Control method based on equipment for deeply removing hexogen in magnesium nitrate solution

Abstract

The invention provides a control method based on equipment for deeply removing hexogen in a magnesium nitrate solution, which comprises the following steps: acquiring operation data of the equipment for deeply removing hexogen in a magnesium nitrate solution in real time, including physical parameters of the inner wall of the equipment, including temperature, pressure and flow velocity physical parameters, and establishing a physical parameter reference value in a scaling state; judging whether the equipment is scaled or not by comparing the physical parameters with a reference value; if the equipment is scaled, micromorphological characteristics of scaling are obtained through a scanning electron microscope, scaling distribution is identified and positioned, component information of a scaling sample is obtained, and component, form and distribution information of scaling is obtained; and comparing the real-time monitored physical parameters of the equipment with the predicted removal effects under different scaling conditions, and if the operation parameters of the equipment deviate from a target range, adjusting the temperature, pressure and flow speed operation parameters of the equipment through a preset control system.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and particularly to a control method for a device for deeply removing RDX in a magnesium nitrate solution. Background Art

[0002] During the production process of RDX, the scaling on the inner wall of the equipment is a long-term problem that plagues the development of the industry. Scaling not only reduces the heat transfer efficiency, increases energy consumption, but also leads to a decline in production capacity and even causes safety accidents. Traditional methods mainly rely on regular shutdown maintenance and manual experience judgment. This method has hysteresis and cannot detect initial scaling in a timely manner, often missing the best treatment opportunity. Relying on experience judgment lacks a scientific basis and is difficult to accurately evaluate the severity and impact of scaling. Furthermore, even if scaling is detected, traditional cleaning methods usually use chemical reagents such as strong acids and alkalis, or physical means such as high-pressure water jets. These methods may not only cause corrosion or damage to the equipment, but also produce secondary pollution. Therefore, how to achieve early warning of scaling in RDX production equipment has become an urgent problem to be solved. Moreover, how to accurately judge the degree, location, and type of scaling without shutting down the machine also faces huge challenges. In addition, how to formulate a precise, efficient, and equipment-non-damaging cleaning plan according to the specific situation of scaling is an objective that the industry has long pursued. Finally, how to suppress the further development of scaling and extend the service life of the equipment by optimizing operating parameters after scaling occurs is also an important research direction. Summary of the Invention

[0003] The present invention provides a control method for a device for deeply removing RDX in a magnesium nitrate solution, mainly including:

[0004] Real-time obtain the operation data of the device for deeply removing RDX, including the physical parameters of the inner wall of the equipment. The physical parameters include temperature, pressure, and flow rate physical parameters. At the same time, establish a reference value of the physical parameters in the scaling state, and judge whether the equipment is scaled by comparing the physical parameters with the reference value;

[0005] If the equipment is scaled, obtain the microscopic morphological characteristics of the scaling through a scanning electron microscope, identify and locate the scaling distribution, and obtain the component information of the scaling sample to obtain the composition, morphology, and distribution information of the scaling;

[0006] Establish a multi-scale numerical model according to the scaling information, simulate the interaction between RDX and the scaling layer under the action of fluid, output the equipment state and the change of the scaling layer, establish a correlation model between the scaling characteristics and the removal efficiency, and predict the removal effect under different scaling conditions;

[0007] Compare the physical parameters of the device monitored in real time with the removal effects under different scaling conditions obtained through prediction. If it is found that the operating parameters of the device deviate from the target range, adjust the operating parameters of the temperature, pressure, and flow rate of the device through a preset control system;

[0008] Continuously monitor the change of the physical parameters of the inner wall of the device to judge whether scaling is inhibited. If the monitored physical parameters tend to be stable or return to the reference value under the preset scale-free state, and the scaling thickness or scaling area stops growing or starts to decrease, it is determined that the scaling is inhibited, then evaluate the removal efficiency of hexogen, and feedback the evaluation result to the associated model for optimization and update;

[0009] If, after adjusting the operating parameters, the physical parameters of the inner wall of the monitored device continue to deviate from the reference value under the scale-free state, or the scaling thickness or scaling area continues to grow, or even accelerates, it is evaluated that the scaling problem has not been alleviated after adjusting the operating parameters, then start the preset device cleaning procedure to clean the inner wall of the device and remove the scaling layer.

[0010] Furthermore, the operating data of the hexogen deep removal device is obtained in real time, including the physical parameters of the inner wall of the device. The physical parameters include temperature, pressure, and flow rate physical parameters. At the same time, a reference value of the physical parameters under the scaling state is established. By comparing the physical parameters with the reference value, it is judged whether the device is scaled, including:

[0011] Obtain the numerical values of the physical parameters of the hexogen deep removal device in the current operating state from the temperature sensor, pressure sensor, and flow rate sensor, and periodically sample and record the physical parameters according to the preset sampling time interval through the data acquisition module. Use the exponentially weighted moving average method to smooth the collected temperature, pressure, and flow rate physical parameters, calculate the statistical mean value according to the smoothed physical parameters, and use it as the reference value of the physical parameters under the scaling state. Calculate the difference between the smoothed physical parameters and the reference value of the physical parameters, and use the least squares method to calculate the standard deviation of the difference sequence to obtain the state parameters of the inner wall of the device. Compare the preset scaling state determination value with the state parameters of the inner wall of the device. If the state parameters of the inner wall of the device are greater than the scaling state determination value, it is determined that the hexogen deep removal device is scaled.

[0012] Furthermore, if the device is scaled, obtain the microscopic morphological characteristics of the scale through a scanning electron microscope, identify and locate the scale distribution, and obtain the composition information of the scale sample to obtain the composition, morphology, and distribution information of the scale, including:

[0013] The fouling surface is scanned and imaged in zones by a scanning electron microscope. Based on a preset scanning area size and a preset pixel resolution, microscopic fouling morphology images are collected to obtain high-resolution scanning electron microscope images of the fouling surface. For the scanning electron microscope images, an image segmentation algorithm is used to identify the fouling area, calculate the proportion of the fouling area and the thickness distribution, and generate a fouling distribution density map. Based on the fouling distribution density map, the area with the highest fouling density is selected as the sampling point, and an energy spectrometer is used for element detection. Peak identification and quantitative analysis are performed on the collected energy spectrum data to obtain the mass fractions of various elemental components in the fouling sample. A three-dimensional reconstruction algorithm is used to perform three-dimensional modeling on the fouling area, and the fouling thickness data is calculated through the height information in the vertical direction to generate a fouling thickness distribution map. According to the fouling distribution density map, the mass fractions of elemental components, and the fouling thickness distribution map, a fouling feature dataset is established, and a convolutional neural network is used to identify the phase structure features of the fouling to obtain the correlation feature data of the fouling composition and distribution. Based on the correlation feature data, the fouling area is classified and labeled to generate a fouling feature map containing fouling composition, morphology, and distribution information.

[0014] Further, a multi-scale numerical model is established according to the fouling information to simulate the interaction between cyclonite and the fouling layer under the action of fluid, output the equipment state and the change of the fouling layer, and establish a correlation model between fouling characteristics and removal efficiency to predict the removal effect under different fouling conditions, including:

[0015] According to the fouling composition, morphology, and spatial distribution data, a multi-scale numerical model of the fouling layer is constructed using the finite element method, the porosity and surface roughness of the fouling layer are meshed to obtain the discretized geometric structure data of the fouling layer. Based on the discretized geometric structure data, a system of fluid mechanics equations is established, including the continuity equation and the momentum equation, and a two-phase flow solver is used to calculate the flow field distribution data at the interface between the fluid and the fouling layer. The cyclonite concentration field is calculated according to the flow field distribution data and the mass transfer equation, and a correlation equation between the concentration gradient and the mass transfer coefficient at the fouling layer interface is established to obtain the interface mass transfer flux data. The local dissolution rate of the fouling layer is calculated using the dissolution kinetics equation, and the heat transfer data is obtained by solving the temperature field equation to obtain the dynamic evolution data of the geometric morphology of the fouling layer. Based on the dynamic evolution data of the geometric morphology of the fouling layer, a dissolution process dataset is established, including the morphological characteristics of the fouling layer, the interface mass transfer flux, and the local dissolution rate, and a deep neural network is used to construct a correlation model between fouling characteristics and removal efficiency. According to the correlation model and preset working condition parameters, the gradient boosting tree algorithm is used to calculate the dissolution amount and dissolution time of the fouling under different fouling layer distribution conditions to obtain the prediction data of the removal efficiency of the fouling layer.

[0016] Further, compare the physical parameters of the device monitored in real time with the removal effects under different fouling conditions obtained by prediction. If it is found that the operating parameters of the device deviate from the target range, adjust the operating parameters of the temperature, pressure, and flow rate of the device through a preset control system, including:

[0017] Collect the current operating parameters of the device from the temperature sensor, pressure sensor, and flow rate sensor, and perform a moving average process on the collected parameter data according to a preset sampling period. The sampling period is set to one-fourth of the response time of the operating parameters to obtain the real-time operating state data of the device. Match the real-time operating state data of the device with the standard condition data in the removal effect prediction database, calculate the deviation value of the operating parameters using the Manhattan distance, and obtain the temperature deviation, pressure deviation, and flow rate deviation data. Establish a parameter adjustment sequence for the temperature deviation, pressure deviation, and flow rate deviation data, predict the change trend of each parameter through a long short-term memory network, and the input features include historical deviation values and adjustment records to obtain the adjustment direction of the operating parameters. Set the target values of the temperature controller, pressure controller, and flow rate controller according to the adjustment direction of the operating parameters, and generate a control instruction sequence based on the adjustment dead zone and adjustment step size of each controller. Optimize the control instruction sequence using a proportional integral derivative operator. The proportional coefficient is determined by the slope of the parameter response curve, the integral time is determined by the parameter adjustment period, and the differential time is determined by the parameter fluctuation period to obtain the corrected control instruction.

[0018] Further, continuously monitor the change of the physical parameters of the inner wall of the device to judge whether the fouling is inhibited. If the monitored physical parameters tend to be stable or return to the reference value under the preset scale-free state, and the fouling thickness or fouling area stops growing or begins to decrease, it is determined that the fouling is inhibited, and then evaluate the removal efficiency of cyclonite, and feedback the evaluation result to the associated model for optimization and update, including:

[0019] Collect the physical parameters of the temperature, pressure, and flow rate of the inner wall of the device according to the preset monitoring period, calculate the short-term change rate of the physical parameters using the exponentially weighted moving average method, and obtain the fluctuation period of the physical parameters through autocorrelation analysis. Decompose the trend of the continuously collected physical parameter sequence, calculate the root mean square error value of the parameter sequence based on the decomposition result, and judge whether the physical parameter sequence enters a stable state. Use an image segmentation algorithm to calculate the fouling thickness and fouling area, obtain the change rate of the fouling characteristic quantity through differential operation, and calculate the recovery degree of the physical parameters by comparing with the reference value of the preset scale-free state. Set the inhibition judgment threshold according to the recovery degree of the physical parameters. If the fluctuation value of the physical parameters is less than the preset threshold within the continuous monitoring period, and the change rate of the fouling characteristic quantity is negative, it is determined that the fouling is inhibited.

[0020] Further, if after adjusting the operating parameters, the monitored physical parameters of the inner wall of the device continuously deviate from the reference values in the scale-free state, or the scale thickness or scale area continuously increases, or even accelerates the growth, it is evaluated that the scale problem has not been alleviated after adjusting the operating parameters, and then a preset device cleaning program is started to clean the inner wall of the device and remove the scale layer, including:

[0021] Collect the physical parameters of the inner wall of the device through temperature sensors, pressure sensors, and flow rate sensors, perform moving average processing on the collected data according to a preset sampling period, and calculate the difference sequence between the physical parameters and the reference values in the scale-free state. Calculate the temperature deviation rate, pressure deviation rate, and flow rate deviation rate according to the difference sequence, and use the weighted summation method to obtain a comprehensive deviation index to judge the deviation state of the physical parameters. Based on the scanning electron microscope, collect images of the fouling area, calculate the fouling thickness and fouling area through an image segmentation algorithm, and use the difference method to obtain the change rate of the fouling characteristic quantity. Use a convolutional neural network to extract the fouling morphology characteristics, and combine the change rate of the fouling characteristic quantity to construct a fouling state evaluation index to judge the growth trend of the fouling layer. Establish a bivariate correlation matrix according to the comprehensive deviation index and the fouling state evaluation index, and predict the development trend of the fouling problem through a Markov chain to obtain the evaluation value of the alleviation of the fouling problem. For the working conditions where the evaluation value of the alleviation of the fouling problem exceeds the preset threshold, generate chemical cleaning agent ratio instructions and mechanical cleaning head position instructions, and start the cleaning execution mechanism. Use an acoustic wave sensor to monitor the peeling state of the fouling layer during the cleaning process, and detect the fouling content in the cleaning liquid through an online turbidimeter to realize the closed-loop control of the cleaning process.

[0022] The technical solution provided by the embodiment of the present invention may include the following beneficial effects:

[0023] The present invention discloses a control method for a device for deeply removing cyclonite in a magnesium nitrate solution. The method monitors the physical parameters of the inner wall of the device in real time, establishes a scale state reference value, and uses a scanning electron microscope to obtain the microscopic characteristics of the scale. According to the scale information, a multi-scale numerical model is constructed to simulate the interaction between cyclonite and the scale layer and predict the removal effect under different fouling conditions. The present invention compares the real-time monitoring data with the prediction results, automatically adjusts the operating parameters of the device, and inhibits the growth of the scale. If the parameter adjustment is ineffective, a preset cleaning program is started. Through continuous monitoring and evaluation, the present invention continuously optimizes the correlation model and improves the cyclonite removal efficiency. The method realizes the intelligent operation of the device for deeply removing cyclonite, effectively solves the fouling problem, and improves the performance and service life of the device. Description of the Drawings

[0024] Figure 1 It is a flowchart of a control method for a device for deeply removing cyclonite in a magnesium nitrate solution according to the present invention. Detailed Embodiments

[0025] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this specification without making creative efforts shall fall within the scope of protection of this specification.

[0026] As Figure 1 , a control method for a device for deep removal of RDX in a magnesium nitrate solution in this embodiment may specifically include:

[0027] S101. If the control function of the device for deep removal of RDX is triggered, the physical parameters during the operation of the device, including temperature, pressure, and flow rate, are collected in real time, and a reference value in the fouling state is established through data processing. Whether fouling occurs on the inner wall of the device is judged according to the comparison result between the real-time parameters and the reference value.

[0028] In the embodiment of the present invention, the device for deep removal of RDX has a real-time monitoring function, which can be started by the built-in control system of the device or remotely activated through an external instruction. The specific triggering method is determined according to the actual application scenario. The collected physical parameters are from the sensors on the inner wall of the device, including temperature sensors, pressure sensors, and flow rate sensors. The temperature sensor selects a high-precision platinum resistance with a measurement range covering 0 to 200 degrees Celsius and an accuracy of 0.1 degree Celsius; the pressure sensor adopts a piezoresistive design with a range of 0 to 10 MPa and an accuracy of 0.01 MPa; the flow rate sensor is an electromagnetic flowmeter with a measurement range of 0 to 15 m / s and an accuracy of 0.1 m / s. These sensors record data periodically at a sampling interval of 10 seconds through a data acquisition module to form time series data for subsequent analysis.

[0029] S1011. The physical parameters are collected by the sensors and data smoothing processing is performed to generate a stable reference value. In the embodiment of the present invention, the data acquisition module stores the temperature, pressure, and flow rate data collected by the sensors as a sampling sequence at a preset time interval. Since there may be random interferences during the operation of the device, such as local eddies or external vibrations, the original data may contain noise. To ensure the reliability of the data, the exponential weighted moving average method is used to smooth the sampling sequence, and the smoothing coefficient is set to 0.3 to retain the parameter change trend and filter out instantaneous fluctuations. The smoothed data calculates the statistical mean based on an 8-hour operation cycle, and the temperature reference value of 85 degrees Celsius, the pressure reference value of 3.5 MPa, and the flow rate reference value of 8 m / s are obtained respectively as the reference standards in the fouling state.

[0030] S1012. Perform a difference analysis on the smoothed data and the reference value, calculate the state parameters of the inner wall of the device, and determine the scaling state.

[0031] After obtaining the smoothed data, in the embodiment of the present invention, the difference between the real-time physical parameters and the reference value is calculated to analyze the change of the state of the inner wall of the device. The standard deviation of the difference sequence is calculated by the least square method to generate the state parameters of the inner wall of the device. When scaling occurs in the device, the attached substances on the inner wall will cause the temperature to increase by 2 to 5 degrees Celsius, the pressure to increase by 0.5 to 1 MPa, and the flow rate to decrease by 1 to 2 m / s. Through the statistical analysis of a large amount of operation data, the scaling determination threshold of the state parameter is set to 0.8. If the calculated state parameter exceeds this threshold, it is determined that scaling has occurred on the inner wall of the device. At this time, the system records the deviation amplitude of the current parameter to provide a basis for subsequent processing.

[0032] In the embodiment of the present invention, if it is determined that the scaling state exists, the abnormal fluctuations of the sensor data are further analyzed. The early warning thresholds of temperature, pressure, and flow rate are set to 95 degrees Celsius, 5 MPa, and 12 m / s respectively, and the data points exceeding these thresholds are marked as abnormal. The marked data is corrected using a convolutional neural network, which includes 3 convolutional layers and 2 fully connected layers. The input is the abnormal data sequence of 128 time points and its time series characteristics, and the output is the corrected physical parameter value. The corrected data is used to dynamically update the reference value to adapt to the long-term changes in the operating state of the device, while maintaining sensitivity to scaling deviations.

[0033] In the embodiment of the present invention, the accurate monitoring of the scaling state of the RDX deep removal device is realized through the above steps. Real-time acquisition and data processing can timely detect the scaling problem on the inner wall, avoiding the decrease of heat transfer efficiency or the blockage of the flow channel caused by the deterioration of scaling. The high-precision design of the sensor and data smoothing processing ensure the reliability of the monitoring results, while the dynamic update mechanism enhances the adaptability of the method to the changes in the device operation. By comparing with the preset threshold, not only can the occurrence of scaling be determined, but also data support can be provided for subsequent control strategies, thereby improving the operation stability of the device and the treatment effect of RDX.

[0034] It can be understood that the embodiment of the present invention does not overly limit the specific implementation methods of the sensor type and sampling interval, which can be adjusted by technicians according to actual needs to adapt to different production environments and device characteristics.

[0035] S102. If it is determined that scaling has occurred in the RDX deep removal device, use a scanning electron microscope to perform high-resolution imaging on the microscopic morphology of the scaling surface on the inner wall of the device and extract features. At the same time, combined with energy spectrum analysis and three-dimensional reconstruction technology, comprehensively obtain the composition, morphological characteristics, and spatial distribution information of the scaling, and establish a detailed scaling feature map to support the formulation of subsequent control strategies.

[0036] In an embodiment of the present invention, when the inner wall of the device is determined to be in a fouling state, a scanning electron microscope is immediately started to perform refined analysis on the fouling area. The scanning electron microscope collects data in a partitioned scanning manner. The scanning area is set to 500 microns × 500 microns, and the pixel resolution reaches 2048 × 2048. The actual size corresponding to each pixel is about 0.24 microns, ensuring that the tiny structural details on the fouling surface can be captured. The working voltage is set to 20 kV, and the working distance is maintained at 10 mm. High-contrast surface topography images are generated through backscattered electron imaging technology, providing reliable basic data for subsequent feature extraction. This high-resolution imaging method can clearly present the microscopic texture and distribution characteristics of fouling, laying a foundation for accurately identifying the fouling area.

[0037] S1021 Perform image segmentation and density distribution analysis based on the scanning electron microscope image to generate a fouling distribution density map.

[0038] In an embodiment of the present invention, for the collected scanning electron microscope images, an image segmentation algorithm is used to automatically identify the fouling area. In the segmentation process, the fouling and non-fouling areas are first distinguished by the gray threshold method. Since the fouling part usually shows a higher gray value in the backscattered electron image, a dynamic threshold can be set to adapt to the image characteristics under different lighting conditions. After the segmentation is completed, the area ratio of the fouling area is calculated. For example, in a certain analysis, the fouling area accounts for 35% of the field of view, and a fouling distribution density map is further generated. The map uses a pseudo-color visualization method, where the areas with a fouling density exceeding 50% are marked in red, and the areas with a density lower than 20% are shown in blue, intuitively reflecting the spatial distribution law of fouling. The analysis shows that the fouling distribution often presents a cluster-like feature, and the local density can reach up to 65%, which is closely related to the fluid scouring and deposition conditions, providing a basis for the selection of subsequent sampling points.

[0039] S1022 Detect the fouling composition through an energy spectrometer and combine it with three-dimensional reconstruction technology to extract the elemental composition and thickness distribution characteristics of the fouling.

[0040] After determining the fouling distribution density map, the embodiment of the present invention selects the area with the highest density as the key analysis area and uses an energy spectrometer to detect the elemental composition. The detector of the energy spectrometer maintains a 40-degree angle with the sample surface to optimize the acquisition efficiency of X-ray signals. Five representative sampling points are selected in the high-density area for analysis, and the mass fractions of each element are obtained through peak identification and quantitative calculation. For example, the detection results show that the fouling is mainly composed of calcium, sulfur, and oxygen, with mass fractions of 32.5%, 15.8%, and 48.2% respectively, and the proportion of trace elements is about 3.5%, indicating that the fouling may be mainly calcium sulfate. Subsequently, a scanning electron microscope is used to collect image pairs at 0-degree and 45-degree tilt angles, and a three-dimensional reconstruction algorithm is combined to calculate the three-dimensional structure of the fouling. The algorithm restores the height information through the principle of stereo vision and generates thickness distribution data. The results show that the fouling thickness varies between 15 microns and 85 microns, with an average value of 45 microns, and a stacking zone with a width of about 200 microns is locally formed. This non-uniformity of the thickness distribution reflects the influence of hydrodynamics on fouling deposition and provides key inputs for subsequent model analysis.

[0041] In the embodiment of the present invention, a fouling feature dataset is constructed by integrating the fouling distribution density map, elemental composition data, and thickness distribution information. To further analyze the phase structure of the fouling, a convolutional neural network is used for feature recognition. The network design includes five convolutional layers, which are used to extract and fuse the morphology, composition, and thickness features respectively. The training inputs include density distribution, elemental mass fraction, and thickness data. After training, the network identifies that the fouling is mainly calcium sulfate crystals, with particle sizes between 10 and 30 microns, and the morphology is mainly a mixture of needle-like and plate-like. This phase feature is consistent with the banded distribution law in the high-flow-rate area, indicating that fluid shear force may be the main driving factor for the formation of fouling morphology.

[0042] Finally, the embodiment of the present invention generates a feature map containing fouling composition, morphology, and distribution information through the above analysis. The map uses color coding to label different phase regions. For example, calcium sulfate fouling is labeled in red, and at the same time, thickness contour lines are superimposed to intuitively show the thickness change trend. The map reveals the banded distribution characteristics of the fouling extending along the flow direction, which not only provides evidence for verifying the hydrodynamics hypothesis but also provides accurate data support for adjusting operating parameters and formulating cleaning strategies. It can be understood that the scanning area size, resolution, and algorithm parameters can be flexibly adjusted by technicians according to the equipment characteristics and fouling complexity to ensure the applicability of the analysis results. Through the implementation of this step, the multi-dimensional characteristics of the fouling can be comprehensively grasped, laying a solid foundation for the dynamic control of the RDX deep removal equipment.

[0043] S103. Establish a multi-scale numerical model based on the scaling information to simulate the interaction between RDX and the scaling layer under the action of fluid, output the equipment state and the change of the scaling layer, establish a correlation model between the scaling characteristics and the removal efficiency, and predict the removal effect under different scaling conditions. The scaling information includes scaling composition, morphology, and distribution information; the equipment state includes flow field, pressure, temperature, and RDX concentration distribution.

[0044] In the embodiment of the present invention, based on the scaling characteristic data obtained in the early stage, including element composition, microscopic morphology, and spatial distribution information, numerical simulation means are used to analyze the evolution process of the scaling layer under the action of fluid. In model construction, the geometric structure of the scaling layer is first discretized. Considering the uneven characteristics of its surface roughness and porosity, the finite element method is used to divide the grid. In the convex area with higher roughness, the grid size is set to 5 microns to capture the detailed features, while in the flat area, a 20-micron grid is adopted, and finally a computational domain with about 500,000 hexahedral elements is generated. This adaptive grid division method can balance the calculation accuracy and efficiency, ensuring that the simulation results accurately reflect the complex geometric characteristics of the scaling layer.

[0045] S1031. Calculate the flow field distribution at the interface of the scaling layer using the hydrodynamic equations and analyze the fluid behavior.

[0046] In the embodiment of the present invention, based on the discretized geometric structure, a set of hydrodynamic equations including the continuity equation and the momentum equation is established to describe the interaction between the fluid and the scaling layer. The inlet boundary of the computational domain is set as the velocity inlet condition, with a fixed flow velocity of 2 m / s, and the outlet boundary adopts the pressure outlet condition to simulate the actual working conditions. The VOF two-phase flow solver is used to calculate the flow field distribution. The results show that the velocity of the fluid on the surface of the scaling layer shows significant spatial differences. The maximum flow velocity in the convex area can reach 3.5 m / s, while the minimum in the concave area drops to 0.5 m / s. This non-uniformity of the flow velocity distribution is due to the geometric blocking effect of the scaling surface, which directly affects the subsequent mass transfer and dissolution processes. During the calculation process, the k-ε turbulence model is adopted to further optimize the simulation of the shear force in the high-flow velocity area, providing guarantee for the reliability of the flow field data. S1032 Calculate the mass transfer flux and dissolution rate of the scaling layer through the mass transfer and heat transfer equations.

[0047] After obtaining the flow field distribution data, the embodiments of the present invention further construct a mass transfer equation to analyze the change of the RDX concentration field. The inlet RDX concentration is set to 500 mg / L, and the concentration gradient and interfacial mass transfer coefficient are calculated through the convective-diffusion equation. The results show that due to the higher flow velocity in the convex region, the mass transfer coefficient can reach 0.002 m / s, while in the concave region it is only 0.0005 m / s, reflecting the significant influence of the flow field on the mass transfer behavior. At the same time, the local dissolution rate of the fouling layer is calculated by combining the dissolution kinetics equation and the temperature field equation. The solution of the temperature field shows that at the working temperature of 85 °C, the surface temperature of the fouling layer is distributed between 80 and 90 °C, and the dissolution rate in the local high-temperature and high-mass transfer coefficient regions can reach up to 0.05 mm / h. This multi-physical field coupling calculation method can comprehensively reveal the synergistic effect of fluid, concentration, and temperature on fouling dissolution, providing multi-dimensional data support for predicting the removal efficiency.

[0048] In the embodiments of the present invention, based on the above calculation results, a dynamic evolution data set including the morphological characteristics of the fouling layer, interfacial mass transfer flux, and local dissolution rate is generated. The data set covers 2000 groups of different working condition data, and each group includes 48 characteristic parameters, such as porosity, surface roughness, flow velocity, and temperature. A deep neural network is used to construct a correlation model between fouling characteristics and removal efficiency. The network is designed with a 5-layer structure, and the number of neurons is 48, 96, 64, 32, and 16 in sequence. The ReLU activation function is used, and the training is completed after 50000 iterations. The trained model combines the gradient boosting tree algorithm to predict the fouling removal effect under different working conditions. For example, tests are carried out in the range of inlet flow velocity from 1 to 4 m / s, temperature from 70 to 100 °C, and RDX concentration from 300 to 700 mg / L. The prediction results show that the fouling layer can be completely dissolved within 4 hours under the optimal working conditions, while the dissolution time is extended to more than 12 hours under the adverse working conditions.

[0049] Through the multi-scale numerical simulation of the embodiments of the present invention, not only can the flow field, pressure, temperature, and RDX concentration distribution in the device be output, but also the geometric changes of the fouling layer can be dynamically tracked, providing a scientific basis for the optimization of process parameters. It can be understood that the grid division strategy and the choice of solver can be adjusted by technicians according to the device characteristics to meet the simulation requirements of different fouling types and operating conditions. This method significantly improves the accuracy of the removal efficiency prediction by coupling physical field calculations with machine learning predictions, laying a foundation for the intelligent control of RDX deep removal devices.

[0050] S104. By comparing the physical parameters monitored in real time by the RDX deep removal device with the predicted fouling removal effect, if it is found that the operating parameters deviate from the expected target, the preset control system is used to dynamically adjust the temperature, pressure, and flow velocity to optimize the device performance and improve the fouling treatment efficiency.

[0051] In the embodiments of the present invention, the monitoring of equipment operating parameters relies on a high-precision sensor group, including a temperature sensor using a platinum resistance thermometer with a measurement range of 0 to 200 degrees Celsius and an accuracy of 0.1 degree Celsius; a piezoresistive pressure sensor with a range of 0 to 10 MPa and an accuracy of 0.01 MPa; and a flow velocity sensor composed of an electromagnetic flowmeter with a measurement range of 0 to 15 m / s and an accuracy of 0.1 m / s. The collected parameter data sets the sampling period according to their respective dynamic response characteristics. Among them, the temperature response time is 60 seconds and the sampling period is set to 15 seconds; the pressure response time is 10 seconds and the sampling period is 2.5 seconds; the flow velocity response time is 5 seconds and the sampling period is 1.25 seconds. The collected data is smoothed by the moving average method, and the window size is set to 5 times the sampling period to filter out instantaneous noise and generate stable real-time operating state data, providing a reliable basis for subsequent deviation analysis.

[0052] S1041, calculate the parameter deviation based on the Manhattan distance and establish an adjustment sequence to predict the adjustment direction.

[0053] After obtaining the real-time operating state data, the embodiments of the present invention match it with the optimal parameter combination in the standard operating condition database. The standard operating condition database stores reference data under different fouling degrees. For example, when the fouling is mild, the temperature is 85 degrees Celsius, the pressure is 3.5 MPa, and the flow velocity is 2 m / s, while when the fouling is severe, the temperature needs to rise to 95 degrees Celsius, the pressure increases to 4.5 MPa, and the flow velocity increases to 3 m / s. The Manhattan distance algorithm is used to calculate the deviation between the real-time data and the standard data. The formula is the sum of the absolute differences of each parameter, and the results respectively generate the temperature deviation, pressure deviation, and flow velocity deviation. If the deviation value exceeds the preset threshold, that is, the temperature deviation is greater than 5 degrees Celsius, the pressure deviation is greater than 0.5 MPa, or the flow velocity deviation is greater than 0.3 m / s, the adjustment process is triggered. Subsequently, the deviation data is organized into a time series and input into a long short-term memory network for trend prediction. The network adopts a double-layer structure, including 128 memory units. The input features include the deviation values in the past 4 hours and the historical adjustment records, and 12 data points are collected per hour. By analyzing the historical trend, the network outputs the direction of parameter adjustment, such as the temperature needs to increase or the flow velocity needs to increase. The prediction accuracy can reach 92% after training, significantly improving the scientificity of the adjustment decision.

[0054] In an embodiment of the present invention, the target values of the temperature, pressure, and flow rate controllers are set according to the predicted adjustment directions, and an initial control instruction sequence is generated. The adjustment dead zones of each controller are set to twice the measurement accuracy, that is, the temperature dead zone is 0.2 degrees Celsius, the pressure dead zone is 0.02 MPa, and the flow rate dead zone is 0.2 m / s; the adjustment step sizes are 2 degrees Celsius, 0.2 MPa, and 0.1 m / s respectively according to the response characteristics. To further optimize the instructions, a proportional-integral-derivative calculator is used to finely adjust the sequence. The proportionality coefficients are determined according to the slopes of the parameter response curves. For example, the temperature control is 2, the pressure is 1.5, and the flow rate is 1.2; the integral times are linked to the adjustment periods and are 180 seconds, 30 seconds, and 15 seconds respectively; the derivative times are related to the fluctuation periods and are 45 seconds, 7.5 seconds, and 3.75 seconds respectively. The optimized control instructions are output through the actuator, and the objective function, that is, the mean square error between the operating parameters and the target values, is calculated using the gradient descent method every 60 seconds for dynamic optimization to ensure that the parameters quickly converge to the target range.

[0055] Through the above control strategy, the embodiment of the present invention realizes the coordinated adjustment of multiple parameters. During actual operation, the temperature fluctuation is controlled within plus or minus 1 degree Celsius, the pressure fluctuation does not exceed plus or minus 0.1 MPa, and the flow rate fluctuation is limited to plus or minus 0.05 m / s. Compared with the traditional single-parameter adjustment method, the stability of equipment operation and the scale removal efficiency are significantly improved. It can be foreseen that the sampling period, adjustment step size, and network parameters can be further adjusted according to the equipment model and working conditions requirements to adapt to a more complex operating environment, thereby providing more flexible technical support for the intelligent management of the RDX deep removal equipment.

[0056] S105 continuously monitors the changes in the physical parameters of the inner wall of the RDX deep removal equipment, analyzes the scale inhibition effect and evaluates the removal efficiency. If the parameters tend to be stable and approach the scale-free state reference value and the scale characteristic quantity no longer increases, the correlation model is optimized to improve the prediction accuracy and control effect.

[0057] In an embodiment of the present invention, the temperature, pressure, and flow rate parameters of the inner wall of the equipment are continuously collected through high-frequency sampling. The sampling periods are set to 15 seconds, 5 seconds, and 2 seconds respectively to capture the dynamic changes of the parameters. The collected data is smoothed using the exponentially weighted moving average method, and the weight coefficient is set to 0.3, which can not only retain the short-term trend but also reduce noise interference. The smoothed data is used to calculate the short-term change rate of the physical parameters, and the fluctuation periods are determined through autocorrelation analysis. For example, the temperature fluctuation period is about 30 minutes, and the pressure and flow rate are 10 minutes and 5 minutes respectively. These periods reflect the dynamic characteristics of equipment operation and scale treatment, providing a time-scale basis for subsequent trend analysis.

[0058] S1051. Decompose the trend of the physical parameter sequence and calculate the stability index to judge the operating state. In the embodiment of the present invention, the continuously collected physical parameter sequence is decomposed by trend with a 24-hour period, and three components, namely, the long-term trend term, the periodic term, and the random term, are decomposed. The long-term trend term reflects the change direction of the parameter with the scale inhibition treatment, the periodic term captures the periodic fluctuations, and the random term quantifies the noise level. The stability of the parameter is evaluated by calculating the root mean square error value of the random term. If the error value is lower than 0.5% and the amplitude of the periodic term is less than 1% of the mean value, it is considered that the parameter sequence enters a stable state. For example, in actual monitoring, after the temperature drops from 95 degrees Celsius to 85 degrees Celsius, the error value drops to 0.3%, indicating that the treatment process has tended to be stable. This decomposition method improves the judgment accuracy of parameter stability through multi-dimensional analysis and lays a foundation for the evaluation of the scale inhibition effect.

[0059] In the embodiment of the present invention, the monitoring of the scale characteristics is realized through an image segmentation algorithm. The scale area is identified by using an adaptive threshold method and its thickness and area are calculated. During the treatment process, the scale area gradually decreases from the initial 350 square centimeters to 150 square centimeters, and the thickness decreases from the maximum of 85 micrometers to 35 micrometers. The change rate of the characteristic quantity is calculated through differential operation. The results show that the area reduction rate is 8 square centimeters per hour and the thickness reduction rate is 2 micrometers per hour, both of which are negative values and continue to decrease. The inhibition judgment threshold is set according to the parameter recovery degree. The temperature fluctuation threshold is plus or minus 1 degree Celsius, the pressure is plus or minus 0.1 megapascal, and the flow rate is plus or minus 0.05 meters per second. When the continuous 8-hour monitoring shows that the temperature is stable at 84 to 86 degrees Celsius, the pressure is at 3.45 to 3.55 megapascals, the flow rate is at 1.95 to 2.05 meters per second, and the scale characteristic quantity no longer increases, it is determined that the scale has been effectively inhibited.

[0060] S1052. Construct an evaluation index for the removal efficiency and optimize the correlation model through machine learning.

[0061] After determining the scale inhibition, the embodiments of the present invention construct an evaluation index for the removal efficiency of cyclonite, including the physical parameter recovery time, the reduction amplitude of the scale characteristic quantity, and the parameter stability degree. The recovery time is defined as the time required for the parameter to reach stability, usually 4 to 6 hours; the reduction amplitude is the percentage difference between the initial value and the final value. For example, the area is reduced by about 57%, and the thickness is reduced by about 59%; the stability degree is characterized by the standard deviation of fluctuations and needs to be less than 5% of the threshold. These indexes are modeled by a support vector machine, using a radial basis kernel function, and the kernel parameter is optimized to 0.8 through cross-validation to generate an evaluation data set. Subsequently, a recursive neural network is used to predict the removal efficiency. The network is designed with 128 hidden layer units, and the parameter change sequence in the first 4 hours is input. The training data comes from 1000 sets of historical records. When optimizing, the momentum method is used to update the weights, the learning rate is 0.01, the momentum factor is 0.9, and the model is updated every 50 sets of new data to ensure that the prediction accuracy exceeds 90%, so as to realize the dynamic optimization of the correlation model and improve the adaptability of subsequent control.

[0062] In the embodiments of the present invention, continuous monitoring and data analysis not only verify the effect of scale inhibition, but also enhance the prediction ability of the model through a feedback mechanism. It can be foreseen that the sampling frequency and threshold setting can be adjusted according to the device operating environment to adapt to different scale types and treatment requirements, ensuring the high efficiency and stability of the cyclonite deep removal device during long-term operation.

[0063] S106. If, after adjusting the operating parameters of the cyclonite deep removal device, the monitored physical parameters of the inner wall still continuously deviate from the scale-free state reference value or the scale characteristic quantity shows an increasing trend, then the scale problem is evaluated as not alleviated through multi-dimensional data analysis, and a cleaning program is started to efficiently remove the scale layer.

[0064] In the embodiments of the present invention, the real-time monitoring of the physical parameters of the inner wall of the device is completed by relying on high-precision sensors. Among them, the temperature sensor covers a range of 0 to 200 degrees Celsius, with an accuracy of 0.1 degrees Celsius, and the sampling period is set to 15 seconds; the pressure sensor has a range of 0 to 10 MPa, with an accuracy of 0.01 MPa, and the sampling period is 5 seconds; the flow rate sensor has a measurement range of 0 to 15 m / s, with an accuracy of 0.1 m / s, and the sampling period is 2 seconds. The collected data is averaged with a 60-second sliding window, and after filtering out random fluctuations, it is compared with the scale-free state reference value to generate a difference sequence. The difference sequence calculates the comprehensive deviation index through weighted summation, and the weights are 0.4 for temperature, 0.3 for pressure, and 0.3 for flow rate. If the index exceeds 0.8, it indicates a serious deviation. For example, when the temperature deviation rate reaches 15%, the pressure deviation rate exceeds 20%, and the flow rate deviation rate drops by 25%, the comprehensive deviation significantly exceeds the warning range, indicating that the scale problem is intensifying.

[0065] S1061. Use scanning electron microscopy and image analysis techniques to monitor the dynamic changes of the scale characteristic quantity and evaluate the growth trend.

[0066] While monitoring the deviation of physical parameters, in the embodiment of the present invention, a scanning electron microscope is used to perform high-resolution imaging on the fouling area. The scanning area is set to 500 microns × 500 microns, and the resolution reaches 2048 × 2048 pixels, which can clearly capture the microscopic morphology of fouling. After the image is segmented by adaptive threshold, the fouling thickness and area are calculated. For example, the thickness increases from 25 microns to 85 microns, and the area expands from 150 square centimeters to 350 square centimeters. The change rate of characteristic quantities is calculated by the difference method, showing that the thickness growth rate reaches 2 microns per hour, and the area growth rate is about 8 square centimeters per hour, presenting an accelerating trend. To deeply analyze the fouling morphology, an 8-layer convolutional neural network is used to extract features. The input is a high-resolution image, and the output is a fouling state vector. The network is trained based on 1000 groups of samples, and an evaluation index is constructed by combining the change rate to confirm that the fouling problem is deteriorating, providing data support for subsequent cleaning decisions.

[0067] In the embodiment of the present invention, a bivariate correlation matrix is composed of a comprehensive deviation index and fouling characteristic quantities, and the development trend of fouling is predicted by a Markov chain. Matrix analysis reveals a high correlation between parameter deviation and fouling growth. The Markov chain has 5 state nodes, from no fouling to severe fouling. The transition probability is trained based on 2000 groups of historical data, and the predicted probability of fouling deterioration in the next 24 hours reaches 80%. If the mitigation evaluation value exceeds the threshold, the cleaning program is started, and instructions for the chemical cleaning agent ratio and the position of the mechanical cleaning head are generated. The pH value of the cleaning agent is controlled between 2.5 and 3.0, the pressure of the mechanical cleaning head is 2.5 MPa, and the rotation speed is 300 revolutions per minute to ensure both cleaning strength and safety.

[0068] S1062, realize the closed-loop control of the cleaning process through acoustic wave and turbidity monitoring and verify the removal effect.

[0069] During the cleaning execution, in the embodiment of the present invention, a 1 MHz acoustic wave sensor is used to monitor the peeling state of the fouling layer. The echo signal intensity weakens as the fouling decreases, reflecting the cleaning progress in real time. At the same time, an on-line turbidimeter detects the content of fouling substances in the cleaning liquid, which increases from the initial 50 mg / L to 350 mg / L and then gradually decreases, indicating that the fouling layer is gradually removed. The cleaning process lasts for 4 hours until the turbidity value tends to be stable, verifying the recovery of the heat transfer efficiency of the equipment. This closed-loop control method optimizes the cleaning parameters through dynamic feedback, avoids potential damage to the equipment caused by over-cleaning, and at the same time ensures the complete removal of fouling.

[0070] In the embodiment of the present invention, the above monitoring and cleaning processes effectively respond to the scenario of parameter adjustment failure and timely restore the equipment performance. It can be foreseen that the cleaning parameters and sensor configurations can be flexibly adjusted according to the fouling type to meet the requirements of different working conditions and improve the long-term operation efficiency of the RDX deep removal equipment.

[0071] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A control method for a device for deeply removing cyclonite in a magnesium nitrate solution, characterized in that, The method includes: Obtaining the operation data of the RDX deep removal device in real time, including the physical parameters of the inner wall of the device. The physical parameters include temperature, pressure, and flow rate physical parameters. At the same time, a physical parameter reference value under the fouling state is established. By comparing the physical parameters with the reference value, it is judged whether the device is fouled; If the device is fouled, the microscopic morphological characteristics of the fouling are obtained by scanning electron microscopy, the fouling distribution is identified and located, and the composition information of the fouling sample is obtained to obtain the composition, morphology, and distribution information of the fouling; According to the fouling information, a multi-scale numerical model is established to simulate the interaction between RDX and the fouling layer under the action of fluid, output the device state and the change of the fouling layer, establish a correlation model between the fouling characteristics and the removal efficiency, and predict the removal effect under different fouling conditions; Compare the physical parameters of the device monitored in real time with the removal effects predicted under different fouling conditions. If it is found that the operating parameters of the device deviate from the target range, adjust the operating parameters of the temperature, pressure, and flow rate of the device through a preset control system; Continuously monitor the change of the physical parameters of the inner wall of the device to judge whether the fouling is inhibited. If the monitored physical parameters tend to be stable or return to the reference value under the preset fouling-free state, and the fouling thickness or fouling area stops growing or begins to decrease, it is determined that the fouling is inhibited. Then evaluate the removal efficiency of RDX and feedback the evaluation result to the correlation model for optimization and update; If, after adjusting the operating parameters, the physical parameters of the inner wall of the monitored device continue to deviate from the reference value under the fouling-free state, or the fouling thickness or fouling area continues to grow, or even accelerates, it is evaluated that the fouling problem has not been alleviated after adjusting the operating parameters. Then start the preset device cleaning program to clean the inner wall of the device and remove the fouling layer.

2. The method according to claim 1, characterized in that, The obtaining the operation data of the RDX deep removal device in real time, including the physical parameters of the inner wall of the device. The physical parameters include temperature, pressure, and flow rate physical parameters. At the same time, a physical parameter reference value under the fouling state is established. By comparing the physical parameters with the reference value, it is judged whether the device is fouled, includes: Obtain the physical parameters from the temperature sensor, pressure sensor, and flow rate sensor, and periodically sample and record the physical parameters by the data acquisition module at a preset sampling time interval to obtain a sampling data sequence; Use the exponentially weighted moving average method to perform data smoothing processing on the sampling data sequence, calculate the statistical mean according to the smoothed physical parameters, and obtain the physical parameter reference value; Calculate the difference between the smoothed physical parameters and the physical parameter reference value, and calculate the standard deviation of the difference sequence by the least squares method to obtain the state parameter of the inner wall of the device; Compare the preset fouling state determination value with the state parameter of the inner wall of the device. If the state parameter of the inner wall of the device is greater than the fouling state determination value, it is determined that the RDX deep removal device is fouled.

3. The method according to claim 1, characterized in that, If the device is fouled, the microscopic morphological characteristics of the fouling are obtained by scanning electron microscopy, the fouling distribution is identified and located, and the composition information of the fouling sample is obtained to obtain the composition, morphology, and distribution information of the fouling, includes: The fouling surface is scanned and imaged in zones using a scanning electron microscope, and the high-resolution scanning electron microscope image of the fouling surface is obtained based on a preset scanning area size; The fouling area is identified using an image segmentation algorithm according to the high-resolution scanning electron microscope image, and the fouling distribution density map is obtained by calculating the proportion of the fouling area; The area with the maximum fouling density is selected from the fouling distribution density map, and the mass fractions of the elemental components in the fouling sample are obtained by elemental detection using an energy spectrometer; According to the mass fractions of the elemental components in the fouling sample, a three-dimensional reconstruction algorithm is used to perform three-dimensional modeling on the fouling area, and the fouling characteristic map is obtained by calculating the height information in the vertical direction.

4. The method according to claim 1, wherein A multi-scale numerical model is established based on the fouling information to simulate the interaction between cyclonite and the fouling layer under the action of fluid, output the equipment state and the change of the fouling layer, establish a correlation model between the fouling characteristics and the removal efficiency, and predict the removal effect under different fouling conditions, including: The fouling layer is meshed using the finite element method to obtain the discretized geometric structure data of the fouling layer; A fluid mechanics equation set is established according to the discretized geometric structure data, and the flow field distribution data at the fouling layer interface is calculated through the fluid mechanics equation set; A mass transfer equation is established according to the flow field distribution data, and the mass transfer flux data at the fouling layer interface is calculated through the mass transfer equation; The local dissolution rate of the fouling layer is calculated according to the interfacial mass transfer flux data, and the removal efficiency prediction data of the fouling layer is obtained through the local dissolution rate.

5. The method according to claim 1, wherein The physical parameters of the equipment monitored in real time are compared with the removal effects predicted under different fouling conditions. If it is found that the equipment operation parameters deviate from the target range, the temperature, pressure, and flow rate operation parameters of the equipment are adjusted through a preset control system, including: The operation parameter data is collected by temperature sensors, pressure sensors, and flow rate sensors, and the real-time operation state data of the equipment is obtained through moving average processing; The Manhattan distance is used to calculate the matching degree between the real-time operation state data of the equipment and the standard working condition database to obtain the temperature deviation data, pressure deviation data, and flow rate deviation data; A parameter adjustment sequence is established for the temperature deviation data, the pressure deviation data, and the flow rate deviation data, and the operation parameter adjustment direction is predicted through a long short-term memory network; The controller target value is set according to the operation parameter adjustment direction, and a proportional integral differential operator is used to optimize the control instruction sequence to obtain the corrected control instruction.

6. The method according to claim 1, wherein The change of the physical parameters of the inner wall of the equipment is continuously monitored to determine whether the fouling is inhibited. If the monitored physical parameters tend to be stable or return to the reference value under the preset fouling-free state, and the fouling thickness or fouling area stops growing or begins to decrease, it is determined that the fouling is inhibited, and then the removal efficiency of cyclonite is evaluated, and the evaluation result is fed back to the correlation model for optimization and update, including: The temperature parameter, pressure parameter, and flow rate parameter of the inner wall of the equipment are collected, and the physical parameter change rate is obtained according to the physical parameters using the exponentially weighted moving average method; Perform trend decomposition on the physical parameter sequence according to the change rate of the physical parameters, and calculate the root mean square error value through the trend decomposition result; Process the physical parameter sequence using an image segmentation algorithm, obtain the scaling thickness value and the scaling area value according to the image segmentation result, and obtain the change rate of the scaling characteristic quantity through differential operation on the scaling thickness value and the scaling area value; Set an inhibition determination threshold according to the root mean square error value. If the change rate of the scaling characteristic quantity is less than the inhibition determination threshold, it is determined that the scaling is inhibited.

7. The method according to claim 1, characterized in that, If, after adjusting the operating parameters, the physical parameters of the inner wall of the monitored device continuously deviate from the reference value in the non-scaling state, or the scaling thickness or the scaling area continuously increases, or even accelerates, it is evaluated that the scaling problem has not been alleviated after adjusting the operating parameters, and a preset device cleaning program is started to clean the inner wall of the device and remove the scaling layer, including: Obtain the physical parameters collected by the temperature sensor, the pressure sensor and the flow rate sensor, and calculate the difference sequence according to the physical parameters and the preset reference value; Obtain the scaling area image corresponding to the difference sequence through a scanning electron microscope, and process the scaling area image using an image segmentation algorithm to obtain the scaling characteristic quantity; Establish a bivariate correlation matrix according to the difference sequence and the scaling characteristic quantity, and calculate the scaling problem alleviation evaluation value through the bivariate correlation matrix; For the area where the scaling problem alleviation evaluation value exceeds the preset threshold, use an acoustic wave sensor to detect the peeling state of the scaling layer, and measure the scaling content in the cleaning liquid through an on-line turbidimeter to obtain the cleaning process control instruction.

Images