Plastic film double-drawing TDO flexible heating control method

Abstract

The invention discloses a plastic film double-drawing TDO flexible heating control method, and relates to the technical field of polyester film production. Comprising the following steps of hardware deployment and data acquisition, dynamic temperature field prediction, adaptive partition division, partition optimization control and the like. The TDO equipment adopts an oil / electricity hybrid heating mode, flexible switching is carried out when energy price fluctuates, heating modes are dynamically distributed based on an oil / electricity hybrid heating power model and a cost objective function, energy price fluctuation is fully utilized, and meanwhile, a solid-state relay matrix supports heating unit rapid recombination and mode switching control; according to the method, the temperature distribution field is predicted, meanwhile, a dynamic partitioning algorithm is adopted, real-time recombination of the heating area along with the change of the temperature field is achieved, the problem of local overheating or underheating caused by traditional fixed partitioning is solved, refined control over the temperature of all the partitions is achieved, and energy cost can be reduced easily.

Description

Technical Field

[0001] The present invention relates to the technical field of polyester film production, and specifically to a flexible heating control method for double-sided stretching TDO of plastic films. Background Art

[0002] The biaxial stretching process is the core link in the production of plastic films. Through longitudinal and transverse stretching, the molecular chains of the film are arranged in an orderly manner, ultimately achieving high mechanical properties and uniform thickness. Heating control is the key to the TDO process, directly affecting the crystallinity, thermal shrinkage rate, and surface quality of the film.

[0003] Most of the existing TDO heating technologies use a single electric heating method or oil heating method. During the enterprise production process, it is difficult to flexibly switch the heating process following the price fluctuations of energy sources (such as electricity, gas, etc.) and select a method with lower cost to reduce the production cost of the enterprise. Moreover, the traditional TDO uses a fixed-zone layout of heating units and cannot adjust the heating range according to the dynamic temperature field during the film forming process. For example, during the film stretching process, due to uneven thickness or fluctuations in stretching speed, local temperature gradients change suddenly, and the fixed zones are difficult to respond quickly, easily causing thermal stress concentration, film breakage, or crystal point defects. Therefore, we provide a flexible heating control method for double-sided stretching TDO of plastic films. Summary of the Invention

[0004] The purpose of the present invention is to provide a flexible heating control method for double-sided stretching TDO of plastic films to solve the above problems.

[0005] The technical problems solved by the present invention are as follows:

[0006] The present invention can be achieved through the following technical solutions: A flexible heating control method for double-sided stretching TDO of plastic films, including the following steps:

[0007] Step 1: Hardware deployment and data collection

[0008] Simultaneously configure oil heaters and electric heaters in each zone of the TDO, control the heating switching logic through PLC, and install an infrared thermal imager in the film forming area to collect temperature distribution data in real time, and synchronously obtain the electric heating power, gas flow rate, film stretching speed, film thickness, and heater working status;

[0009] Step 2: Construct a dynamic temperature field prediction model

[0010] Based on historical temperature distribution data, establish an LSTM prediction model. The input variables include the temperature distribution matrix, heater status matrix, film thickness distribution matrix, and stretching speed, and the output variable is the future one-minute temperature field prediction matrix. Data preprocessing uses spatio-temporal joint interpolation method and Min-Max normalization;

[0011] Step 3: Dynamically divide the heating zones

[0012] Extract gradient features based on the predicted temperature field, perform dynamic partitioning through an improved K-means clustering algorithm, introduce a gradient direction consistency penalty term into the objective function, and set partitioning splitting and merging constraint conditions, and implement the heating unit recombination control through a solid-state relay array;

[0013] Step 4: Perform heating optimization control on each heating zone

[0014] Calculate the oil / electric heating power demand, construct an optimization function with the goal of minimizing cost, determine the optimal heating mode ratio in combination with the particle swarm algorithm, and dynamically adjust the heater state matrix.

[0015] A further technical improvement of the present invention lies in: the elements in the heater state matrix are set as the state value pairs (z d , z y ), z d represents the working state of the electric heater, z y represents the working state of the oil heater, and both take values in 0, 1, where 0 means not working and 1 means in the heating state.

[0016] A further technical improvement of the present invention lies in: using the spatio-temporal joint interpolation method to process missing values, and the formula is:

[0017]

[0018] where λ represents the time decay coefficient, and X and Y respectively represent the output variable and the output variable.

[0019] A further technical improvement of the present invention lies in: the clustering objective function in Step 3 is:

[0020]

[0021] where N represents the number of current dynamic heating zones, i represents the dynamic heating zone index, R i represents the i-th dynamic heating zone, represents the average temperature of this dynamic heating zone, represents the temperature gradient amplitude matrix of the dynamic heating zone, μ represents the penalty weight coefficient, represents the Frobenius norm corresponding to the dynamic heating zone, and (x, y) represents the coordinates of the predicted sampling point;

[0022] The constraint conditions include temperature constraints and partitioning range constraints. Combining the optimization situation, when the constraint conditions are broken, the heating zones are split and merged.

[0023] A further technical improvement of the present invention lies in that: the temperature constraint is that the standard deviation of the temperature in the current partition exceeds a set threshold; if the standard deviation of the temperature in an adjacent partition exceeds the set threshold, a splitting operation is triggered to divide the partition into two sub - partitions;

[0024] The partition range constraint is that the maximum area of the partition range does not exceed four heating units and the distance between the "temperature centroids" of two adjacent partitions is not less than a set distance;

[0025] When the distance between the "temperature centroids" of two adjacent partitions is less than the set distance, a merging operation is triggered to merge the two partitions into one partition and the maximum partition area does not exceed the coverage working area of four heating units.

[0026] A further technical improvement of the present invention lies in that: a solid - state relay is used to form a switch matrix array, which is reorganized and bound according to the partition result, and the type and working state of the heater are controlled by switching the state of the solid - state relay.

[0027] A further technical improvement of the present invention lies in that: the method for calculating the oil / electric heating power demand in step four includes:

[0028] The oil heating power is:

[0029] The electric heating power is:

[0030] Among them, η oil represents the boiler thermal efficiency, H oil represents the calorific value of gas combustion; Q represents the required heat load, ρ represents the density of polyester film, c p represents the specific heat capacity of polyester film, ε represents the thermal conductivity; represents the temperature transient change rate in the corresponding heating partition, is the Laplace operator of the temperature field in the heating partition; η elc represents the electro - thermal conversion efficiency.

[0031] A further technical improvement of the present invention lies in that: with the goal of minimizing cost, an optimization function is constructed as:

[0032]

[0033] Among them, cost oil (t) represents the gas price at the current moment, cost elc (t) represents the electricity price at the current moment, φ i (t) represents the proportion of using an oil heater for heating; is the heating mode switching penalty factor, where β represents the switching penalty coefficient;

[0034] The constraints include:

[0035] Temperature stability constraint: |T(R i ) - T set (R i )| ≤ ΔT 允 , where ΔT 允 is the maximum allowable temperature difference; T set represents the required temperature driven by the film stretching speed and film thickness;

[0036] Energy supply constraint: That is, the oil heating energy supply should not exceed the upper limit of the total oil circuit energy supply; That is, the electric heating energy supply should not exceed the upper limit of the maximum circuit load;

[0037] Heat load balance constraint: Q(R i , t) = φ i (t)η oil H oil + (1 - φ i (t))η elc P elc , that is, the oil / electric heating hybrid power of a certain heating zone R i at the corresponding time t meets the heat demand of the heating zone;

[0038] Heating mode switching frequency: ∑RoundUp(|φ i (t) - φ i (t - 1)|, 0) ≤ K, where K is the limit value of the mode switching frequency.

[0039] A further technical improvement of the present invention lies in that the calculation formula for the required temperature is:

[0040]

[0041] In the formula, T g represents the polyester glass transition temperature, k represents the material viscoelastic coefficient, δ represents the draw ratio, d represents the film thickness; v represents the film stretching speed.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The TDO device of the present invention adopts an oil / electric hybrid heating method, which can be flexibly switched when the energy price fluctuates, so as to produce at a lower energy cost; based on the power model and cost objective function of oil / electric hybrid heating, the heating mode is dynamically allocated, making full use of the energy price fluctuation. At the same time, the solid-state relay matrix supports the rapid reorganization and mode switching control of the heating unit.

[0044] 2. The present invention predicts the temperature distribution field in the future period by constructing an LSTM prediction model, and at the same time adopts a dynamic zoning algorithm to realize the real-time reorganization of the heating zone with the change of the temperature field, solve the problems of local overheating or underheating caused by traditional fixed zoning, and based on the precise control and heating mode switching after dynamic zoning, while ensuring the production quality, the energy cost can be refinedly controlled. Brief Description of the Drawings

[0045] For the convenience of those skilled in the art to understand, the present invention will be further described below in conjunction with the accompanying drawings.

[0046] Figure 1 It is a schematic diagram of the execution process of the method of the present invention. Detailed Embodiments

[0047] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following describes in detail the specific embodiments, structures, features and effects of the present invention in conjunction with the accompanying drawings and preferred embodiments.

[0048] Please refer to Figure 1 As shown, a flexible heating control method for double-stretching TDO of plastic film specifically includes the following steps:

[0049] Step 1. Hardware deployment and data collection

[0050] Install oil heaters and electric heaters in each partition of TDO at the same time, and control the switching logic of the heating systems in each partition through PLC. At the same time, install an infrared thermal imager in the corresponding film forming area to obtain the temperature distribution, and obtain the electric heating power, gas flow rate, film stretching speed, film thickness and the working status of each heater in real time.

[0051] Step 2. Construct an LSTM dynamic temperature field prediction model based on the historical temperature distribution data in the film forming area

[0052] Define input variables:

[0053] The temperature distribution matrix T(t) of each partition of TDO at time t m×n , where m represents the number of natural horizontal partitions, and n represents the number of sampling points selected longitudinally;

[0054] The heater status matrix H(t) in each partition at time t m×1 , and the elements in the heater status matrix are a pair of status value pairs (z d , z y ), z d represents the working status of the electric heater, taking values in 0 and 1; z y represents the working status of the oil heater, also taking values in 0 and 1; thus (zd , z y ) has four states, namely (0, 0), (0, 1), (1, 0), and (1, 1), where 0 indicates not working and 1 indicates the heating state;

[0055] The film thickness distribution matrix δ(t) of each zone at time t m×j , where j represents the number of thickness sampling point distributions in the stretching direction;

[0056] In addition, the input variables also include the stretching speed and some optional parameters: the inlet air temperature, the inlet air flow rate, the outlet air temperature, the outlet air flow rate, and the oligomer deposition amount (measured by a laser scattering sensor);

[0057] Define the output variable: the temperature field prediction matrix for the next minute;

[0058] Perform missing value processing and standardization processing on the historical operation data:

[0059] Among them, the missing value processing uses the spatio-temporal joint interpolation method:

[0060] Among them, λ represents the time decay coefficient, which takes values between 0.6 and 0.8, and x and y represent the output variables and output variables respectively;

[0061] The standardization processing adopts the Min - Max normalization processing method.

[0062] Construct the input set and output set according to the historical operation data, and divide the training set and test set according to a certain ratio;

[0063] During training, arrange the input set in chronological order, set the batch size of the input to 32, set the sequence length of the input to 20, that is, corresponding to 200 seconds of historical operation data, and the prediction step is 60, that is, the temperature field for the next minute.

[0064] Step 3: Dynamically divide the heating zone based on the temperature field predicted by the model

[0065] S31: Extract the gradient features of the predicted temperature field to obtain the gradient amplitude matrix;

[0066] Specifically, use the Sobel operator to extract the gradient features of the predicted temperature field T pred That is:

[0067]

[0068] Then the gradient amplitude matrix is The gradient direction matrix is

[0069] S32. Construct clustering objective function based on K-means clustering algorithm for dynamic partitioning

[0070] The gradient direction consistency penalty term is introduced to avoid the thermal stress concentration problem caused by the drastic change of temperature gradient within the partition. The objective function is:

[0071] Where N represents the number of current dynamic heating partitions, i represents the dynamic heating partition index, and R i represents the ith dynamic heating partition, represents the average temperature of the dynamic heating zone, represents the temperature gradient amplitude matrix of the dynamic heating partition, μ represents the penalty weight coefficient, and takes values between 0.48 and 0.65; Represents the Frobenius norm of the corresponding dynamic heating partition, which forces the algorithm to generate partitions with smooth temperature transitions by penalizing high gradient areas; (x, y) represents the coordinates of the predicted sampling point.

[0072] It should be noted that the initial cluster centers are generated based on the extreme points of the temperature field (including high temperature extreme points and low temperature extreme points), and are recalculated at fixed time intervals. And update the partition;

[0073] In the adaptive dynamic change of partitions, set the region splitting and combination constraints:

[0074] The temperature constraint is: the temperature difference between adjacent partitions does not exceed the set threshold;

[0075] If the temperature standard deviation of the current adjacent partition exceeds the set threshold, the split operation is triggered to form two sub-partitions;

[0076] Partition range constraints:

[0077] The maximum area of the zoned area does not exceed the coverage area of four heating units;

[0078] When the distance between the "temperature centroid" positions of two adjacent partitions is less than the set distance, the merge operation is triggered and the two partitions are merged into the same partition; and the area of the largest partition does not exceed the coverage working area of four heating units; where the "temperature centroid" represents the average position of the temperature distribution within the partition.

[0079] S33, automatic partition reorganization control based on the partition result

[0080] The switch matrix array formed by solid-state relays is reorganized according to the partition results and bound to the latest partition. The type of heating unit (oil heating or electric heating) and the working state of the heating unit can be controlled by switching the state of the solid-state relays.

[0081] Step 4: Independently control the determined heating zones within a certain period to minimize costs

[0082] S41. Calculate the oil heating power P oil and the electric heating power P elc :

[0083] Oil heating power:

[0084] where η oil represents the boiler thermal efficiency, H oil represents the calorific value of gas combustion, which can be calculated according to the gas flow rate per unit time; Q represents the required heat load, ρ represents the density of polyester film, c p represents the specific heat capacity of polyester film, and ε represents the thermal conductivity; represents the rate of transient temperature change within the corresponding heating zone, is the Laplace operator of the temperature field within the heating zone, characterizing the intensity of heat diffusion;

[0085] Electric heating power:

[0086] where η elc represents the electro-thermal conversion efficiency;

[0087] S42. Construct an optimization objective function with the goal of minimizing the heating cost:

[0088]

[0089] where cost oil (t) represents the gas price at the current moment, cost elc (t) represents the electricity price at the current moment, φ i (t) represents the proportion of heating using an oil heater, which can be quickly determined according to the heater status matrix; is the heating mode switching penalty factor, where β represents the switching penalty coefficient, generally taking a value of 0.2;

[0090] The constraint conditions include:

[0091] Temperature stability constraint: |T(R i ) - T set (R i )| ≤ ΔT 允 , ΔT 允 is the allowable temperature difference, determined according to the film thickness and stretching speed, generally taking values between 1.5°C and 3.5°C; the required temperature In the formula, T g represents the polyester glass transition temperature, k represents the material viscoelasticity coefficient, δ represents the draw ratio, and d represents the film thickness;

[0092] Energy supply constraint: That is, the energy supply of oil heating should not exceed the upper limit of the total energy supply of the oil circuit (this upper limit value is calibrated according to the oil circuit system);

[0093] That is, the energy supply of electric heating should not exceed the upper limit of the maximum load of the circuit (this upper limit value is calibrated according to the line specification parameters);

[0094] Thermal load balance constraint: Q(R i ,t) = φ i (t)η oil H oil +(1 - φ i (t))η elc P elc , that is, the oil / electric heating mixed power of a certain heating zone R i at the corresponding moment t meets the thermal demand of the heating zone;

[0095] Heating mode switching frequency: ∑RoundUp(|φ i (t) - φ i (t - 1)|, 0) ≤ K, where RoundUp(value, number of decimal places) is the ceiling function, K is the limit value of the mode switching frequency, and in this embodiment, K is taken as 5, that is, the heating mode switching frequency within one hour should not exceed 5 times;

[0096] S43. Use the particle swarm algorithm to obtain the optimal proportion φ i * (t) of using the oil heater for heating, and adjust the state values of each heating unit pair (z d ,z y ) in the heater state matrix of this heating zone according to this optimal proportion, so as to complete the adjustment of the heating mode.

[0097] The above is only a preferred embodiment of the present invention, and does not impose any form of limitation on the present invention. Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to make equivalent embodiments with equivalent changes, but as long as it does not depart from the technical content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A flexible heating control method for double-drawn TDO of plastic film, characterized in that , including the following steps: Step 1, hardware deployment and data collection Simultaneously configure oil heaters and electric heaters in each partition of TDO, control the heating switching logic through PLC, and install an infrared thermal imager in the film forming area to collect temperature distribution data in real time, and synchronously obtain the electric heating power, gas flow rate, film stretching speed, film thickness, and heater working status; Step 2, construct a dynamic temperature field prediction model Establish an LSTM prediction model based on historical temperature distribution data. The input variables include the temperature distribution matrix, heater status matrix, film thickness distribution matrix, and stretching speed. The output variable is the future one-minute temperature field prediction matrix. Data preprocessing uses spatio-temporal joint interpolation method and Min-Max normalization; Step 3, dynamically divide the heating zone Extract gradient features based on the predicted temperature field, and perform dynamic partitioning through an improved K-means clustering algorithm. The objective function introduces a gradient direction consistency penalty term, and sets partition splitting and merging constraint conditions. The heating unit recombination control is realized through a solid-state relay array; Step 4, perform heating optimization control on each heating partition Calculate the oil / electric heating power demand, construct an optimization function with the goal of minimizing cost, combine the particle swarm algorithm to determine the optimal heating mode ratio, and dynamically adjust the heater status matrix.

2. A flexible heating control method for double-drawn TDO of plastic film according to claim 1, characterized in that, The elements in the heater status matrix are set to the status value pairs (z d , z y ), where z d represents the working status of the electric heater, and z y represents the working status of the oil heater. Both take values in 0, 1, where 0 means not working and 1 means in the heating state.

3. A flexible heating control method for double-drawn TDO of plastic film according to claim 1, characterized in that, The spatio-temporal joint interpolation method processes missing values, and the formula is: Among them, λ represents the time decay coefficient, and X and Y represent the output variable and the output variable respectively.

4. A flexible heating control method for double-drawn TDO of plastic film according to claim 1, characterized in that, The clustering objective function described in Step 3 is: Where, N represents the number of current dynamic heating zones, i represents the dynamic heating zone index, R i represents the i-th dynamic heating zone, represents the average temperature of the dynamic heating zone, represents the temperature gradient amplitude matrix of the dynamic heating zone, μ represents the penalty weight coefficient, represents the Frobenius norm of the corresponding dynamic heating zone, (x, y) represents the coordinates of the predicted sampling point; The constraint conditions include temperature constraints and partition range constraints. Combining the optimization situation, when the constraint conditions are broken, the heating partition is split and merged.

5. A flexible heating control method for double-drawn TDO of plastic film according to claim 4, characterized in that The temperature constraint is that the temperature standard deviation of the current partition exceeds the set threshold; if the temperature standard deviation of the adjacent partition exceeds the set threshold, a splitting operation is triggered to divide the partition into two sub-partitions; The partition range constraint is that the maximum area of the partition range does not exceed four heating units and the distance between the "temperature centroids" of two adjacent partitions is not less than the set distance; When the distance between the "temperature centroids" of two adjacent partitions is less than the set distance, a merging operation is triggered to merge the two partitions into one partition and the maximum partition area does not exceed the covering working area of four heating units.

6. A flexible heating control method for double-drawn TDO of plastic film according to claim 5, characterized in that Use a solid-state relay to form a switch matrix array to be reorganized and bound according to the partition result, and control the type of heater and the working status of the heater by switching the status of the solid-state relay.

7. A flexible heating control method for double - drawn TDO of plastic film according to claim 1, characterized in that, The method for calculating the oil / electric heating power demand in Step 4 includes: The oil heating power is: The electric heating power is: Among them, η oil represents the boiler thermal efficiency, H oil represents the calorific value of gas combustion; Q represents the required heat load, ρ represents the density of the polyester film, c p represents the specific heat capacity of the polyester film, and ε represents the thermal conductivity; represents the temperature transient change rate in the corresponding heating zone, ▽ 2 T pred (R i ) is the Laplace operator of the temperature field in the heating zone; η elc represents the electrothermal conversion efficiency.

8. A flexible heating control method for double-drawn TDO of plastic film according to claim 7, characterized in that, With the goal of minimizing cost, the constructed optimization function is: Among them, cost oil (t) represents the gas price at the current moment, cost elc (t) represents the electricity price at the current moment, φ i (t) represents the proportion of heating using an oil heater; is the heating mode switching penalty factor, where β represents the switching penalty coefficient; The constraint conditions include: Temperature stability constraint: |T(R i ) - T set (R i )| ≤ ΔT 允 , ΔT 允 is the maximum allowable temperature difference; T set represents the required temperature driven by the film stretching speed and the film thickness; Energy supply constraint: That is, the energy supply for oil heating should not exceed the upper limit of the total energy supply of the oil circuit; That is, the energy supply for electric heating should not exceed the upper limit of the maximum load of the circuit; Heat load balance constraint: Q(R i , t) = φ i (t)η oil H oil + (1 - φ i (t))η elc P elc , that is, the oil / electricity hybrid heating power of a certain heating zone R i at the corresponding time t meets the heat demand of the heating zone; Heating mode switching frequency: ∑RoundUp(|φ i (t)-φ i (t - 1)|, 0) ≤ K, where K is the limit value of the mode switching frequency.

9. A flexible heating control method for double-drawn TDO of plastic film according to claim 8, characterized in that The calculation formula for the required temperature is: Wherein, T g represents the glass transition temperature of the polyester, k represents the viscoelastic coefficient of the material, δ represents the draw ratio, d represents the film thickness; v represents the film drawing speed.

Images