A method and system for refrigeration control for a filter rod making machine

Abstract

The present application relates to filter rod cooling control technical field, especially in kind of refrigeration control method and system for filter rod making machine, the method comprises: calculating the cooling time of filter rod in cooling section;Acquire cooling fluid temperature and filter rod specification parameters, the cooling time, cooling fluid temperature and filter rod specification parameters obtained are input into the cooling fluid flow rate prediction model pre-constructed, and the predicted cooling fluid flow rate is output.The present application provides a kind of refrigeration control method for filter rod making machine, by combining the heat conduction simulation model of filter rod unit volume, temperature distribution curve in cooling process can be simulated, accurately identify supercooling region and high gradient region, and it is mapped to front section, middle section, rear section three independent control regions, based on abnormal score, the cooling fluid flow rate of each region is differentially adjusted, so that cooling process is more uniform and controllable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of filter rod cooling control, in particular to a refrigeration control method and system for a filter rod forming machine. BACKGROUND

[0002] In the production process of filter rods, acetate fiber tows need to be heated to a specified temperature to be plasticized with triacetin, and the plasticized filter rods need to be quickly cooled and shaped. The refrigeration control method is used to control this cooling link. The hot plasticized filter rods, i.e. acetate fiber tows, are softened and plasticized under the action of triacetin plasticizer and need to be quickly cooled and solidified to fix their shapes such as circumference, pressure drop and hardness, prevent deformation or adhesion, and when the cooling is insufficient, the filter rods will be too soft, affecting the pressure drop and taste during smoking, and even blocking the cigarette machine. The basic process of filter rod forming includes tow opening, plasticizer application, heating and plasticizing, shaping and cooling, and slitting.

[0003] The existing filter rod cooling control method mainly uses preset cooling temperature and cooling flow rate to control the cooling of filter rods, and the cooling time is mainly determined passively by the conveying line speed. In the existing cooling control method for filter rods, there is generally a lack of analysis and prediction of the temperature distribution and temperature gradient inside and outside the filter rods, and there are problems such as excessive temperature difference between the inside and outside and uneven temperature gradient in the cooling process, which can easily form internal stress concentration during shaping, affecting key quality indicators such as circumference, pressure drop and hardness. SUMMARY

[0004] To solve the above problems, the present application provides a refrigeration control method and system for a filter rod forming machine.

[0005] The present application adopts the following technical scheme, a refrigeration control method for a filter rod forming machine, on the conveying line of the filter rod forming section, the current line speed is collected in real time by an encoder or a laser speed meter, the length of the cooling section is collected, and the cooling time of the filter rod in the cooling section is calculated;

[0006] The cooling fluid temperature and the filter rod specification parameters are collected, and the obtained cooling time, cooling fluid temperature and filter rod specification parameters are input into a pre-constructed cooling fluid flow rate prediction model to output a predicted cooling fluid flow rate;

[0007] A heat conduction simulation model of the unit volume of the filter rod is constructed to simulate the temperature distribution curve of the filter rod along the conveying and cooling of the cooling section, and to identify abnormal areas, the abnormal areas including supercooling areas and high gradient areas;

[0008] The cooling section is divided into three independent control areas, i.e. a front section area, a middle section area and a rear section area, along the conveying direction of the filter rod, and each independent control area is equipped with an independent air cooler unit;

[0009] Map the super-cooling region and high gradient region to the divided front section region, middle section region and rear section region, obtain the anomaly scores of the three independent control regions, generate the regulation instruction based on the anomaly scores, adjust the cooling fluid flow rate of the three independent control regions based on the generated regulation instruction, and control the filter rod to enter the cooling section for cooling.

[0010] As a further description of the above technical solution: the control method further comprises:

[0011] A target temperature of the filter rod entering the rear section region is preset, the surface temperature of the filter rod entering the rear section region and the cooling fluid flow rate of the current rear section region are obtained in real time, a temperature difference value and a flow rate difference value are obtained, and the flow rate difference value is the predicted cooling fluid flow rate minus the cooling fluid flow rate of the current rear section region.

[0012] The filter rod specification parameter, the temperature difference value and the current flow rate difference value are input into a pre-constructed cooling time prediction model, and a cooling time is output, the length of the cooling section in the current rear section region is calculated based on the cooling time, and the linear velocity of the conveying line in the cooling section is adjusted.

[0013] As a further description of the above technical solution: the method for constructing the heat conduction simulation model of the unit volume of the filter rod comprises:

[0014] establishing a transient convection-conduction equation of the filter rod cooling section length and the radial length r;

[0015] Setting the convection boundary condition, the axisymmetric condition and the inlet condition of the surface boundary based on the cooling fluid temperature and the predicted cooling fluid flow rate;

[0016] Dividing the cooling section length into sections, dividing the radial length r into sections, and obtaining (i, j) grid nodes, wherein i is the index of the number of sections of the cooling section length , j is the index of the number of sections of the radial length r, i∈ , j∈ ;

[0017] Based on the extracted surface temperature curve , the core temperature curve and the surface-core temperature difference , the super-cooling region and the high gradient region are identified.

[0018] As a further description of the above technical solution: based on the extracted surface temperature curve , the core temperature curve and the surface-core temperature difference , the method for identifying the super-cooling region and the high gradient region comprises:

[0019] When , the first segment is marked as a supercooling region, wherein, the segment is marked as a supercooling region, wherein, is a filter rod glass transition or embrittlement reference temperature, is a safety margin;

[0020] When ; the first segment is marked as a high gradient region, wherein, the segment is marked as a high gradient region, wherein, is a preset temperature difference threshold.

[0021] As a further description of the above technical solutions: the training method of the cooling fluid flow rate prediction model comprises:

[0022] Pre-collect Q sets of training data, Q is a positive integer greater than 0, the training data includes cooling time, cooling fluid temperature and filter rod specification parameters, and corresponding cooling fluid flow rate;

[0023] Select a deep learning model as the cooling fluid flow rate prediction model, use the training data to train the cooling fluid flow rate prediction model, take the cooling time, cooling fluid temperature and filter rod specification parameters as the input of the cooling fluid flow rate prediction model, take the cooling fluid flow rate as the output of the cooling fluid flow rate prediction model, adopt the stochastic gradient descent method, adjust the weight and bias of the cooling fluid flow rate prediction model through the back propagation algorithm, so that the error between the prediction result of the cooling fluid flow rate prediction model and the actual result is minimized, set the loss function, the loss function is mean square error, when the loss function value reaches convergence, stop training the cooling fluid flow rate prediction model, take the cooling fluid flow rate prediction model corresponding to the loss function value reaching convergence as the trained cooling fluid flow rate prediction model.

[0024] As a further description of the above technical solutions: the method for obtaining the abnormal score of the three independent control regions and generating the regulation and control instruction based on the abnormal score comprises:

[0025] Obtain the number of abnormal regions and the corresponding abnormal region type in the front segment region, the middle segment region and the rear segment region in sequence;

[0026] Different weights are set for different abnormal region types, and the abnormal scores S y of the front segment region, the middle segment region and the rear segment region are calculated in sequence.

[0027] Preset abnormal score gradient threshold values S1, S 2和 S 3， Compare and analyze the abnormal scores S y of the front segment region, the middle segment region and the rear segment region with the corresponding abnormal score gradient threshold values to generate the regulation and control instruction.

[0028] As a further description of the above technical solution: the control instruction includes a first control instruction, a second control instruction and a dangerous control instruction; wherein the first control instruction and the second control instruction are both of a level of reducing the flow rate of the cooling fluid, and the second control instruction is of a greater level of reduction than the first control instruction, and when the dangerous control instruction is generated, the line is stopped and an alarm is given.

[0029] The abnormal score S of the front section, the middle section and the rear section is compared and analyzed with the corresponding abnormal score gradient threshold value to generate a control instruction. y The method for generating a control instruction by comparing and analyzing the abnormal score gradient threshold value includes:

[0030] When S y When S

[0031] When S y When S

[0032] When S y When S

[0033] When S y When S

[0034] As a further description of the above technical solution: the filter rod specification parameters include the fiber bundle density, the ATG content and the specification diameter.

[0035] As a further description of the above technical solution: the training method of the cooling time prediction model includes:

[0036] P sets of training data are collected in advance, P is a positive integer greater than 0, the training data includes filter rod specification parameters, temperature difference and current flow rate difference, and corresponding cooling time;

[0037] A deep learning model is selected as the cooling time prediction model, the training data is used to train the cooling time prediction model, the filter rod specification parameters, the temperature difference and the current flow rate difference are taken as the input of the cooling time prediction model, the cooling time is taken as the output of the cooling time prediction model, the random gradient descent method is adopted, the weights and biases of the cooling time prediction model are adjusted through the back propagation algorithm to minimize the error between the prediction result of the cooling time prediction model and the actual result, a loss function is set, the loss function is mean square error, when the loss function value reaches convergence, the training of the cooling time prediction model is stopped, and the cooling time prediction model corresponding to the loss function value reaching convergence is taken as the trained cooling time prediction model.

[0038] A refrigeration control system for a filter rod making machine is used to implement the refrigeration control method for the filter rod making machine, and the system includes:

[0039] The first data acquisition module acquires the current line speed in real time through an encoder or a laser speed meter on the conveying line of the filter rod forming section, acquires the length of the cooling section, and calculates the actual cooling time of the filter rod in the cooling section;

[0040] The flow rate generation module acquires filter rod specification parameters, inputs the acquired cooling time, cooling fluid temperature and filter rod specification parameters into a pre-constructed cooling fluid flow rate prediction model, and outputs a predicted cooling fluid flow rate;

[0041] The heat conduction simulation module constructs a heat conduction simulation model of a unit volume of the filter rod, simulates a temperature distribution curve of the filter rod along the conveying and cooling of the cooling section, and identifies abnormal regions, the abnormal regions including supercooling regions and high gradient regions;

[0042] The segmentation module divides the cooling section into three independent control regions, i.e., a front region, a middle region and a rear region, along the conveying direction of the filter rod, and each independent control region is equipped with an independent cooling fan unit;

[0043] The flow rate adjustment module maps the supercooling regions and the high gradient regions to the divided front region, middle region and rear region, acquires abnormal scores of the three independent control regions, generates a control instruction based on the abnormal scores, adjusts the cooling fluid flow rate of the three independent control regions based on the generated control instruction, and controls the filter rod to enter the cooling section for cooling;

[0044] The second data acquisition module presets a target temperature of the filter rod entering the rear region, acquires the surface temperature of the filter rod entering the rear region and the cooling fluid flow rate of the current rear region in real time, acquires a temperature difference value and a flow rate difference value, and the flow rate difference value is the predicted cooling fluid flow rate minus the cooling fluid flow rate of the current rear region;

[0045] The time adjustment module inputs the filter rod specification parameters, the temperature difference value and the current flow rate difference value into a pre-constructed cooling time prediction model, outputs a cooling time, acquires the line speed of the conveying line in the cooling section based on the length of the current rear region of the cooling time, and adjusts the line speed of the conveying line in the cooling section.

[0046] Beneficial effects:

[0047] The application provides a refrigeration control method for a filter rod forming machine, which realizes dynamic matching of the flow rate of the cooling fluid by collecting the conveying line speed and the cooling section length in real time, calculating the actual cooling time, combining the cooling fluid temperature and the filter rod specification parameters, and inputting into a cooling flow rate prediction model, avoids the phenomenon of overcooling or insufficient cooling that is prone to occur when the production rhythm changes in the traditional constant-speed cooling mode, and further combines the heat conduction simulation model of the unit volume of the filter rod, can simulate the temperature distribution curve in the cooling process, accurately identifies the overcooling area and the high gradient area, and maps them to the front section, the middle section and the rear section three independent control areas, differentiates the cooling fluid flow rate of each area based on the abnormal score, and makes the cooling process more uniform and controllable.

[0048] In addition, a target temperature control link is arranged in the rear section area, the surface temperature of the filter rod entering the rear section is monitored in real time, the flow rate difference and the temperature difference are calculated by combining the current flow rate and the predicted flow rate, the cooling time prediction model is input, and the conveying line speed is adjusted reversely, so that the double closed-loop control of temperature and residence time is realized at the end of cooling, the method effectively prevents the problem that the cooling is not up to standard due to the reduction of the flow rate of the cooling fluid in the front section area and the middle section area, effectively prevents the defects such as softening deformation, surface brittle cracking or internal stress unevenness in the cooling process of the filter rod, ensures the stability of the key indicators such as the circumference, the pressure drop and the hardness, and improves the product quality and the adaptability and stability of the production process. BRIEF DESCRIPTION OF DRAWINGS

[0049] The application will be further explained in combination with the drawings and embodiments:

[0050] Figure 1 A flowchart of a refrigeration control method for a filter rod forming machine is provided for the application embodiment 1;

[0051] Figure 2 A flowchart of a refrigeration control method for a filter rod forming machine is provided for the application embodiment 2;

[0052] Figure 3 A flowchart of a method for obtaining the abnormal scores of the three independent control areas and generating the control instructions is provided for the application embodiment 1;

[0053] Figure 4 A module connection diagram of a refrigeration control system for a filter rod forming machine is provided for the application embodiment 3. DETAILED DESCRIPTION

[0054] In order to make the technical means, creative features, purposes and effects realized by the application easy to understand, the application will be further described in combination with specific drawings. It should be noted that the embodiments in the application and the features in the embodiments can be combined with each other without conflict.

[0055] Embodiment 1

[0056] Please refer to Figure 1 and Figure 3 , the embodiment of the application provides a technical scheme: a refrigeration control method for a filter rod forming machine, comprising:

[0057] On the conveying line of the filter rod forming section, the current line speed is collected in real time by an encoder or a laser speedometer, the length of the cooling section is collected, and the cooling time of the filter rod in the cooling section is calculated;

[0058] The calculation method of the cooling time is: ; In the formula, is the cooling time, is the total length of the cooling section, is the line speed;

[0059] The filter rod specification parameters and the cooling fluid temperature are collected, the filter rod specification parameters including the fiber bundle density, the ATG content and the specification diameter; It should be noted that the cooling fluid temperature and the filter rod specification parameters are directly obtained by consulting the preparation manual.

[0060] The obtained cooling time, cooling fluid temperature and filter rod specification parameters are input into a pre-constructed cooling fluid flow rate prediction model, and a predicted cooling fluid flow rate is output;

[0061] The training method of the cooling fluid flow rate prediction model comprises:

[0062] Q sets of training data are collected in advance, Q is a positive integer greater than 0, the training data includes the cooling time, the cooling fluid temperature and the filter rod specification parameters, and the cooling fluid flow rate corresponding to the cooling time, the cooling fluid temperature and the filter rod specification parameters;

[0063] A deep learning model is selected as the cooling fluid flow rate prediction model, the training data is used to train the cooling fluid flow rate prediction model, the cooling time, the cooling fluid temperature and the filter rod specification parameters are used as the input of the cooling fluid flow rate prediction model, the cooling fluid flow rate is used as the output of the cooling fluid flow rate prediction model, the random gradient descent method is adopted, the weights and biases of the cooling fluid flow rate prediction model are adjusted through the back propagation algorithm, so that the error between the prediction result of the cooling fluid flow rate prediction model and the actual result is minimized, a loss function is set, the loss function is mean square error, when the loss function value reaches convergence, the training of the cooling fluid flow rate prediction model is stopped, and the cooling fluid flow rate prediction model corresponding to the loss function value reaching convergence is used as the trained cooling fluid flow rate prediction model.

[0064] A heat conduction simulation model of the filter rod unit volume is constructed, a temperature distribution curve of the filter rod along the cooling section conveying and cooling is simulated, and an abnormal area is identified, the abnormal area including a supercooling area and a high gradient area;

[0065] The method for constructing the heat conduction simulation model of the filter rod unit volume comprises:

[0066] The length of the cooling section of the filter rod is established and the transient convection-conduction equation of the radial length r is written for the temperature field of the moving filter rod:

[0067] ;

[0068] In the formula, = (r, ,t), represents the temperature, is the cooling section length, that is, the axial coordinate, from the cooling section inlet =0 to the outlet =L (m), L is the total length of the cooling section, and the unit is m (meter), r is the radial length, that is, the sectional radius direction coordinate, 0≤r≤R (m), R is the radius of the filter rod, and the unit is m (meter), t is time (unit: second); is the specific heat capacity (unit: J / (kg·K)); is the axial conveying speed of the filter rod; is the thermal conductivity of the filter rod (unit: W / (m·K)); wherein, = t.

[0069] In the formula, is the radial heat conduction, is the radial temperature gradient, is the axial heat conduction;

[0070] The convection boundary condition of the surface boundary (r=R) is set based on the cooling fluid temperature and the predicted cooling fluid flow rate;

[0071] ;

[0072] In the formula, is the heat exchange coefficient (W / (m²·K)), is the cooling fluid temperature; is the heat flux density at the radius R (filter rod surface) (unit: W / m²), is the filter rod surface temperature;

[0073] The heat exchange coefficient is an index for measuring the ability of the cold air to take away heat, and the larger the value is, the stronger the cooling is. The heat exchange coefficient can be obtained by field calibration method or estimation method;

[0074] The axisymmetric condition (r=0):

[0075] ;

[0076] Inlet condition (T(r,0,t)) =0): directly measured T(r,0,t) temperature when entering into the cooling section.

[0077] Divide the cooling section length L into L segments, and divide the radial length r into r segments, to obtain (i,j) grid nodes, wherein i is the segment index of the cooling section length L division, j is the segment index of the radial length r division, i∈ , j∈ ; Extract the surface temperature curve T(r,t), the core temperature curve Tc(r,t), and the surface-core temperature difference ΔT(r,t); ; ;

[0078] wherein T(r,t) represents the outer surface temperature at each axial position r; , Tc(r,t) represents the temperature at the axis at each axial position r; , and ΔT(r,t) represents the surface-core temperature difference.

[0079] ; ; ; ; ; ;

[0080] Based on the extracted surface temperature curve T(r,t), the core temperature curve Tc(r,t), and the surface-core temperature difference ΔT(r,t), identify the supercooling region and the high gradient region; ; ; ;

[0081] The identification method of the supercooling region and the high gradient region comprises:

[0082] When T(r,t) < Tg - ΔTg, mark the i-th segment as a supercooling region; wherein Tg is a filter rod vitrification or embrittlement reference temperature, ΔTg is a safety margin, and optionally, ΔTg is 2-5℃. ; ; ; ; ; ;

[0083] ; ; ; ; ;

[0084] Divide the cooling section into three independent control regions of a front section region, a middle section region, and a rear section region along the conveying direction of the filter rod, and each independent control region is equipped with an independent cooling fan unit.

[0085] Optionally, when the front section area, the middle section area and the rear section area are divided, the division lengths of the front section area, the middle section area and the rear section area are divided in an increasing order of length, and the length of the rear section area is divided to be the longest, thereby increasing the temperature regulation capability of the rear section area.

[0086] The subcooling region and the high gradient region are mapped to the divided front section area, the middle section area and the rear section area, the abnormal scores of the three independent control regions are obtained, the regulation instruction is generated based on the abnormal scores, the cooling fluid flow rate of the three independent control regions is adjusted based on the generated regulation instruction, and the filter rod is controlled to enter the cooling section for cooling.

[0087] The method for obtaining the abnormal scores of the three independent control regions includes:

[0088] The number of abnormal regions and the corresponding abnormal region types in the front section area, the middle section area and the rear section area are obtained in sequence;

[0089] Different weights are set for different abnormal region types, and the abnormal scores S of the front section area, the middle section area and the rear section area are obtained in sequence y ;

[0090] Optionally, the abnormal region types are the subcooling region and the high gradient region, and the subcooling region is set to 1 and the high gradient region is set to 1.2, so that the abnormal score of each independent control region is recorded as the number of high gradient regions multiplied by 1.2, and then the number of subcooling regions is added.

[0091] The preset abnormal score gradient threshold values S1, S 2和 S 3， The abnormal scores S y of the front section area, the middle section area and the rear section area are compared and analyzed with the abnormal score gradient threshold values to generate the regulation instruction, the regulation instruction includes a first regulation instruction, a second regulation instruction and a dangerous regulation instruction; wherein the first regulation instruction and the second regulation instruction are both to reduce the level of the cooling fluid flow rate, and the reduction level of the second regulation instruction is greater than that of the first regulation instruction; optionally, the first regulation instruction is to reduce the predicted cooling fluid flow rate by 5%-9%, and the second regulation instruction is to reduce the predicted cooling fluid flow rate by 10%-20%.

[0092] The abnormal scores S y of the front section area, the middle section area and the rear section area are compared and analyzed with the abnormal score gradient threshold values to generate the regulation instruction, the regulation instruction includes a first regulation instruction, a second regulation instruction and a dangerous regulation instruction; wherein the first regulation instruction and the second regulation instruction are both to reduce the level of the cooling fluid flow rate, and the reduction level of the second regulation instruction is greater than that of the first regulation instruction; optionally, the first regulation instruction is to reduce the predicted cooling fluid flow rate by 5%-9%, and the second regulation instruction is to reduce the predicted cooling fluid flow rate by 10%-20%.

[0093] When S y <S1, no regulation instruction is generated;

[0094] When S1≤S y <S2, the first regulation instruction is generated;

[0095] When S2≤S y When S3, a second control instruction is generated;

[0096] When S y When S3, a dangerous control instruction is generated.

[0097] It should be noted that when the dangerous control instruction is generated, the line is stopped and an alarm is given.

[0098] In this embodiment, by collecting the conveying line speed and the cooling section length in real time, the actual cooling time is calculated, and the cooling fluid temperature and the filter rod specification parameters are input into the cooling flow rate prediction model to realize dynamic matching of the cooling fluid flow rate, thereby avoiding the overcooling or insufficient cooling phenomenon that is prone to occur when the production rhythm changes in the traditional constant-speed cooling method. Further, in combination with the heat conduction simulation model of the filter rod unit volume, the temperature distribution curve in the cooling process can be simulated to accurately identify the overcooling area and the high gradient area, and map them to the front section, middle section, and rear section three independent control areas. Based on the abnormal score, the cooling fluid flow rate of each area is differentially adjusted to make the cooling process more uniform and controllable.

[0099] Embodiment 2

[0100] Please refer to Figure 2 The control method further includes: presetting a target temperature of the filter rod entering the rear section area, acquiring the surface temperature of the filter rod entering the rear section area and the cooling fluid flow rate of the current rear section area in real time, acquiring a temperature difference value and a flow rate difference value, the flow rate difference value being the predicted cooling fluid flow rate minus the cooling fluid flow rate of the current rear section area;

[0101] The filter rod specification parameters, the temperature difference value, and the current flow rate difference value are input into a pre-constructed cooling time prediction model to output the cooling time, and the line speed of the conveying line in the cooling section is calculated based on the length of the current rear section area according to the output cooling time, and the line speed of the conveying line in the cooling section is adjusted.

[0102] The training method of the cooling time prediction model includes:

[0103] P sets of training data are collected in advance, P being a positive integer greater than 0, and the training data including filter rod specification parameters, temperature difference values, and current flow rate difference values, and the cooling time corresponding to the filter rod specification parameters, temperature difference values, and current flow rate difference values;

[0104] The deep learning model is selected as the cooling time prediction model, the cooling time prediction model is trained using training data, the filter rod specification parameters, the temperature difference and the current flow rate difference are taken as the inputs of the cooling time prediction model, the cooling time is taken as the output of the cooling time prediction model, the random gradient descent method is adopted, the weight and bias of the cooling time prediction model are adjusted through the back propagation algorithm, so that the error between the prediction result of the cooling time prediction model and the actual result is minimized, the loss function is set, the loss function is the mean square error, when the loss function value reaches convergence, the training of the cooling time prediction model is stopped, and the cooling time prediction model corresponding to the loss function value reaching convergence is taken as the trained cooling time prediction model.

[0105] The cooling fluid flow rate of the rear section is the cooling fluid flow rate obtained by adjusting the predicted cooling fluid flow rate through the control instruction.

[0106] The surface temperature of the filter rod entering the rear section is measured by the set thermal imager.

[0107] In the embodiment, the target temperature control link is arranged in the rear section, the surface temperature of the filter rod entering the rear section is monitored in real time, the flow rate difference and the temperature difference are calculated in combination with the difference between the current flow rate and the predicted flow rate, the cooling time prediction model is input, and the conveying line speed is adjusted reversely, so that the double closed-loop control of the temperature and the residence time is realized at the cooling end. The method effectively prevents the problem that the cooling is not up to standard due to the reduction of the cooling fluid flow rate in the front section and the middle section, effectively prevents the defects such as softening deformation, surface brittle cracking or internal stress unevenness in the cooling process of the filter rod, ensures the stability of the key indexes such as the circumference, the pressure drop and the hardness, and improves the product quality and the adaptability and stability of the production process.

[0108] Embodiment 3

[0109] A refrigeration control system for a filter rod forming machine is used to realize the refrigeration control method for the filter rod forming machine. The system comprises a first data acquisition module, a flow rate generation module, a heat conduction simulation module, a sectioning module, a flow rate adjustment module, a second data acquisition module and a time adjustment module, and the modules are connected through wired / wireless connection.

[0110] The first data acquisition module acquires the current line speed in real time through an encoder or a laser speedometer on the conveying line of the filter rod forming section, acquires the length of the cooling section, and calculates the actual cooling time of the filter rod in the cooling section.

[0111] The flow rate generation module acquires the filter rod specification parameters, inputs the acquired cooling time, cooling fluid temperature and filter rod specification parameters into a pre-constructed cooling fluid flow rate prediction model, and outputs a predicted cooling fluid flow rate.

[0112] The heat conduction simulation module constructs a heat conduction simulation model of a filter rod unit volume, simulates a temperature distribution curve of the filter rod along a cooling section during conveying and cooling, and identifies abnormal areas, which include supercooling areas and high gradient areas.

[0113] The segmentation module divides the cooling section into three independent control areas, i.e., a front area, a middle area and a rear area, along the conveying direction of the filter rod, and each independent control area is equipped with an independent cooling fan unit.

[0114] The flow rate adjustment module maps the supercooling areas and the high gradient areas to the divided front area, middle area and rear area, obtains abnormal scores of the three independent control areas, generates a control instruction based on the abnormal scores, adjusts the cooling fluid flow rate of the three independent control areas based on the generated control instruction, and controls the filter rod to enter the cooling section for cooling.

[0115] The second data acquisition module presets a target temperature of the filter rod entering the rear area, acquires the surface temperature of the filter rod entering the rear area and the cooling fluid flow rate of the current rear area in real time, obtains a temperature difference value and a flow rate difference value, and the flow rate difference value is the predicted cooling fluid flow rate minus the cooling fluid flow rate of the current rear area.

[0116] The time adjustment module inputs the filter rod specification parameter, the temperature difference value and the current flow rate difference value into a pre-constructed cooling time prediction model, outputs a cooling time, calculates the line speed of the conveying line in the cooling section based on the length of the current rear area according to the output cooling time, and adjusts the line speed of the conveying line in the cooling section.

[0117] The above shows and describes the basic principles, main features and advantages of the present application. It should be understood by those skilled in the art that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method of refrigeration control for a filter rod making machine, characterized in that, The application relates to a control method for a cooling section of a filter rod forming section. The current line speed is collected in real time by an encoder or a laser speed meter on a conveying line of the filter rod forming section, the length of the cooling section is collected, and the cooling time of the filter rod in the cooling section is calculated; The cooling fluid temperature and the filter rod specification parameters are collected, the obtained cooling time, cooling fluid temperature and filter rod specification parameters are input into a pre-constructed cooling fluid flow rate prediction model, and the predicted cooling fluid flow rate is output; A heat conduction simulation model of a unit volume of the filter rod is constructed, a temperature distribution curve of the filter rod along the conveying direction of the cooling section is simulated, and an abnormal area is identified, the abnormal area including a supercooling area and a high gradient area; The cooling section is divided into three independent control areas, i.e., a front area, a middle area and a rear area along the conveying direction of the filter rod, and each independent control area is equipped with an independent cooling fan unit; The supercooling area and the high gradient area are mapped to the divided front area, middle area and rear area, the abnormal scores of the three independent control areas are obtained, the regulation and control instructions are generated based on the abnormal scores, the cooling fluid flow rates of the three independent control areas are adjusted based on the generated regulation and control instructions, and the filter rod is controlled to enter the cooling section for cooling; The control method further comprises: A target temperature of the filter rod entering the rear area is preset, the surface temperature of the filter rod entering the rear area and the cooling fluid flow rate of the current rear area are collected in real time, a temperature difference value and a flow rate difference value are obtained, and the flow rate difference value is the predicted cooling fluid flow rate minus the cooling fluid flow rate of the current rear area; The filter rod specification parameters, the temperature difference value and the current flow rate difference value are input into a pre-constructed cooling time prediction model, the cooling time is output, the line speed of the conveying line in the cooling section is obtained based on the length of the current rear area and the output cooling time, and the line speed of the conveying line in the cooling section is adjusted; The method for constructing the heat conduction simulation model of the unit volume of the filter rod comprises: A transient convection-conduction equation of the filter rod in the axial direction x and the radial direction r is established; A convection boundary condition of a surface boundary, an axial symmetry condition and an inlet condition are set based on the cooling fluid temperature and the predicted cooling fluid flow rate; dividing the cooling section length into segments, dividing the radial direction r into segments, to obtain (i, j) grid nodes, where i ∈ , j ∈ ; Temperature profile of extraction Core temperature profile Difference between surface-core temperature Identify sub-cooling, super-heating and high gradient regions; based on the extracted temperature profile , core temperature profile and shell-core temperature difference , the method of identifying the undercooling region, the overheating region and the high gradient region comprises: When then the segment is labeled as a supercooling region, where is the filter rod glassification or embrittlement reference temperature, is a safety margin; when Then The segment is marked as a high gradient region, where, This is a preset temperature difference threshold.

2. A method of refrigeration control for a filter rod making machine according to claim 1, characterized in that, The training method of the cooling fluid flow rate prediction model comprises: Q sets of training data are collected in advance, Q is a positive integer greater than 0, and the training data include the cooling time, the cooling fluid temperature and the filter rod specification parameters, and the cooling fluid flow rate corresponding to the cooling time, the cooling fluid temperature and the filter rod specification parameters; A deep learning model is selected as the cooling fluid flow rate prediction model, the training data are used to train the cooling fluid flow rate prediction model, the cooling time, the cooling fluid temperature and the filter rod specification parameters are taken as the input of the cooling fluid flow rate prediction model, the cooling fluid flow rate is taken as the output of the cooling fluid flow rate prediction model, the random gradient descent method is adopted, the weight and bias of the cooling fluid flow rate prediction model are adjusted through the back propagation algorithm, so that the error between the prediction result of the cooling fluid flow rate prediction model and the actual result is minimized, a loss function is set, the loss function is mean square error, the training of the cooling fluid flow rate prediction model is stopped when the loss function value reaches convergence, and the cooling fluid flow rate prediction model corresponding to the loss function value reaching convergence is taken as the trained cooling fluid flow rate prediction model.

3. A method of refrigeration control for a filter rod making machine according to claim 1, characterized in that, The method for obtaining the abnormal score of the three independent control areas and generating the control instruction comprises: The number of abnormal areas and the corresponding abnormal area type in the front section area, the middle section area and the rear section area are obtained in sequence; Different weights are set for different abnormal region types, and abnormal scores S of the front region, the middle region and the rear region are sequentially calculated y ; A preset abnormal score gradient threshold, S1, S 2和 S 3， The abnormal scores S y The abnormal scores S of the front, middle and rear regions are compared with the abnormal score gradient threshold to generate a control instruction.

4. A method of refrigeration control for a filter rod making machine according to claim 3, characterized in that, The control instruction comprises a first control instruction, a second control instruction and a dangerous control instruction; the first control instruction and the second control instruction are both levels of reducing the flow rate of the cooling fluid, and the reducing level of the second control instruction is greater than that of the first control instruction; when the dangerous control instruction is generated, the line is stopped and an alarm is given; The abnormal score S of the front section region, the middle section region, and the rear section region is calculated according to the following formula: y The method for generating a control instruction by comparing and analyzing the abnormal score gradient threshold value comprises: When S y When S1, no modulation instruction is generated; When S1≤S y When S2, generate the first control instruction; When S2≤S y When <S3, a second control instruction is generated; When S y When S3, generate a dangerous control instruction.

5. A method of refrigeration control for a filter rod making machine according to claim 1, characterized in that, The filter rod specification parameters comprise the fiber bundle density, the ATG content and the specification diameter.

6. A method of refrigeration control for a filter rod making machine according to claim 1, characterized in that, The training method of the cooling time prediction model comprises: P groups of training data are collected in advance, P is a positive integer greater than 0, the training data comprise filter rod specification parameters, temperature difference values and current flow rate difference values, and cooling times corresponding to the filter rod specification parameters, the temperature difference values and the current flow rate difference values; A deep learning model is selected as the cooling time prediction model, the training data are used to train the cooling time prediction model, the filter rod specification parameters, the temperature difference values and the current flow rate difference values are taken as the input of the cooling time prediction model, the cooling time is taken as the output of the cooling time prediction model, the random gradient descent method is adopted, the weight and bias of the cooling time prediction model are adjusted through the back propagation algorithm, so that the error between the prediction result of the cooling time prediction model and the actual result is minimized, a loss function is set, the loss function is mean square error, when the loss function value reaches convergence, the training of the cooling time prediction model is stopped, and the cooling time prediction model corresponding to the loss function value reaching convergence is taken as the trained cooling time prediction model.

7. A refrigeration control system for a filter rod making machine for implementing a refrigeration control method for a filter rod making machine according to any one of claims 1 to 6, characterized in that The system comprises: A first data acquisition module, on the conveying line of the filter rod forming section, a current line speed is acquired in real time through an encoder or a laser speed meter, the length of the cooling section is acquired, and the actual cooling time of the filter rod in the cooling section is calculated; A flow rate generation module, filter rod specification parameters are acquired, the acquired cooling time, cooling fluid temperature and filter rod specification parameters are input into a pre-constructed cooling fluid flow rate prediction model, and a predicted cooling fluid flow rate is output; A heat conduction simulation module, a heat conduction simulation model of a unit volume of the filter rod is constructed, a temperature distribution curve of the filter rod along the conveying direction of the cooling section is simulated, and an abnormal area is identified, the abnormal area comprising a supercooling area and a high gradient area; A segmentation module, the cooling section is divided into three independent control areas, i.e., a front section area, a middle section area and a rear section area, along the conveying direction of the filter rod, and each independent control area is equipped with an independent cooling fan unit; A flow rate adjustment module, the supercooling area and the high gradient area are mapped to the divided front section area, middle section area and rear section area, the abnormal score of the three independent control areas is obtained, the control instruction is generated based on the abnormal score, the cooling fluid flow rate of the three independent control areas is adjusted based on the generated control instruction, and the filter rod is controlled to enter the cooling section for cooling. The second data acquisition module is configured to preset a target temperature of a filter rod entering a rear section area, acquire a surface temperature of the filter rod entering the rear section area and a current cooling fluid flow rate of the rear section area in real time, acquire a temperature difference value and a flow rate difference value, and the flow rate difference value is a predicted cooling fluid flow rate minus the current cooling fluid flow rate of the rear section area; The time adjustment module is configured to input the filter rod specification parameter, the temperature difference value and the current flow rate difference value into a pre-constructed cooling time prediction model, output a cooling time, acquire a line speed of a conveying line in the cooling section based on a length of the rear section area according to the output cooling time, and adjust the line speed of the conveying line in the cooling section.

Images