Production optimization control system of mining pipeline

Abstract

The invention discloses a mining pipeline production optimization control system, which belongs to the technical field of industrial automatic control, and comprises the following steps: collecting powder thermal characteristics and steel pipe base material parameters, constructing a preheating temperature model, setting a preheating adjustment method, calculating a target adjustment amount, dynamically adjusting heating power, and monitoring a heating curve in different time periods to judge a preheating state; constructing an injection parameter mapping model, setting an injection adjustment method, dynamically adapting to powder particle size fluctuation and pipeline curved surface change, identifying powder type switching, and calculating transition parameters according to batch weights for real-time correction; real-time data pairs are obtained, a leveling control method is set, a relation model is constructed, the leveling temperature and the constant temperature time are dynamically adjusted, segmented control is carried out, and graded correction is carried out based on leveling indexes during multi-variety switching; cross-linking activation energy is fitted, a hierarchical curing method is set, hierarchical control is carried out, cross-linking time is dynamically adjusted, dual-mode control real-time compensation is started during multi-variety switching, curing quality is monitored in real time, and errors are corrected in a hierarchical mode.

Description

TECHNICAL FIELD

[0001] The present application relates to a production optimization control system of mine pipeline, and belongs to the technical field of industrial automatic control. BACKGROUND

[0002] As a key component for fluid transportation in mines, the production quality and efficiency of mine pipelines affect the safety of mine production. The traditional production mode has problems such as setting the preheating temperature by experience, fixed spraying parameters, and lack of coupling control of melting and flow leveling, which cannot meet the modern mining precise control requirements.

[0003] In recent years, through the integration of Internet of Things, sensors and powder thermal analysis technology, by embedding distributed temperature sensors, online viscosity equipment and high-speed camera devices, the powder thermal properties and production data are collected, and a data interaction network is built relying on industrial Ethernet. Combined with the preheating temperature model driven by thermal analysis, the fluid dynamics spraying regulation and control model, and the solidification control model for different powders, the whole process chain is precisely controlled. At the same time, three-dimensional modeling and digital twin technology are integrated to simulate the coating preparation process, optimize the heating, spraying and solidification parameters, improve the coating performance, reduce powder loss and energy consumption, promote the transformation of mining production to fine and intelligent, and ensure the reliability of the pipeline under complex working conditions.

[0004] However, the existing technology is still limited by the dynamic difference of material properties, and faces multiple challenges: due to the difference in production process and raw material purity, the melting index and glass transition temperature of different batches of powder fluctuate by ±5℃ to ±10℃, resulting in fluctuations in the thermal stability of the powder batches. For example, when the melting point of a batch of epoxy powder decreases from the standard value of 180℃ to 175℃, if the traditional system still sets the preheating temperature at 180℃, the powder will not melt sufficiently on the surface of the steel pipe, forming a "granular" coating or adhesion defects. In addition, the sudden change of material properties when switching between multiple types of powder is difficult to handle, especially in the context of emergency order insertion. For example, thermoplastic PE and thermosetting epoxy powder have significantly different molecular reaction mechanisms. Even if the solidification process has self-adaptive regulation and control function, the lack of dynamic optimization of parameter transition strategy often leads to insufficient cross-linking or uncontrolled crystallinity of the intermediate batch of coating, causing continuous quality defects. SUMMARY

[0005] In view of the deficiencies of the prior art, the purpose of the present application is to provide a production optimization control system of mine pipeline, which adjusts the preheating temperature and spraying parameters by establishing a dynamic model based on real-time monitoring of powder thermal properties and substrate parameters, and optimizes the parameter transition through whole-process data linkage according to the difference in powder type, to solve the coating quality problems caused by the dynamic difference of material properties.

[0006] To achieve the above purpose, the present application provides the following technical solutions:

[0007] The production optimization control system of the mine pipeline comprises:

[0008] Collecting powder thermal characteristic parameters and steel pipe base material parameters, constructing a preheating temperature model, setting a preheating adjustment method, obtaining a target preheating temperature, calculating a target adjustment amount, and dynamically adjusting heating power, monitoring the temperature curve in time segments, and determining the preheating state through the target adjustment amount percentage and the temperature deviation;

[0009] Constructing a spraying parameter mapping model, setting a spraying adjustment method, dynamically adapting to powder particle size fluctuations and pipeline surface changes, identifying powder type switching, and calculating transition parameters according to batch weights for real-time correction, and through ultrasonic thickness closed-loop feedback and parameter iterative optimization;

[0010] Obtaining real-time data pairs, setting a flow leveling control method, constructing a relationship model, dynamically adjusting flow leveling temperature and constant temperature time, and controlling in segments, generating a gradient transition curve based on batch weight calculation of mixed activation energy when switching multiple varieties, and correcting according to flow leveling index classification;

[0011] Fitting the crosslinking activation energy, setting a hierarchical curing method, controlling and dynamically adjusting the crosslinking time, starting double-mode control for real-time compensation when switching multiple varieties, real-time monitoring of curing quality, and hierarchical correction of errors.

[0012] Preferably, the specific steps of constructing the preheating temperature model comprise:

[0013] Obtaining powder thermal characteristic parameters and steel pipe base material parameters, generating a multi-dimensional parameter group, and forming a multi-source data set through cumulative sample data;

[0014] Preprocessing the multi-source data set, removing outliers and standardizing, and labeling powder types according to DSC curve characteristics;

[0015] Constructing a preheating temperature model to output a target preheating temperature, and training the preheating temperature model using the multi-source data set.

[0016] Preferably, the specific steps of constructing the preheating temperature model further comprise:

[0017] Filtering and denoising the powder characteristic spectrum vector, reducing dimension through principal component analysis, calculating the Euclidean distance with existing characteristic vectors in the database and the difference rate of DSC curve characteristics;

[0018] If the Euclidean distance is greater than the preset distance threshold and the difference rate exceeds the preset proportion, it is determined as a new powder; otherwise, it is classified as an existing powder type;

[0019] The new powder data is added to the multi-source data set, triggering an incremental update mechanism, using transfer learning to retain the original weights, fine-tuning only the output layer parameters, and updating the preheating temperature model through incremental learning algorithm;

[0020] If the prediction error of a continuous batch exceeds the preset error value, a full update mechanism is triggered.

[0021] Preferably, the preheating adjustment method comprises:

[0022] The preheating temperature model is called, the current powder thermal characteristic parameters and the steel pipe substrate parameters are input, and the target preheating temperature is obtained;

[0023] The actual temperature of the substrate is obtained in real time by using a distributed temperature sensor, a heat conduction delay compensation coefficient is introduced, and the target adjustment amount ΔT aim is calculated.

[0024] The actual temperature of the substrate is collected in real time, the melting peak temperature of the current batch of powder is compared with the standard value, and if the fluctuation exceeds the preset temperature difference, the preheating temperature model is called again to update the target preheating temperature, and a time constant is introduced to correct the target adjustment amount;

[0025] The two-stage power switching threshold ΔT th.1 , ΔT th.2 , and ΔT th.1 > ΔT th.2 , and the heating power is dynamically adjusted.

[0026] If |ΔT aim | > ΔT th.1 , the maximum heating power of the heating device is started;

[0027] If |ΔT aim | ≤ ΔT th.2 , switch to the holding power;

[0028] If ΔT th.2 < |ΔT aim | ≤ ΔT th.1 , an incremental PID algorithm is used to calculate the heating power according to the target adjustment amount.

[0029] Preferably, the preheating adjustment method further comprises:

[0030] According to the heating power and the target adjustment amount, a substrate temperature rise curve is drawn, the heating process is divided into multiple heating periods according to time intervals, the target adjustment amount percentage of each heating period is extracted, and a target adjustment list is generated;

[0031] The calculated heating power is used to drive the actuator, and after the end of each heating period, the actual temperature of the substrate is obtained synchronously, the preheating actual completion rate and the temperature deviation are calculated.

[0032] If the actual completion rate is not less than 1 for two consecutive times and the temperature deviation is less than the preset deviation threshold, it is determined that the preheating is successful, and the spraying process is triggered; otherwise, the preheating is continued.

[0033] After the preheating is successful, the expected adjustment amount in the target adjustment list is read, the actual adjustment amount is obtained, the target difference ΔT is calculated, and the target difference threshold is set as ΔT th , to determine whether the current temperature adjustment is effective;

[0034] If ΔT> ΔT th , the three-level error tracing is started; otherwise, it is determined that the temperature adjustment is effective.

[0035] Preferably, the spraying adjustment method comprises:

[0036] The powder particle size distribution and dielectric constant are obtained by a laser particle size analyzer, and the pipeline geometric parameters are obtained by a three-dimensional scanner to construct a historical database;

[0037] Based on fluid dynamics and convolutional neural networks, a spraying parameter mapping model is constructed, the spraying parameter mapping model is trained by data in the historical database, and working parameters are output;

[0038] The spray gun is driven to perform a spraying operation, and a sensing data stream is obtained during the spraying process;

[0039] The target deposition density is calculated based on the electrostatic spraying adsorption theory, the actual deposition density is obtained by an infrared densitometer and an ultrasonic thickness gauge, and the absolute deviation and the relative deviation rate of the deposition are calculated respectively;

[0040] According to the pipe diameter and the curvature of the pipeline, the angle deflection amount and the moving rate of the spray gun are calculated;

[0041] A two-level deviation threshold is set, a hierarchical response strategy is established, and hierarchical correction is started based on the current operating state.

[0042] Preferably, the spraying adjustment method further comprises:

[0043] The powder molecular vibration spectrum is collected by an online infrared spectrometer, the characteristic peak change is identified, and the powder type switching is determined in combination with the sudden change of the thermal gravimetric analysis curve slope;

[0044] Order information is obtained, and once an emergency order insertion is detected, a switching batch number is automatically assigned, and a parameter transition value is calculated according to the batch weight;

[0045] After spraying a pipeline of a preset length, the thickness distribution standard deviation is obtained, and once the thickness distribution standard deviation exceeds a preset thickness distribution value, the tracing is started, and the parameters are adjusted according to the correction priority;

[0046] The correction parameter is fed back to a target deposition density calculation step to update the reference process coefficient periodically.

[0047] Preferably, the flow leveling control method comprises:

[0048] The powder melt viscosity is monitored by an online viscosity sensor, and the coating temperature field distribution is obtained by an infrared thermal imaging device to generate real-time data pairs.

[0049] The Arrhenius equation is used to describe the viscosity-temperature mapping relationship, and the flow activation energy and pre-exponential factor are fitted by the least squares method to construct a relationship model.

[0050] When the measured data deviates from the standard curve beyond the preset range, it is determined that the corresponding relationship deviates, triggering logarithmic transformation and linear fitting, recalculating the flow activation energy and pre-exponential factor and updating the relationship model parameters.

[0051] The softening interval parameters are obtained, the heating rate is calculated, and the heating device is driven;

[0052] Based on the relationship model, the optimal flow leveling temperature and constant temperature time are calculated to generate control instructions.

[0053] For thermoplastic powder, the cooling rate is calculated according to the crystallization onset temperature and activation energy, and the cooling device is driven.

[0054] The batch weight of spray adjustment is obtained, the mixing activation energy is calculated, the gradient transition curve is generated, and the constant temperature time is adjusted.

[0055] The flow leveling index is calculated, and based on the preset secondary threshold, the hierarchical correction is triggered, and the heating rate is adjusted preferentially and the parameters are calibrated.

[0056] Preferably, the hierarchical solidification method comprises:

[0057] The powder melting peak temperature, crosslinking exothermic peak characteristics, molecular vibration spectrum, and melting index are collected in real time, and the crosslinking activation energy is fitted based on the DSC exothermic curve to update the Arrhenius equation parameters.

[0058] The solidification process is divided into gelation, crosslinking expansion, and post-curing stages, and the solidification parameters of the three stages are calculated respectively.

[0059] The first-stage cooling promotes crystal nucleus formation, the isothermal crystallization section dynamically calculates the constant temperature time, and the second-stage cooling controls the crystallization stress.

[0060] When multiple varieties are switched, a dual-mode control of the original parameter base and the target parameter increment is adopted, and when the crosslinking degree of the transition batch is insufficient, the temperature is automatically compensated.

[0061] The solidification quality is monitored in real time, the solidification deviation is calculated, and through the preset secondary solidification deviation threshold, hierarchical correction is performed.

[0062] Advantages of the present application:

[0063] Based on the real-time collected powder thermal properties, pipe geometry parameters and deposition data, the model fits the nonlinear relationship, combines the hierarchical correction and dynamic updating mechanism, accurately controls the preheating temperature, spraying parameters, leveling and solidification process; when the powder particle size fluctuates, the activation energy changes or the pipe curvature is different, the parameter self-correction and gradient transition are automatically triggered to avoid the problems of insufficient crosslinking and uneven deposition caused by traditional fixed parameters; through real-time monitoring and closed-loop feedback of infrared spectrum, ultrasonic thickness and the like, the rapid identification and iterative optimization of coating quality defects are realized, the qualified rate of intermediate batches is improved when multiple varieties of powders are switched, the waste rate is reduced, and the performance stability of the mine pipeline coating is ensured. BRIEF DESCRIPTION OF DRAWINGS

[0064] Fig. 1 Figure 1 is a structure diagram of a production optimization control system for a mine pipeline;

[0065] Fig. 2 Figure 2 is a flowchart for constructing a preheating temperature model of the present application;

[0066] Fig. 3 Figure 3 is a flowchart of a preheating adjustment method of the present application;

[0067] Fig. 4 Figure 4 is a flowchart of a leveling control method of the present application. DETAILED DESCRIPTION

[0068] The technical scheme of the present application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, and are not limitations of the technical scheme of the present application. In the case of no conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0069] Reference Figs. 1 to 4 The present embodiment introduces a production optimization control system for a mine pipeline, which comprises a preheating decision module, a spraying optimization module, a leveling control module and a solidification module.

[0070] The preheating decision module is used to collect powder thermal characteristic parameters and steel pipe base material parameters through various devices, construct a preheating temperature model, combine the nonlinear relationship between the powder thermal characteristic parameters and the steel pipe base material parameters to realize intelligent calculation of the target temperature, avoid coating quality defects caused by molecular structure changes through new powder identification and model incremental updating mechanism, set a preheating adjustment method, output the target preheating temperature based on the preheating temperature model, monitor the actual temperature in real time through a distributed temperature sensor, combine the heat conduction delay compensation and the equipment thermal inertia correction to calculate the target adjustment amount, dynamically adjust the heating power by using the two-stage power switching threshold and the incremental PID algorithm, reduce the energy waste and temperature overshoot of constant power heating, solve the preheating temperature misalignment problem, monitor the temperature rise curve in time segments, construct a target adjustment list through the target adjustment percentage, determine whether the preheating is successful through the actual completion rate and the temperature deviation double condition to trigger the spraying process, ensure that the base material surface temperature and the powder melting demand are accurately matched, cooperate with the error traceability mechanism to ensure the preheating precision, solve the problems of insufficient cross-linking or uncontrollable crystallinity caused by the lack of dynamic optimization of the parameter transition strategy during the switching of multiple varieties of powder, and ensure the stability of the coating quality;

[0071] The spraying optimization module is used to construct a powder characteristic and pipe geometry parameter database through multi-source data acquisition, combine fluid dynamics and convolutional neural network to construct a spraying parameter mapping model, set a spraying adjustment method, dynamically adapt to powder particle size fluctuations and pipe surface changes, solve the problem of large deposition density uniformity error caused by traditional fixed parameter spraying, realize the reduction of deposition density uniformity error, simultaneously identify powder type switching, dynamically calculate transition parameters according to batch weight and real-time correction, solve the parameter transition lag problem during the switching of multiple varieties of powder, improve the intermediate batch coating thickness compliance rate, in addition, through ultrasonic thickness closed-loop feedback and parameter iterative optimization, reduce the waste rate during the switching of multiple varieties of powder, effectively deal with the deposition defects caused by sudden changes in material properties;

[0072] The leveling control module is used to obtain real-time data pairs, set a leveling control method, construct a relationship model, dynamically adjust the leveling temperature and constant temperature time to ensure that the melt is in the best leveling interval, simultaneously control the rapid heating section to shorten the low temperature stay to prevent powder caking, control the crystallinity of thermoplastic powder through gradient cooling, calculate the mixed activation energy based on the batch weight during the switching of multiple varieties to generate a gradient transition curve, compensate and adjust the leveling parameters to cope with sudden changes in material properties, and according to the leveling index grading correction, calibrate the sensor or fine-tune the parameters, improve the coating uniformity, reduce the leveling defect rate, and ensure the quality of the intermediate batch during the switching of multiple varieties;

[0073] The solidification module is used for setting a staged solidification method by using DSC and infrared spectrum fitting crosslinking activation energy, dynamically adjusting crosslinking time, solving crosslinking deficiency caused by activation energy fluctuation, segmentally cooling thermoplastic powder according to crystallization activation energy, controlling crystallinity in a target range, generating a gradient transition curve according to batch weight when multiple varieties are switched, starting double-mode control and real-time compensation, solving intermediate batch defects caused by parameter transition lag, simultaneously monitoring solidification quality in real time through an infrared densitometer, correcting errors in stages, solving defect discovery lag problem, reducing waste rate, and guaranteeing coating quality stability.

[0074] Preferably, the specific steps of constructing the preheating temperature model include:

[0075] Differential scanning calorimeter (DSC) is used to obtain powder melting peak temperature T m , glass transition temperature T g , melt flow rate tester is used to test melt index, near-infrared spectrometer is used to generate powder characteristic spectrum vector, eddy current sensor is used to detect carbon content in steel pipe, ultrasonic thickness gauge and laser thermal conductivity instrument are used to obtain pipe diameter and thermal conductivity respectively, energy conservation formula is used to calibrate heating equipment thermal efficiency, multi-dimensional parameter group is generated based on the above parameters, and multi-source data set is formed by accumulating sample data;

[0076] The data in the multi-source data set is preprocessed, abnormal values are removed based on the 3σ principle, and parameters are standardized by Z-Score to eliminate dimension influence, and the powder type is automatically labeled according to the DSC curve characteristics, wherein the thermoplastic powder has no exothermic platform after the melting peak, and is labeled as 0, and the thermosetting powder has obvious crosslinking exothermic peak, and is labeled as 1;

[0077] Because the BP neural network has strong fitting ability for nonlinear relationship, it can capture the complex coupling relationship between powder thermal characteristics and substrate parameters, a preheating temperature model is constructed based on a three-layer BP neural network, including an input layer, a hidden layer and an output layer, wherein the input layer is used to receive the multi-source data set, the hidden layer is used to fuse the powder type characteristics and the thermal conduction delay characteristics, the hidden layer is activated in two stages, first, the ReLU function is used to fuse the powder type characteristics, and then the LeakyReLU function is used to introduce the thermal conduction delay characteristics, and the output layer generates the target preheating temperature of the substrate surface through linear activation, and the preheating temperature model is trained using the multi-source data set, the mean square error and L1 regularization are combined to determine the loss function, the Adam optimizer and the cosine annealing learning rate decay strategy are combined to improve the generalization ability;

[0078] Since the molecular structure and thermal decomposition behavior of the new powder change significantly, the temperature mapping relationship of the original model is invalid. If the preheating temperature is not updated, the preheating temperature will still be set according to the existing setting, which will lead to insufficient melt flowability and high error of coating thickness uniformity. The characteristic spectrum vector of the powder is filtered and denoised by Savitzky-Golay filtering, and the dimension is reduced to an 8-dimensional characteristic vector by principal component analysis. The Euclidean distance between the characteristic vector and the existing powder characteristic vector in the database is calculated, and the difference rate of the DSC curve characteristics is also calculated.

[0079] If the Euclidean distance is greater than the preset distance threshold and the difference rate exceeds the preset difference proportion, such as the change of the melting peak shape, it is determined that the new powder is a new powder; otherwise, it is classified as an existing powder type.

[0080] Once the new powder is determined, the thermal characteristics of the new powder and the substrate parameters are collected, a number of training samples are generated, and the data of the new powder is added to the multi-source data set for training. The incremental update mechanism is triggered, the transfer learning strategy is adopted, the original weight is retained, only the output layer parameters are fine-tuned, and the preheating temperature model is updated by incremental learning algorithm to avoid catastrophic forgetting.

[0081] If the prediction error of the last 5 batches exceeds the preset error value, the full update mechanism is triggered, the weight of the new powder data is set to 1.5, the weight of the historical data is set to 0.8 to balance the influence of new and old samples, and the first training of the new powder is set to 200 rounds, and the subsequent incremental update is 50 rounds / time to ensure the long-term adaptability of the model to the dynamic difference of material characteristics and solve the coating quality defects caused by parameter lag. Wherein, the absolute error between the target preheating temperature output by the preheating temperature model and the actual measured temperature of the substrate surface after preheating is calculated, and at this time the ratio of the absolute error to the measured temperature is defined as the prediction error, which reflects the proportion of the error relative to the actual temperature, and is suitable for error comparability in different temperature intervals.

[0082] Preferably, the preheating adjustment method comprises:

[0083] The preheating temperature model is called, the current powder thermal characteristic parameters and the steel pipe substrate parameters are input, and the target preheating temperature T aim in the current state is obtained. act The actual temperature T aim of the substrate is obtained in real time by using a distributed temperature sensor to form a basic temperature difference. At the same time, if the heat conduction delay of the substrate is not considered, such as the temperature difference between the inner and outer walls when the thick-walled steel pipe is heated, the adjustment amount and the actual demand will be deviated. By introducing a heat conduction delay compensation coefficient, the heat conduction lag deviation of the thick-walled steel pipe is corrected, the problems of excessive or insufficient heating caused by the fluctuation of the characteristics of the powder batches (such as the shift of the melting peak) are solved, and the formation of granular coating due to insufficient melting is avoided. The target adjustment amount ΔT aim is calculated based on the heat conduction delay compensation coefficient a. aim The expression is as follows:

[0084] ΔTaim = a x (T aim - T act )

[0085] To avoid excessive or insufficient temperature rise caused by powder batch characteristics fluctuation (such as melting peak shift), the actual temperature of the base material is collected in real time by a distributed temperature sensor, updated every 10 seconds, and compared with the standard value of the current batch of powder melting peak temperature. If the fluctuation exceeds the preset temperature difference, the target preheating temperature is recalculated by calling the preheating temperature model, considering the thermal inertia of the heating equipment, introducing the time constant τ, combining the remaining heating time, to correct the target adjustment amount, as shown in the following expression:

[0086] ΔT aim.new = ΔT aim.old x (1 - e -t / τ )

[0087] In the formula, ΔT aim.old , ΔT aim.new are the target adjustment amounts before and after correction, t is the remaining heating time, which is the difference between the total preset heating time and the heated time, used to predict the temperature change trend at different times, and the time constant τ is obtained by step response test;

[0088] Since constant power heating easily leads to energy waste and serious temperature overshoot in the heating stage, the two-stage power switching threshold values are set as ΔT th.1 , ΔT th.2 , and ΔT th.1 > ΔT th.2 , to dynamically adjust the heating power under different conditions;

[0089] If |ΔT aim | > ΔT th.1 , start the maximum heating power to quickly heat, then P = P max ; where P is the actual heating power of the heating equipment, and P max is the maximum heating power of the heating equipment;

[0090] If |ΔT aim | ≤ ΔT th.2 , switch to the holding power to maintain temperature stability, that is, P = P keep ; the expression is as follows:

[0091] P keep = η -1 x m x c x ΔT samll

[0092] In the formula, P keep is the holding power of the heating equipment, η is the thermal efficiency, m is the mass of the steel pipe, c is the specific heat capacity, and ΔT samllThe minimum adjustment amount;

[0093] If ΔT th.2 <|ΔT aim |≤ΔT th.1 , the incremental PID algorithm is adopted to calculate the heating power according to the target adjustment amount, avoiding energy waste and temperature overshoot caused by constant power heating, and the expression is as follows:

[0094]

[0095] In the formula, K p , K i , K d are adjustment parameters, K p is used to control the response speed, K i is used to eliminate static error, and K d is used to suppress overshoot.

[0096] Because the one-time temperature rise cannot control the temperature rise rate, it is easy to cause the powder to stay in the softening interval for too long, and monitoring by time period helps to accurately control the powder residence time. According to the heating power and the target adjustment amount, the temperature rise curve of the base material is drawn to intuitively show the relationship between time and adjustment amount. The horizontal coordinate is time, and the vertical coordinate is the expected adjustment amount at the corresponding time point. According to the time interval Δt aim , the heating process is divided into several heating periods, and the target adjustment percentage β j of each heating period is extracted, such as 40% for the first segment, 60% for the second segment, and 100% for the last segment, so as to obtain the expected adjustment amount ΔT aim.j corresponding to each heating period. A target adjustment list is generated; wherein β j is the target adjustment percentage corresponding to the jth heating period, and ΔT aim.j is the expected adjustment amount corresponding to β j .

[0097] According to the calculated heating power, the actuator is driven, and after the jth heating period ends, the corresponding actual temperature T real.j is synchronously obtained, the actual completion rate γ j of preheating and the temperature deviation ΔT j are calculated, and the expression is as follows:

[0098]

[0099] ΔT j =|T real.j -T aim |

[0100] If the actual completion rate of two consecutive times is not less than 1 and the temperature deviation is less than the preset deviation threshold, it is determined that the preheating is successful, the spraying process is triggered, and the powder softening interval is avoided to stay too long due to uncontrollable temperature rising rate; otherwise, the preheating operation is continued;

[0101] After the preheating is successful, the corresponding expected adjustment amount ΔT is read from the target adjustment list aim , and the actual adjustment amount ΔT real is obtained, the target difference ΔT is calculated, and the target difference threshold is set as ΔT th to determine whether the current temperature adjustment is effective; the expression is as follows:

[0102] ΔT = | ΔT real - ΔT aim |

[0103] If ΔT > ΔT th , the current temperature adjustment has an error, the error tracing is started, and the three-level inspection of the sensor, the preheating temperature model and the heating equipment is performed to avoid the preheating temperature misalignment caused by the change of the base material parameters; otherwise, the current temperature adjustment has no error, which means that the current temperature adjustment is effective.

[0104] Preferably, the spraying adjustment method comprises:

[0105] The powder particle size distribution and the dielectric constant are obtained by the laser particle size analyzer, the geometric parameters of the pipeline are obtained by the three-dimensional scanner, including the pipe diameter, the curvature radius and the elbow angle, and the historical database is constructed based on the collected historical data;

[0106] Since the fixed parameter spraying cannot adapt to the powder particle size fluctuation and the pipeline curved surface change, the deposition density uniformity error is large, the spraying parameter mapping model is constructed based on the fluid dynamics and the convolutional neural network, including the establishment of the spraying flow field basic equation by using the Reynolds number, the introduction of the convolutional neural network to process the high-speed camera data, the identification of the powder deposition density distribution, the output of the parameter correction coefficient, and the input of the data in the historical database to the spraying parameter mapping model for training, the working parameters of the electrostatic spray gun are output, and a dynamic updating mechanism is set, once the powder particle size fluctuation exceeds 15% or the pipeline curvature changes by not less than 30°, such as the powder D50 increases from 50 μm to 60 μm, the 90° elbow of the pipe diameter 159 mm, the model parameter self-correction is triggered, and the genetic algorithm is used to optimize the CFD simulation boundary conditions;

[0107] The working parameters output based on the jet parameter mapping model drive the electrostatic spray gun to perform a spraying operation, and during the spraying process, a high-speed camera is used in combination with a backlight source to capture the motion trajectory of the powder particles, the particle displacement between adjacent frames is calculated through an optical flow method, the particle speed is obtained by combining the frame rate and the spraying distance, and a speed distribution array is output. Meanwhile, to reduce the interference of reflected light, a blue LED backlight source is used, and a YOLOv5 model is used to identify the particles. A piezoresistive pressure sensor is arranged at a distance from the outlet of the spray gun to measure the dynamic pressure at the outlet of the spray gun, and the dynamic pressure is temperature-compensated. A static induction probe is used to synchronously monitor the charge density. The above data are combined to form a timestamp-aligned sensor data stream, which provides basic physical quantities for subsequent deviation calculation.

[0108] Based on the electrostatic spraying adsorption theory, the target deposition density ρ is calculated by comprehensively considering the powder dielectric constant, voltage, particle size and spraying distance. aim Meanwhile, an infrared densitometer and an ultrasonic thickness gauge are used to obtain the actual deposition density ρ. real The absolute deviation Δρ and the relative deviation rate δ of the deposition are calculated. ρ

[0109] Since the smaller the pipe curvature is, the greater the spray gun angle deflection is and the lower the moving speed is, to ensure uniform deposition of each part of the curved surface, once the pipe curvature radius is less than 2 times the pipe diameter, the spray gun angle deflection θ is automatically calculated adj , and the spray gun moving speed is simultaneously reduced to v adj to avoid powder accumulation at the edge of the curved surface and sparseness at the center, and the servo motor is controlled to adjust the spray gun posture and moving speed. The expression is as follows:

[0110]

[0111] In the formula, D is the pipe diameter, R is the curvature radius, v0 is the initial moving speed, R min is the critical curvature radius, and R min = 2D.

[0112] Two-level deviation threshold values δ ρ.th1 and δ ρ.th2 are set, and δ ρ.th1 < δ ρ.th2 , a hierarchical response strategy is established, including first-level correction, second-level correction and third-level correction, and the hierarchical correction is started based on the current operation state. When δ ρ ≤ δ ρ.th1 , the first-level correction is started, only the voltage is adjusted, and the voltage adjustment amount ΔU is calculated. When δ ρ.th1 < δ ρ ≤ δ ρ.th2 , the second-level correction is started, the voltage and the pressure are simultaneously adjusted, the voltage adjustment amount ΔU and the pressure adjustment amount ΔV are calculated. When δ ρ > δ ρ.th2 ​, start the third level correction, start the spray gun trajectory re-planning, calculate the spray gun angle deflection combined with the pipe curvature; The expression is as follows:

[0113]

[0114] ΔV=k V ×Δρ

[0115] In the formula, k U is the pressure correction coefficient, k V is the pressure correction coefficient, and ε is the powder dielectric constant;

[0116] Collect powder molecular vibration spectrum by online infrared spectrometer, identify characteristic peak change, combined with thermal analysis curve slope mutation, determine powder type switching; Once the spectral similarity of the characteristic peak is less than 80% and the weight loss rate changes more than 30%, it is marked as type switching to be processed;

[0117] Get order information, if emergency order insertion is detected, such as epoxy powder order queuing PE production, automatically promote switching priority, start fast transition mode, and automatically assign batch number according to order priority and production order, for example, orders switching from PE to epoxy will be marked as switching batch 1, and subsequent orders of the same type will be switching batch 2, switching batch 3, and once the powder type in the order is inconsistent with the type being processed in the current production line, a new switching batch number is automatically generated;

[0118] Based on the switching batch number, such as the first batch number, the original powder and target powder characteristic parameters are obtained, and the transition value of each type of parameter is calculated according to the batch weight, such as in the first batch production, the weight is set to 50%, and the first batch parameter is transitioned from the original powder parameter to the target powder characteristic parameter by 50% proportion, such as 50% of the voltage from PE's 60kV to epoxy's 75kV, that is, 67.5kV, and the subsequent batch is gradually adjusted according to the weight, while the intermediate batch is monitored and parameter corrected in real time, the deposition density is monitored in real time by infrared densitometer, and once the epoxy powder is insufficient due to insufficient charging, the voltage compensation and pressure compensation are automatically triggered, combined with the online infrared monitoring result dynamic adjustment, to avoid intermediate batch quality defects, solve the parameter switching lag problem, and improve the intermediate batch coating thickness compliance rate;

[0119] After spraying a certain length of pipe, the thickness distribution standard deviation is obtained by using an ultrasonic thickness gauge, and if the thickness distribution standard deviation exceeds the preset thickness distribution value, the source is started, the correction priority is set as powder characteristic fluctuation, equipment parameter deviation, and trajectory accuracy, the powder characteristic data is checked first, then the pressure sensor and electrostatic voltage source are calibrated, and finally the spray gun trajectory is fine-tuned;

[0120] The step of driving the injection actuator based on the corrected parameters to make parameter adjustment and feeding the corrected parameters to the target deposition density is iteratively updated, and after completing 10 groups of effective correction, the latest target deposition density p aim The measured deposition density p real The data is refitted with the process coefficient, the reference process coefficient is updated, the prediction accuracy of the subsequent batch parameters is improved, the waste rate during the switching of multiple varieties of powder is reduced, and the deposition defects caused by sudden changes in material properties are effectively addressed; The expression is as follows:

[0121]

[0122] k=k0x[1+0.02(MI-10)]

[0123] In the formula, k is the process coefficient, which is dynamically corrected by the powder melting index, U is the electrostatic voltage, D m is the median particle size of the powder, L is the spraying distance, k0 is the reference process coefficient, the initial value is 0.8, and MI is the powder melting index.

[0124] Preferably, the flow leveling control method comprises:

[0125] The powder melt viscosity is continuously monitored by an online viscosity sensor, the coating temperature field distribution is obtained synchronously by using an infrared thermal imaging device, real-time data pairs are generated, the real-time mapping relationship between viscosity and temperature is described by using the Arrhenius equation, the flow activation energy E a and the pre-exponential factor η0 are fitted by the least square method as the input of subsequent segmented control to construct a relationship model for describing the viscosity and temperature relationship of the current powder; The initial parameters of the relationship model are usually obtained by inverse calculation based on a standard curve, and the standard curve is stored in a history library for defining a normal fluctuation range, and the viscosity and temperature standard curves of different powder varieties are stored in the history library;

[0126] Once the deviation of the measured data points in the real-time data set from the standard curve exceeds the preset range, it is determined that the corresponding relationship deviates. Due to the differences in molecular chain structure, particle size distribution or additive content of different batches of powder, the flow activation energy and the pre-exponential factor are likely to change, causing the measured relationship to deviate from the standard curve. At this time, the model parameter refitting mechanism is triggered, the flow activation energy and the pre-exponential factor are calculated by logarithmic transformation and linear fitting, and the relationship model parameters are dynamically updated. At this time, the model curve deviates from the standard curve, but is closer to the real characteristics of the current powder, so that the relationship model can adapt to the flow characteristics of the current powder in real time, ensuring the calculation accuracy of the subsequent flow leveling temperature, constant temperature time and other parameters, and avoiding flow leveling defects caused by parameter lag;

[0127] The softening interval parameters between batches are obtained to determine the starting and ending temperatures of the temperature rise, and the optimal temperature rise rate vT1 and drive the heating device to perform rapid heating section control to shorten the residence time in the low-temperature softening zone and avoid powder caking; the expression is as follows:

[0128]

[0129] wherein k T1 is an empirical correction coefficient related to the heating efficiency of the device and the bulk density of the powder, calibrated through historical process data, T g is the glass transition temperature, indicating the transition temperature of amorphous material from the glass state to the high-elastic state, is the powder softening temperature, indicating the critical temperature of the transition from solid to liquid, t soft is the softening time reference;

[0130] Based on the parameters in the relationship model, the optimal leveling temperature T otp and the constant temperature time t hold are calculated to generate control instructions for precise control of the optimal leveling section, ensuring that the melt viscosity is in the optimal leveling interval, and the liquid droplets are driven to merge by surface tension to eliminate pores; the expression is as follows:

[0131]

[0132] wherein Θ is the gas constant, η opt is the optimal leveling viscosity, ι is the target leveling degree, n is the leveling reaction order, k hold is the scale parameter determined by the inherent characteristics of the material, determined through experiments or modeling, and A is the reaction pre-exponential factor related to the reaction kinetics, and exp(·) represents the exponential function;

[0133] For thermoplastic powder, the gradient cooling rate v cool is calculated according to the crystallization onset temperature and the activation energy to promote the ordered arrangement of molecular chains and control the crystallinity within the target range, thereby generating an adjustment signal to drive the cooling device; the expression is as follows:

[0134]

[0135] wherein E a_c is the crystallization activation energy, indicating the energy required for molecular arrangement in the crystallization process, T c is the crystallization onset temperature, k c is the crystallization rate constant;

[0136] The corresponding batch weight in the jet adjustment method is obtained synchronously, the mixed activation energy E a_mix is calculated, the leveling temperature curve of the gradient transition is generated, and the constant temperature time is recalculated, and the leveling parameters of the transition batch are compensated and adjusted in combination with the real-time viscosity monitoring result; the expression is as follows:

[0137] E a_mix = w con E a_1 + (1-w con )E a_2

[0138]

[0139] wherein E a_1 , E a_2 are the activation energies of the original powder and the target powder respectively, the equivalent melting point of the mixed powder, also calculated by batch weight, w con is the batch weight, representing the proportion of the original powder parameters in the current batch during the switching process, t hold.min is the flow leveling constant temperature time of the transition batch, t hold.old is the standard flow leveling constant temperature time of the original powder, T a_mix is the equivalent melting point of the mixed powder;

[0140] The temperature field uniformity is evaluated using infrared thermal imaging, the coating roughness is monitored by laser confocal microscope, quality evaluation data is generated, the flow leveling index is calculated, once the flow leveling index is less than the preset value, the correction is triggered, the heating rate is adjusted preferentially, and the hierarchical response is performed, by setting the secondary flow leveling threshold, the first error, the second error and the third error are divided, the activation energy is refitted when the first error, the sensor is calibrated when the second error, and the control parameters are fine-tuned when the third error.

[0141] Preferably, the hierarchical curing method comprises:

[0142] Since the fixed parameters cannot adapt to the fluctuation of activation energy between batches, leading to insufficient crosslinking or overcuring, the DSC differential scanning calorimeter and infrared spectrometer are used to collect the powder melting peak temperature, crosslinking exothermic peak characteristics and molecular vibration spectrum in real time, the melt flow rate tester is used to test the melt index at the same time, and based on the crosslinking reaction kinetics equation, the crosslinking activation energy is fitted through the DSC exothermic curve, the Arrhenius equation parameters are updated, and the traditional fixed parameters leading to insufficient melting are avoided;

[0143] To cope with the fluctuation of crosslinking activation energy between batches, avoid under-crosslinking or over-crosslinking caused by fixed parameters, for thermosetting powder, based on real-time crosslinking activation energy, the curing process is divided into three stages, including gelation stage, crosslinking expansion stage and post-curing stage, and the curing parameters of the three stages are calculated respectively, such as the gelation stage is heated to the glass transition temperature plus 10 DEG C, and the preliminary crosslinking is promoted by holding, the crosslinking expansion stage is dynamically adjusted according to the optimal crosslinking temperature and time formula to update the holding time, when the crosslinking activation energy increases, the holding time is automatically extended, the under-crosslinking caused by the difference of activation energy between batches is solved, and the crosslinking degree is stabilized in the target interval, and the post-curing stage is monitored by an infrared spectrometer in real time to compensate the holding time, so as to eliminate internal stress;

[0144] For thermoplastic powder, segmented cooling is implemented according to the crystallization activation energy and Avrami index, primary cooling is promoted to form crystal nucleus, and the holding time is dynamically calculated according to the activation energy in the isothermal crystallization section, secondary cooling is controlled to control crystallization stress, the crystallinity is controlled in the target range to solve the problem of uncontrolled crystallinity caused by the fluctuation of melt index;

[0145] When multiple varieties are switched, the batch weight and mixed activation energy are read, and the dual-mode control is started, including the original parameter base and the target parameter increment, the holding temperature is automatically compensated when the crosslinking degree of the transition batch is insufficient, the parameter transition lag problem is solved, and the intermediate batch qualification rate is improved;

[0146] In each preset time period, the curing quality is monitored in real time by an infrared spectrum and a densimeter, and the curing deviation is calculated by a preset curing standard, and the curing deviation is corrected based on the curing deviation, the second curing deviation threshold is set, the first error is fitted to the activation energy, the second error is calibrated to the instrument, and the third error is fine-tuned to the parameters, so that the defect discovery lag problem is solved, and the waste rate is reduced.

[0147] In the above embodiments, the powder thermal properties and steel pipe parameters are collected by DSC, infrared spectrum and other equipment, the preheating temperature model and the injection parameter mapping model are constructed, the intelligent calculation is realized by combining BP neural network and convolutional neural network, the data are obtained by using distributed temperature sensors, high-speed cameras and the like, the preheating power is adjusted by heat conduction delay compensation and PID algorithm, the injection parameters are dynamically adjusted by fluid dynamics and hierarchical response strategy, and the temperature is controlled in sections according to the viscosity-temperature relationship model, the curing is adjusted in stages according to the activation energy, the gradient transition curve is generated according to the batch weight when multiple varieties are switched, the infrared monitoring and ultrasonic thickness measurement are combined to realize closed-loop feedback, and the process coefficients are iteratively updated.

[0148] The above merely describes the preferred embodiments of the present application, and the protection scope of the present application is not limited to the above-described embodiments. Any technical solution falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principles of the present application shall also be considered as falling within the protection scope of the present application.

Claims

1. A production optimization control system for mining pipelines, characterized in that, include: Collect powder thermal characteristic parameters and steel pipe substrate parameters, construct a preheating temperature model, set a preheating adjustment method, obtain the target preheating temperature, calculate the target adjustment amount, dynamically adjust the heating power, monitor the temperature rise curve in different time periods, and determine the preheating status by the percentage of the target adjustment amount and the temperature deviation. A spraying parameter mapping model is constructed, a spraying adjustment method is set, and the powder particle size fluctuation and pipeline surface change are dynamically adapted. The powder type switching is identified, and the transition parameters are calculated according to batch weight for real-time correction. The parameters are optimized through ultrasonic thickness measurement closed-loop feedback and parameter iteration. Acquire real-time data pairs, set up leveling control methods, construct a relational model, dynamically adjust leveling temperature and constant temperature time, and implement segmented control. When switching between multiple varieties, calculate the mixing activation energy based on batch weight to generate a gradient transition curve, and make graded corrections according to the leveling index. The crosslinking activation energy is fitted, a staged curing method is set, the crosslinking time is controlled and dynamically adjusted in stages, dual-mode control is activated for real-time compensation when switching between multiple varieties, the curing quality is monitored in real time, and the error is corrected in stages.

2. The production optimization control system for mining pipelines according to claim 1, characterized in that: The specific steps for constructing a preheating temperature model include: Obtain powder thermal property parameters and steel pipe substrate parameters, generate multi-dimensional parameter sets, and form a multi-source dataset by accumulating sample data; The multi-source dataset is preprocessed to remove outliers and standardize the data, while the powder type is labeled according to the DSC curve characteristics. A preheating temperature model is constructed to output the target preheating temperature, and the preheating temperature model is trained using the multi-source dataset.

3. The production optimization control system for mining pipelines according to claim 2, characterized in that: The specific steps for constructing a preheating temperature model also include: The powder feature spectral vectors are filtered and denoised, and dimensionality is reduced by principal component analysis. The Euclidean distance and the difference rate of DSC curve features with existing feature vectors in the database are calculated. If the Euclidean distance is greater than a preset distance threshold and the difference rate exceeds a preset ratio, it is determined to be a new powder; otherwise, it is classified as an existing powder type. New powder data is added to the multi-source dataset to trigger an incremental update mechanism. Transfer learning is used to retain the original weights and only the output layer parameters are fine-tuned. The preheating temperature model is updated through an incremental learning algorithm. If the prediction error of consecutive batches a exceeds the preset error value, a full update mechanism will be triggered.

4. The production optimization control system for mining pipelines according to claim 3, characterized in that: The preheating adjustment method includes: Call the preheating temperature model, input the current powder thermal characteristic parameters and steel pipe substrate parameters, and obtain the target preheating temperature; The actual temperature of the substrate is obtained in real time using distributed temperature sensors. A thermal conduction delay compensation coefficient is introduced to calculate the target adjustment amount ΔT. aim ; The actual temperature of the substrate is collected in real time and compared with the melting peak temperature of the current batch of powder. If the fluctuation exceeds the preset temperature difference, the preheating temperature model is called again to update the target preheating temperature. At the same time, a time constant is introduced to correct the target adjustment amount. Set the secondary power switching threshold ΔT th.1 ΔT th.2 And ΔT th.1 >ΔT th.2 The heating power is dynamically adjusted. If |ΔT aim >ΔT th.1 Start the heating equipment to its maximum heating power; If |ΔT aim |≤ΔT th.2 Switch to heat preservation power; If ΔT th.2 <|ΔT aim |≤ΔT th.1 An incremental PID algorithm is used to calculate the heating power based on the target adjustment amount.

5. The production optimization control system for mining pipelines according to claim 4, characterized in that: The preheating adjustment method further includes: Based on the heating power and the target adjustment amount, a substrate temperature rise curve is plotted. The heating process is divided into multiple heating periods according to time intervals. The percentage of the target adjustment amount for each heating period is extracted, and a target adjustment list is generated. Drive the actuator according to the calculated heating power. After each heating period ends, synchronously obtain the actual temperature of the substrate and calculate the actual preheating completion rate and temperature deviation. If the actual completion rate is not less than 1 for two consecutive times and the temperature deviation is less than the preset deviation threshold, the preheating is considered successful and the spraying process is triggered; otherwise, preheating continues. After successful preheating, the expected adjustment amount in the target adjustment list is read, the actual adjustment amount is obtained, the target difference ΔT is calculated, and the target difference threshold is set to ΔT. th To determine whether the current temperature adjustment is effective; If ΔT>ΔT th If the error is detected, a level 3 error tracing mechanism is initiated; otherwise, the temperature adjustment is deemed effective.

6. The production optimization control system for mining pipelines according to claim 5, characterized in that: The injection adjustment method includes: Powder particle size distribution and dielectric constant are obtained by laser particle size analyzer, and pipe geometric parameters are obtained by 3D scanner to build historical database; Based on fluid dynamics and convolutional neural networks, a jet parameter mapping model is constructed. The jet parameter mapping model is trained using data from the historical database and outputs working parameters. Drive the spray gun to perform spraying operation and acquire sensor data streams during the spraying process; The target deposition density was calculated based on the electrostatic spraying adsorption theory. The measured deposition density was obtained by an infrared densitometer and an ultrasonic thickness gauge. The absolute deviation and relative deviation rate of the deposition were calculated respectively. Calculate the spray gun angle deflection and moving speed based on the pipe diameter and curvature. Set a secondary deviation threshold, establish a graded response strategy, and initiate graded correction based on the current operating status.

7. The production optimization control system for mining pipelines according to claim 6, characterized in that: The injection adjustment method further includes: By acquiring the vibrational spectrum of powder molecules using an online infrared spectrometer, identifying changes in characteristic peaks, and combining this with abrupt changes in the slope of the thermogravimetric analysis curve, the powder type switch can be determined. Obtain order information, and once an urgent order is detected, automatically assign a switching batch number and calculate parameter transition values ​​according to batch weight; After each preset length of pipe is sprayed, the standard deviation of the thickness distribution is obtained. Once the standard deviation of the thickness distribution exceeds the preset thickness distribution value, the source tracing is initiated and the parameters are adjusted according to the correction priority. The corrected parameters are fed back into the target deposition density calculation step, and the baseline process coefficients are updated periodically.

8. The production optimization control system for mining pipelines according to claim 7, characterized in that: The leveling control method includes: The viscosity of the powder melt is monitored by an online viscosity sensor, and the temperature field distribution of the coating is obtained by an infrared thermal imaging device to generate real-time data pairs. The Arrhenius equation is used to describe the viscosity-temperature mapping relationship, and the flow activation energy and pre-exponential factor are fitted by the least squares method to construct a relationship model; When the measured data deviates from the standard curve by more than the preset range, it is determined that the correspondence has deviated, triggering logarithmic transformation and linear fitting, recalculating the flow activation energy and pre-exponential factor, and updating the relational model parameters. Obtain the softening range parameters, calculate the heating rate, and drive the heating equipment; Based on the aforementioned relationship model, the optimal leveling temperature and isothermal time are calculated, and control commands are generated. For thermoplastic powders, the cooling rate is calculated based on the crystallization initiation temperature and activation energy to drive the cooling equipment; Obtain the batch weights for injection regulation, calculate the mixing activation energy, generate a gradient transition curve, and adjust the isothermal time; Calculate the leveling index, trigger graded correction based on a preset secondary threshold, prioritize adjusting the heating rate and calibrating parameters.

9. The production optimization control system for mining pipelines according to claim 8, characterized in that: The staged curing method includes: Real-time acquisition of powder melting peak temperature, crosslinking exothermic peak characteristics, molecular vibrational spectrum and melt index; and fitting of crosslinking activation energy based on DSC exothermic curve to update Arrhenius equation parameters. The curing process is divided into three stages: gelation, cross-linking propagation, and post-curing. Curing parameters for each stage are calculated. The process involves first-stage cooling to promote crystal nucleation, dynamic calculation of isothermal crystallization time during the isothermal crystallization stage, and second-stage cooling to control crystallization stress. When switching between multiple varieties, a dual-mode control system with the original parameter base and the target parameter increment is adopted, and automatic temperature rise compensation is given when the crosslinking degree of the transition batch is insufficient. Real-time monitoring of curing quality, calculation of curing deviation, and graded correction based on preset secondary curing deviation thresholds.

Images