulate the environmental conditions that UHPC is in during actual construction or maintenance, the environmental simulation chamber 3 integrates a high-precision temperature and humidity control module, which specifically includes a humidification device 31 and a cooling and heating device 32.

[0030] Humidification device 31: Used to regulate ambient humidity. A humidifier can be used, and its steam outlet preferably adopts a porous media structure with a gradient pore size distribution, wherein the area with a pore size ≤50μm accounts for 60-70%, so as to achieve uniform humidity diffusion and avoid local overhumidification.

[0031] Cooling and heating device 32: used to regulate ambient temperature. Preferably, a semiconductor cooling module is used, with a cooling efficiency of 2.8-3.2, to achieve efficient and energy-saving temperature control.

[0032] Temperature and humidity closed-loop control: The device is equipped with an independent temperature and humidity closed-loop control module, which can accurately regulate the conditions within the environmental chamber 30, with temperature control accuracy up to ±2℃ and humidity control accuracy up to ±4%RH.

[0033] 3. Optimized design of key components To ensure the accuracy of test results and the reliability of the device, the following key components were specifically designed: Insulation and anti-adhesion design of the mixing chamber: The inner wall of the mixing chamber 20 is coated with a fluorocarbon hydrophobic coating to reduce the adhesion of UHPC slurry to the wall surface and ensure the accuracy of slurry dosage. Its outer wall is equipped with a double-layer vacuum insulation structure with a layer spacing of 4±0.5mm to minimize heat exchange between the slurry inside the mixing chamber 20 and the inner wall of the environmental chamber 30, ensuring the stability and response speed of temperature control.

[0034] Dual-section stirring blades: The stirring blades 23 adopt a dual-section design. The lower blade 231 features a "fishbone-shaped cross" design, with its radially radiating branched protrusions at an angle of 15°-30°. During rotation, it generates strong radial and axial shear forces, effectively breaking down the three-dimensional thixotropic network structure formed after the slurry has settled. It also utilizes the downward axial component force to counteract fiber settling, ensuring the slurry quickly returns to a uniform and measurable rheological equilibrium state. The upper blade 230 is a standard flat cross-shaped blade, forming a stable laminar shear field in the homogenized slurry, providing reliable conditions for rheological parameter regression analysis based on torque-speed data.

[0035] Lifting mechanism: The lifting mechanism 33 preferably adopts a telescopic cylinder, specifically any one of a pneumatic cylinder, a hydraulic cylinder, or an electric telescopic cylinder. The cylinder body of the telescopic cylinder is fixedly connected to the environmental chamber 30, and the end of its piston rod is fixedly connected to the mixing chamber 20, thereby driving the mixing chamber 20 to rise and fall smoothly.

[0036] The stirring motor 21, torque sensor, and temperature control device are all electrically connected to the control and computing unit 4 for controlling action execution, data acquisition, data preprocessing, rheological equilibrium determination, and running the rheological-workability mapping model.

[0037] 4. Workflow Summary During testing, the prepared UHPC slurry is first injected into the mixing chamber 20. The mixing chamber 20 is then lowered into the environmental simulation chamber 3 and sealed using the lifting mechanism 33. The conditions inside the environmental chamber are quickly adjusted to the target temperature and humidity using the humidification device 31 and the cooling and heating device 32. After the preset settling time is reached, the mixing chamber 20 is raised, the mixing motor 21 is started, and the performance indicators of the slurry are measured using the mixing blades 23.

[0038] Example 2 This embodiment aims to illustrate, by way of example, the construction and data preparation process of the "rheology-workability mapping model".

[0039] 1. Experimental Design Strategies (1) Global space design Based on the general material composition range of UHPC, the variation range of each major component was determined as the initial boundary for data acquisition: Total cementitious material: 700–1200 kg / m³ 3 Water-cement ratio: 0.16–0.22; steel fiber volume content: 0–3%; silica fume and ultrafine powder content (by mass of cementitious material): 5%–15%; high-efficiency water-reducing agent content (by mass of cementitious material): 0.8%–2.2%.

[0040] (2) Generation of mix proportions based on uniform design To achieve the most uniform coverage of the factor space with the fewest number of experiments, experimental design methods (such as uniform design, factor design, etc.) can be used to generate a series of mix proportion schemes covering the above-mentioned variation range. This embodiment uses the uniform design method to generate a highly representative basic mix proportion. Five key factors in the above global design are selected: water-cement ratio (A), steel fiber content (B), silica fume and ultrafine powder content (C), high-efficiency water-reducing agent content (D), and total amount of cementitious materials (E). Based on the number of levels for each factor and the expected number of experiments (e.g., 60 groups), a corresponding uniform design table U is selected or constructed. 60 (5 10 (or its equivalent table). The value range of each factor is evenly divided into the required levels, and the combination of each factor level is arranged according to the uniform design table, thereby generating 60 test points evenly distributed in the five-dimensional mix proportion space.

[0041] (3) Key factor response surface design For the three key factors most significantly affecting rheological behavior—water-binder ratio, steel fiber content, and ambient temperature—twenty additional experiments were generated using a central composite response surface design within their respective typical ranges (water-binder ratio: 0.18-0.20; steel fiber content: 1.0-2.0%; temperature: 20-35℃). This data aims to finely characterize the nonlinear impact of the interactions of these key factors on workability, enhancing the model's predictive accuracy in the core engineering range.

[0042] In summary, the model training was based on 80 sets of basic mix design experiments.

[0043] 2. Coupling of multi-time series data acquisition with environmental variables For each of the above combination ratios, simultaneous tests were conducted at multiple time points: Time point settings: 5 time points in total, including freshly mixed state (t0) and standing for 30, 75, 90, and 110 minutes (t1, t2, t3, t4).

[0044] Environmental variable coupling: Each combination experiment was repeated under three different environmental conditions (normal temperature and humidity: 25°C / 60%RH; high temperature and medium humidity: 35°C / 50%RH; medium temperature and low humidity: 20°C / 40%RH) to couple the effects of environmental factors.

[0045] Data volume: 80 mix proportions × 5 time points × 3 environments = 1200 raw data points. Each data point contains a complete input vector (mix proportion, environment T / RH, time t, rheological parameter τ / η) and an output label (measured slump S or spread L).

[0046] 3. Data Preprocessing and Physically Guided Data Augmentation (1) Standardization and cleaning Z-score standardization is applied to all continuous input variables (such as τ, η, T, t, etc.) to eliminate the influence of dimensions and accelerate model convergence. Outlier data points caused by operational errors or transient equipment failures are removed based on the 3σ criterion.

[0047] (2) Data augmentation based on physical rules To significantly improve the model's generalization ability without increasing the actual experimental cost, the following physics-guided data augmentation techniques are employed: Rheological parameter interpolation amplification: Based on rheological principles, between measured data from two adjacent time points (e.g., t1 and t2) under the same mix proportion and environment, assuming that the rheological parameters (τ, η) change exponentially or linearly with time, linear / spline interpolation is performed to generate synthetic data for intermediate time points. The corresponding amplification labels are obtained using the same interpolation function. This method can amplify the amount of time series data by approximately 30%.

[0048] Mix proportion amplification with minor perturbations: For existing mix proportion data, random minor perturbations are applied within the allowable range of raw material measurement errors (e.g., cement dosage ±1.5%, water-cement ratio ±0.005) to generate new, similar mix proportion data points. Simultaneously, based on close packing theory or empirical formulas, the expected trends of rheological parameters and workability labels are adjusted accordingly. This simulates the minor fluctuations in mix proportions during actual production, greatly enhancing the model's robustness to "unknown" mix proportions.

[0049] Environmental parameter extrapolation and amplification: Using the empirical relationship between viscosity and temperature, existing data are extrapolated within a limited range (±5°C) in the temperature dimension to generate synthetic data corresponding to the new temperature, and the working labels are adjusted accordingly.

[0050] Through the aforementioned enhancement techniques, the original 1200 data points can be expanded to approximately 2000-2500 high-quality training samples.

[0051] 4. Model architecture, training, and physical constraints (1) Model selection and architecture Taking a gradient boosting decision tree as the basic model architecture as an example. Input features: unit usage of each raw material (cement, silica fume, quartz sand, water, fiber, water-reducing agent), ambient temperature, ambient humidity, settling time, yield stress τ, and plastic viscosity η.

[0052] Output: Slump S (unit: mm) or spread L (unit: mm).

[0053] (2) Physical constraint loss function To prevent the model from learning absurd relationships that violate the laws of materials science, a physical constraint penalty term is introduced into the standard mean squared error loss function: Total loss = MSE(predicted value, actual value) + λ*max(0, ∂S / ∂τ) Where ∂S / ∂τ is the partial derivative of the workability predicted by the model with respect to the yield stress. λ is the penalty coefficient. This ensures that during training, the model is forced to learn the fundamental physical law that "as yield stress increases, workability decreases" (i.e., ∂S / ∂τ should be negative).

[0054] (3) Training and validation strategies The dataset was divided using stratified sampling based on the combination ratio, and the model was trained, validated, and tested according to standard machine learning procedures to ensure that the model has good generalization and prediction capabilities for novel combination ratios.

[0055] 5. Model Update and Adaptation Mechanism The system is designed with a model update mechanism. After accumulating enough new measured data, the existing model can be incrementally trained or fine-tuned to adapt to new materials or new processes.

[0056] Example 3 This embodiment details how to perform UHPC time-lapse loss testing, such as... Figure 4 As shown.

[0057] The test subject was a typical UHPC slurry, with the following raw material parameters and mix proportions: cement dosage 800 kg / m³. 3 Silica fume 100kg / m 3 Quartz sand 1100kg / m 3 1.5% high-efficiency water-reducing agent (as a percentage of the mass of the cementitious material), water-cement ratio 0.18.

[0058] S1. Reference Calibration (1) Initial workability test According to the "Technical Specification for Application of Ultra-High Performance Concrete" (T / CECS864-2021), the test object (UHPC slurry) was tested three times using a slump cone / spread plate to obtain L. 01 =620mm, L 02 =615mm, L 03 =625mm, arithmetic mean L0=620mm, difference <5% median, data is valid.

[0059] The slurry is loaded into the mixing chamber 20 of the device, the mixing motor 21 is started, and the mixing blades 23 are controlled to shear the slurry at multiple speeds (e.g., 35 rpm, 80 rpm), and torque and speed data are collected. After processing the original torque signal using wavelet threshold denoising (db4 wavelet, 3-level decomposition), regression analysis is performed using the Herschel-Bulkley model to calculate the initial yield stress τ0 = 25 Pa and the initial plastic viscosity η0 = 3.5 Pa·s.

[0060] Model calibration Input the above mix proportion information, τ0, η0 and initial expansion L0 into the calculation unit (control computer) 4, call the preset mapping model library for calibration, and obtain the calibrated rheological-workability mapping model M specifically for this slurry.

[0061] The above mixing information includes the types of raw materials and mixing parameters.

[0062] Furthermore, the aforementioned raw materials include any combination of several of the following: cement, silica fume, quartz sand, fiber, fly ash, mineral powder, and water-reducing agent.

[0063] Furthermore, the above-mentioned proportioning parameters include raw material dosage, water-cement ratio, and paste-aggregate ratio.

[0064] S2. Environmental Simulation and Static Conditioning Place the UHPC slurry or freshly mixed UHPC slurry of the same type from step S1 into a closed test environment. According to the preset time test plan, adjust the temperature and humidity in the test environment to allow the slurry to stand for a preset time.

[0065] Specifically, the evolution over time under a controlled environment: The target temperature for gradient environment control is 30℃ and the relative humidity is 60%RH. The temperature rise rate is 4℃ / min, and the humidity rise rate is 10%RH / min. After reaching the set value, the PID stabilizes at ±2℃ / ±4%RH and is left to stand for 120 minutes.

[0066] S3. Rheological Recovery and Secondary Testing: After settling, the slurry was stirred using a rheological testing device. Torque and rotational speed data were collected during the stirring process, and the time-dependent rheological characteristic parameters of the slurry were extracted after processing. Specifically, the rotational speed n1 was set to 80 rpm, and the stirring chamber speed was increased by 1 mm / s. Through real-time torque monitoring (sampling frequency 25 Hz), when the relative change in torque was <2% within 10 consecutive seconds, it was determined to be in rheological equilibrium. After equilibrium was reached, stirring was stopped, and the slurry was allowed to stand for 40 seconds to ensure that the internal stress of the slurry was completely released.

[0067] Adjust the rotation speed to n2 = 35 rpm, and continue raising the mixing chamber until the upper blades are completely submerged. Collect the yield stress τ over time. t With plastic viscosity η t τ t =35Pa、η t =6.7 Pa·s S4. Intelligent Prediction: Input the time-varying rheological characteristic parameters, mix proportion information, and environmental parameters obtained in step S3 into the rheology-workability mapping model calibrated in step S1, and output the predicted physical workability index of the slurry at that moment. Specifically: τ t η tThe calibrated mapping model M is input for calculation, and the time-dependent expansion Lt = 500 mm is obtained.

[0068] S5. Calculation of time loss: Based on the measured physical workability index obtained in step S1 and the predicted physical workability index obtained in step S4, calculate the time loss rate of UHPC slurry.

[0069] Taking the calculation of the expansion loss rate over time as an example: ΔL%=(620-500) / 620×100%=19.35% Note: The calculation of slump loss rate is similar.

[0070] In step S1 above, after the rheology-workability mapping model calibration is completed, a historical mapping model corresponding to the UHPC slurry is generated. It also includes a model applicability determination step: when testing again, if the UHPC slurry in this test belongs to the same type as the UHPC slurry corresponding to the existing historical mapping model, the historical mapping model is directly called and step S2 is executed; that is, there is no need to perform actual testing and calibration again. Otherwise, the rheology-workability mapping model needs to be calibrated again.

[0071] "Same type" means that the types of raw materials in the mix proportion information of UHPC slurry remain unchanged, and the fluctuation of the water-reducing agent dosage in the mix proportion parameters is ≤ the benchmark value ± 0.2 percentage points.

[0072] Example 4 This embodiment provides an experimental method for verifying the rationality of setting the fluctuation threshold of high-efficiency water-reducing agent dosage.

[0073] 1. Experimental Methods (1) Prepare a reference UHPC slurry with the following mix proportion: total cementitious material 850 kg / m³ 3 The water-cement ratio was 0.19, the volumetric steel fiber content was 1.8%, and the high-efficiency water-reducing agent content was 1.6% (as a percentage of the cementitious material's mass). A rheology-workability mapping model M0 was obtained by training this baseline slurry.

[0074] (2) Keeping all raw materials except for the high-efficiency water-reducing agent in the benchmark mix ratio unchanged, five experimental groups were prepared with the following high-efficiency water-reducing agent dosages: 1.3%, 1.4%, 1.6% (benchmark), 1.8%, and 2.0%.

[0075] (3) The following tests were performed on each group of slurries: After standing for 90 minutes in the freshly mixed state and in an environment of 25°C and 60%RH, their spread and rheological parameters (yield stress τ and plastic viscosity η) were measured simultaneously. The measured rheological parameters, corresponding dosage and environmental parameters were input into model M0 to obtain the predicted spread value.

[0076] 2. Experimental Results After standing for 60 minutes, the test results for each slurry and the model prediction errors are shown in the table below: As shown in the table, when the dosage of the high-efficiency water-reducing agent fluctuates within ±0.2 percentage points of the baseline value of 1.6% (i.e., 1.4% to 1.8%), the relative error between the prediction extension of model M0 and the measured value does not exceed 4.8%. Within this range, the rheological behavior of the slurry changes continuously, and the model exhibits good generalization ability. When the dosage deviation reaches -0.3 or +0.4 percentage points, the relative prediction error increases to 9.4% and 7.0%, respectively, showing a significant increase in error. This error inflection point indicates that fluctuations exceeding ±0.2 percentage points have caused the rheological state of the slurry to deviate from the basic mapping relationship learned by model M0, entering a region where prediction uncertainty increases dramatically.

[0077] 3. Recalibration procedure and effect verification For the 2% doping group, the system triggered a recalibration procedure according to step S1, and the test results are as follows: Experiments show that when raw material fluctuations exceed a set threshold ±0.2 percentage points, the rheology-workability mapping relationship of the slurry shifts significantly. By triggering a recalibration procedure using the applicability determination mechanism of this invention, the prediction error can be reduced from 7.9% to 1.0%. The above description only illustrates preferred embodiments of the present invention, but the present invention is not limited to the above embodiments.

Claims

1. A smart testing method for the performance loss of a UHPC over time, characterized in that, Includes the following steps: S1. Benchmark Calibration: Obtain the mix proportion information of UHPC slurry, determine its physical workability index at the initial moment, and simultaneously determine its initial rheological characteristic parameters; Using the initial rheological characteristic parameters, mix proportion information, and measured physical workability indicators, the parameters of the prefabricated rheology-workability mapping model are fine-tuned to complete the calibration of the rheology-workability mapping model. The parameter fine-tuning refers to adjusting the weight parameters of the model through optimization algorithms while keeping the main structure of the model fixed, so that the output of the rheology-workability mapping model matches the initial measured data of the current slurry. S2. Environmental simulation and settling: Place the UHPC slurry from step S1 or the same type of freshly mixed UHPC slurry in a closed test environment. According to the preset time test plan, adjust the temperature and humidity in the test environment to allow the slurry to settle for a preset time. S3. Rheological recovery and secondary testing: After settling, the slurry was stirred using a rheological testing device. Torque and speed data during the stirring process were collected, and the time-dependent rheological characteristic parameters of the slurry were extracted after processing. S4. Intelligent prediction: Input the time-dependent rheological characteristic parameters obtained in step S3, the mix ratio information and environmental parameters into the rheology-workability mapping model calibrated in step S1, and output the predicted physical workability index of the slurry at that moment. S5. Calculation of time loss: Based on the measured physical workability index obtained in step S1 and the predicted physical workability index obtained in step S4, calculate the time loss rate of UHPC slurry.

2. The method according to claim 1, characterized in that, In step S3, the specific method for extracting the time-dependent rheological characteristic parameters includes: controlling the stirring blades of the rheological testing device to shear the slurry at multiple speeds; using the Bingham model or the Herschel-Bulkley model to perform regression analysis on the collected torque and speed data, and calculating the yield stress (τ) and plastic viscosity (η) as the time-dependent rheological characteristic parameters; In the data processing, a wavelet threshold denoising algorithm is used to process the original torque signal. The algorithm selects db series or sym series wavelet basis functions for multi-level decomposition.

3. The method according to claim 1, characterized in that, The rheology-workability mapping model in step S1 adopts the following construction strategy: The input layer vector includes: yield stress, plastic viscosity, mix proportion information, ambient temperature, ambient humidity, and settling time; the output layer vector includes: slump (S) or spread (L); the model architecture is selected from at least one of deep neural network (DNN), support vector regression (SVR), random forest (RF), or gradient boosting decision tree (GBDT); during model training, a physical constraint loss function is introduced to ensure that the output predicted physical workability index is negatively correlated with the yield stress; The mixing ratio information includes the types of raw materials and the mixing ratio parameters.

4. The method according to claim 1 or 3, characterized in that, The initial measured data in step S1 refer to slump or spread in workability and viscosity and yield stress in rheology. In step S5, the time loss rate is the percentage of the difference between the initial workability index value and the workability index value after time, relative to the initial value; the workability index value refers to the spread or slump of the UHPC slurry.

5. The method according to claim 1, characterized in that, In step S1, after the rheology-workability mapping model calibration is completed, a historical mapping model corresponding to the UHPC slurry is generated. It also includes a model applicability determination step: when testing again, if the UHPC slurry in this test is of the same type as the UHPC slurry corresponding to the existing historical mapping model, the historical mapping model is directly called and step S2 is executed; otherwise, the rheology-workability mapping model is calibrated again. "Same type" means that the types of raw materials in the mix proportion information of UHPC slurry remain unchanged, and the fluctuation of water-reducing agent dosage in the mix proportion parameters is ≤ the benchmark value ± 0.2 percentage points.

6. The method according to claim 1, characterized in that, In step S2, the test environment is controlled using a gradient change mode; in the initial static setting period, the temperature and humidity are controlled to reach the target set value in a controlled gradual manner; during the static holding period, the temperature fluctuation range is controlled to be ≤±2℃ and the humidity fluctuation range is controlled to be ≤±4%RH; the time test scheme includes setting multiple non-equal interval test time points within the initial condensation time range.

7. The method according to claim 1, characterized in that, In step S3, the rheological testing device uses a stirring blade with a dual-segment asymmetric structure, including an upper blade and a lower blade. During the test, the stirring chamber is raised to immerse the lower blade in the slurry for low-speed pre-shearing. After the rheological equilibrium state is determined, the blade continues to descend until the upper blade is fully immersed in the slurry for full parameter acquisition.

8. A UHPC time-loss testing apparatus for implementing the method of any one of claims 1-7, characterized in that, It includes a fixed support (1), a rheological testing device (2), an environmental simulation chamber (3), and a control and computing unit (4); The rheological testing device (2) includes a stirring chamber (20), a stirring motor (21) and a stirring shaft (22). The outer shell of the stirring motor (21) is connected to the fixed bracket (1). The output shaft of the stirring motor (21) is fixedly connected to the stirring shaft (22) through a torque sensor. The stirring shaft (22) is provided with stirring blades (23). The stirring blades (23) are dual-section stirring blades (23). The environmental simulation chamber (3) includes an environmental chamber (30) and a temperature control device; the environmental chamber (30) has an open top and is connected to the mixing chamber (20) via a lifting mechanism (33); the humidity and temperature inside the environmental chamber (30) are adjusted by the temperature control device. The stirring motor (21), torque sensor and temperature control device are all electrically connected to the control and calculation unit (4) for controlling action execution, data acquisition, data preprocessing, rheological equilibrium determination, and running the rheological-workability mapping model.

9. The testing apparatus according to claim 8, characterized in that, The dual-section stirring blade (23) includes an upper blade (230) and a lower blade (231) arranged coaxially; the lower blade (231) has a fishbone-shaped cross structure, including a central axis and inclined branch protrusions radiating outwards; the upper blade (230) has a flat cross structure; the ratio of the projected area of ​​the upper blade (230) and the lower blade (231) on the plane of rotation is 1:(0.36-0.81).

10. The testing apparatus according to claim 8, characterized in that: The environmental chamber (30) is equipped with an automatically opening and closing sealing cover (34) at the top. When a rheological test is performed, the sealing cover (34) opens to allow the stirring shaft (22) to extend in. When in a static state, the sealing cover (34) closes to maintain the microenvironment inside the environmental chamber (30). The inner wall of the stirring chamber (20) is coated with a fluorocarbon hydrophobic coating.

Images