A method for optimizing the performance parameters of a composite membrane based on a neural network

Abstract

The present invention provides a method for optimizing the performance parameters of a composite membrane based on a neural network, which relates to the technical field of composite membranes. By setting an optimization target set, it is used to dynamically adjust the optimization priority of the target; a two-stage optimization structure is constructed. The first layer uses a neural network input layer with additional weight coding, and the second layer automatically adjusts the hidden layer activation function based on the target sensitivity to adapt to different optimization requirements; a dynamic error compensation mechanism is adopted to jointly analyze the target errors respectively and adjust the neural network loss function to control the error under different optimization modes; through two consecutive rounds of optimization iterations, the target weights are adjusted, and the weight coding is corrected according to the error change trend to achieve the adaptive switching of the optimization target and improve the comprehensive performance of the composite membrane. The present invention can achieve higher optimization accuracy and adaptability in optimizing the performance parameters of the composite membrane and improve the comprehensive performance of the composite membrane.

Description

Technical Field

[0001] The present invention relates to the technical field of composite membranes, and more specifically, to a method for optimizing the performance parameters of composite membranes based on neural networks. Background Art

[0002] With the development of advanced manufacturing technologies, the preparation process of composite membranes has been continuously improved, resulting in enhanced material properties in terms of light transmittance, mechanical strength, durability, etc. However, since composite membranes usually involve the coordinated optimization of multiple physical performance parameters, how to achieve a balance among different performance indicators has become an important research direction in the current optimization design of composite membranes. Traditional optimization methods mainly rely on empirical parameter adjustment or optimization models based on experimental data, and there are many deficiencies. On the one hand, most optimization methods fail to fully consider the complex coupling relationships among different performance parameters, and local optimal problems are prone to occur during the optimization process. For example, when optimizing the light transmittance, the mechanical strength often decreases, and the existing methods lack a mechanism for dynamically adjusting the optimization weights, resulting in limited optimization effects. On the other hand, the adaptive ability of existing optimization methods is weak, mostly using fixed parameters or single optimization strategies, and it is difficult to make real-time adjustments according to different application requirements. In addition, although traditional neural network optimization methods can improve the performance of composite membranes to a certain extent, due to the fixed key parameters such as activation functions and loss functions, the optimization objectives are difficult to adaptively adjust with the change of errors, and the optimization efficiency is low. Therefore, there is an urgent need for an optimization method that can comprehensively consider the multi-objective optimization requirements and has the ability of dynamic regulation to improve the comprehensive performance of composite membranes. Summary of the Invention

[0003] In order to solve the above technical problems, the present invention is proposed. The present invention provides a method for optimizing the performance parameters of composite membranes based on neural networks.

[0004] According to one aspect of the present invention, there is provided a method for optimizing the performance parameters of composite membranes based on neural networks, which includes: setting an optimization target set for dynamically regulating the optimization priorities of the targets; constructing a two-stage optimization structure, where the first layer uses a neural network input layer with additional weight encoding, and the second layer automatically adjusts the hidden layer activation function based on the target sensitivity to adapt to different optimization requirements; adopting a dynamic error compensation mechanism to jointly analyze the target errors respectively and adjust the neural network loss function to control the errors in different optimization modes; through two consecutive rounds of optimization iterations, adjusting the target weights and correcting the weight encoding according to the error change trend to achieve the adaptive switching of the optimization objectives and improve the comprehensive performance of the composite membranes.

[0005] Further, the targets include light transmittance and mechanical strength.

[0006] Further, the optimization target set is , where, is the light transmittance target, is the mechanical strength target.

[0007] Furthermore, the optimization priorities of the dynamic regulation targets include:

[0008] Set the optimization priority parameters , control the weights of light transmittance and mechanical strength during the optimization process; perform adaptive target weight adjustment based on experimental data; and identify the target weight configuration adopted in the current optimization process through optimized target encoding.

[0009] Furthermore, the two-stage optimization structure includes: The first-stage optimization: By attaching optimized target weight encoding to the input layer of the neural network, the network can understand the current optimization priority and distinguish different optimization requirements; The second-stage optimization: Adopt the target sensitivity mechanism, and adaptively adjust the activation function of the hidden layer according to the influence degree of the optimization target on the result, so that the network can adapt to different optimization requirements.

[0010] Furthermore, the second-stage optimization includes: Using the set G of optimization targets to take the derivative of the composite film structure parameters to represent the influence degree of the changes in different composite film structure parameters on the target optimization, that is, the target sensitivity; Establishing a target sensitivity matrix based on the target sensitivity; Adjusting the activation function of the hidden layer according to the calculation result of the target sensitivity matrix.

[0011] Furthermore, the adoption of the dynamic error compensation mechanism and the joint analysis of the target errors respectively include: Calculating the light transmittance error and the mechanical strength error respectively; Ensuring the correct optimization direction according to the error ratios of the light transmittance error and the mechanical strength error; Calculating the dynamic error compensation factor to dynamically adjust the optimization weights of the light transmittance error and the mechanical strength error.

[0012] Furthermore, the light transmittance error is measured by the difference between the predicted value of the neural network and the true value measured experimentally; The mechanical strength error is uniformly calculated by comparing the predicted value with the true value.

[0013] Furthermore, adjusting the neural network loss function to control errors in different optimization modes includes: The optimization modes include the light transmittance priority mode, the mechanical strength priority mode, and the balanced optimization mode; In different optimization modes, dynamically adjust the loss function, including: In the light transmittance priority mode, the loss function increases the weight of the light transmittance error; In the mechanical strength priority mode, the loss function increases the weight of the mechanical strength error;

[0014] In the balanced optimization mode, the loss function equalizes the light transmittance error and the mechanical strength error, so that the optimization weights of the two remain stable.

[0015] Further, the continuous two - round optimization iteration includes: in the first - round optimization process, read the optimized target weight encoding, calculate the error compensation factor β, and dynamically adjust the optimized target weight based on the error change trend to ensure the balance of the optimization strategy; in the second - round optimization process, use the optimized target weight adjusted in the first round for training, continuously and dynamically adjust the error compensation factor β to optimize the error balance, and calculate the error convergence situation. If the error change trend slows down, further adjust the weight; otherwise, judge the optimization convergence to ensure the correct final optimization direction.

[0016] Compared with the prior art, the method and system for optimizing the performance parameters of a composite film based on a neural network provided by the present invention enable the network to identify different optimization requirements by attaching the optimized target weight encoding to the input layer of the neural network in the first layer, and adjust the optimization strategy accordingly; in the second layer, adaptively adjust the activation function of the hidden layer based on the target sensitivity to improve the adaptability of the neural network to different optimization targets. In addition, the introduction of a dynamic error compensation mechanism ensures the balanced adjustment of the light transmittance and mechanical strength errors during the optimization process, avoiding the overall performance degradation caused by excessive bias towards a certain optimization target. The continuous two - round optimization iteration further improves the self - adaptability of the optimization strategy. By adjusting the target weight encoding according to the error change trend, the optimization process becomes more intelligent and avoids falling into local optimal solutions. The present invention can achieve higher optimization accuracy and adaptability in optimizing the performance parameters of the composite film, and improve the comprehensive performance of the composite film. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings:

[0018] Figure 1 It is a flowchart of the method for optimizing the performance parameters of a composite film based on a neural network according to an embodiment of the present invention.

[0019] Figure 2 It is a schematic diagram of optimizing the priority of dynamic regulation targets in the method for optimizing the performance parameters of a composite film based on a neural network according to an embodiment of the present invention.

[0020] Figure 3 It is a schematic diagram of a two - stage optimization structure in the method for optimizing the performance parameters of a composite film based on a neural network according to an embodiment of the present invention.

[0021] Figure 4 It is a schematic diagram of two - round optimization iteration in the method for optimizing the performance parameters of a composite film based on a neural network according to an embodiment of the present invention. Detailed implementation manners

[0022] Next, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited by the exemplary embodiments described herein.

[0023] As mentioned in the above background art, since composite films usually involve the collaborative optimization of multiple physical performance parameters, how to achieve a balance between different performance indicators has become an important research direction for the current optimization design of composite films. Traditional optimization methods mainly rely on empirical parameter adjustment or optimization models based on experimental data, such as response surface methodology (RSM), genetic algorithm (GA), and finite element analysis (FEA), etc. These methods can improve certain performances of the composite film to a certain extent. However, due to the lack of precise adjustment of multi-objective trade-off in the optimization process, it may sacrifice other key performances while improving a certain performance, and it is difficult to meet the high-precision and intelligent optimization requirements.

[0024] There are multiple deficiencies in the existing composite film optimization technologies. On the one hand, most optimization methods fail to fully consider the complex coupling relationships between different performance parameters, and local optimum problems are likely to occur during the optimization process.

[0025] Therefore, an optimization scheme for composite film performance parameters based on a neural network is needed.

[0026] Figure 1 FIG. is a flowchart of an optimization method for composite film performance parameters based on a neural network according to an embodiment of the present invention. As Figure 1 shown, in the optimization method for composite film performance parameters based on a neural network, the following operating steps are included:

[0027] S1: Set an optimization target set for dynamically regulating the optimization priority of the target; S2: Construct a two-level optimization structure. The first layer uses a neural network input layer with additional weight encoding, and the second layer automatically adjusts the hidden layer activation function based on the target sensitivity to adapt to different optimization requirements; S3: Adopt a dynamic error compensation mechanism to jointly analyze the target errors respectively, and adjust the neural network loss function to control the error in different optimization modes; S4: Through two consecutive rounds of optimization iterations, adjust the target weights, and correct the weight encoding according to the error change trend to achieve the adaptive switching of the optimization target and improve the comprehensive performance of the composite film.

[0028] In the production and application of composite films, light transmittance and mechanical strength are the two most important performance indicators, and there is usually a conflict between the two. However, these two goals have a certain conflict. For example, improving light transmittance often means reducing the thickness of the film or changing the material ratio, which may lead to a decrease in mechanical strength; conversely, improving mechanical strength may affect the light transmittance of the material. Therefore, it is difficult for traditional single-objective optimization methods to find the optimal balance between the two goals, which easily leads to optimization biased towards a certain characteristic and failure to take into account the overall performance. Therefore, how to find a suitable optimization balance between the two is the core problem that the present invention needs to solve.

[0029] Traditional optimization methods often adopt a single-objective optimization strategy, that is, optimizing transmittance or mechanical strength separately, and cannot flexibly adjust the target priority under different demand scenarios. In addition, the traditional optimization method lacks a data-driven intelligent control mechanism, which makes the optimization direction easily deviate from the actual application requirements. This step S1 sets an optimization target set and introduces a mechanism for dynamically controlling the optimization priority, so that the optimization process can adapt to different application requirements and provide reasonable target settings for subsequent optimization strategies. In an embodiment of the present invention, the step S1 specifically includes:

[0030] In order to construct a reasonable set of optimization targets, it is first necessary to collect the performance data of the composite film, including light transmittance and mechanical strength. The specific method of data collection is as follows:

[0031] Using a UV-visible spectrophotometer, the composite film sample was fixed on an optical test stand, and the sample was illuminated by a light source. The transmittance was calculated by detecting the intensity of the transmitted light, and the transmittance at different wavelengths was recorded, and the average transmittance was determined. Using an electronic universal testing machine, the composite film was subjected to a tensile test in accordance with the ASTM D882 standard, the maximum tensile strength (MPa) was recorded, the tensile strength of multiple samples was measured, and the average value was determined. The 3σ principle was used to eliminate abnormal experimental data to avoid measurement errors affecting the optimization process, and the transmittance and mechanical strength were normalized to ensure that the data was within the range of 0-1, making the neural network training more stable.

[0032] Set optimization goal set ,in, It is the light transmittance target, which is used to measure the light transmittance of the composite film. Its unit is percentage (%). The higher the value, the better the light transmittance performance. It is a mechanical strength target, used to measure the tensile and tear resistance of the composite film. Its unit is MPa (megapascal). The higher the value, the better the mechanical strength.

[0033] Optional, optimization target set The setting mode makes this method suitable for many different application scenarios, such as: Display screen film: Prioritize light transmittance ; Protective film material: prioritize mechanical strength ; Comprehensive application film material: balance and .

[0034] Furthermore, after setting the optimization target set, as Figure 2 shown, it is necessary to define the optimization priority so that the optimization process can be flexibly adjusted between different targets:

[0035] Optimization priority parameter , used to control the weights of light transmittance and mechanical strength during the optimization process, and satisfy the following conditions: , only optimize the light transmittance , ignore the mechanical strength ; 0, only optimize the mechanical strength , ignore the light transmittance ; , equal-weight optimization of light transmittance and mechanical strength, that is, the balance mode.

[0036] Preferably, an adaptive target weight adjustment method based on experimental data is adopted to calculate the initial value. Set the average light transmittance and average mechanical strength in the experimental data, and calculate the light transmittance and mechanical strength of the current film material, then the calculation formula of

[0037]

[0038] Among them, and are the light transmittance and mechanical strength of the current film material respectively, and are the average light transmittance and average mechanical strength in the experimental data respectively.

[0039] If the light transmittance of the current film material deviates greatly from the historical average, the light transmittance will be preferentially increased during the optimization process, that is tends to 1; if the mechanical strength of the current film material deviates greatly from the historical average, the mechanical strength will be preferentially enhanced during the optimization process, that is tends to 0; if the deviations of both are similar, then is close to 0.5 to achieve balanced optimization.

[0040] In order to enable the neural network to process the optimization target set G, the present invention introduces an optimization target encoding for identifying the target weight configuration adopted in the current optimization process:

[0041] 01 means >0.5, that is, preferential optimization of light transmittance; 10 means < 0.5, that is, the mechanical strength is preferentially optimized; 11 means = 0.5, that is, the two are balanced and optimized. This encoding method can be used in the input layer of the neural network to make the regulation of the optimization strategy more intelligent and structured.

[0042] It should be noted that there is a contradiction between the light transmittance and the mechanical strength target. This step makes the optimization direction more in line with the actual needs by dynamically adjusting the optimization priority, and through adaptive weight adjustment , reducing unnecessary iterations in the training process and improving the convergence speed of the optimization algorithm. The optimization target encoding enables the neural network to adapt to different types of membrane materials and improves the applicable range of the optimization method. Compared with the linear weight assignment that requires manual setting of weights and does not have self-adaptability, the multi-objective optimization method of the present invention adopts a data-driven dynamic regulation mechanism, which can automatically adjust the optimization target priority, not only taking into account the light transmittance and mechanical strength, but also improving the optimization efficiency, and has irreplaceability.

[0043] In the embodiment of the present invention, the step S2 needs to construct a two-level optimization structure adapted to different optimization requirements based on the optimization target setting of S1, so that the neural network can more accurately optimize the performance parameters of the composite membrane, such as Figure 3 As shown, specifically including, the first-level optimization (input layer optimization): By adding an optimization target weight encoding to the input layer of the neural network, the network can understand the current optimization priority and distinguish different optimization requirements; the second-level optimization (hidden layer optimization): Adopting a target sensitivity mechanism, according to the influence degree of the optimization target on the result, adaptively adjust the activation function of the hidden layer, so that the network can adapt to different optimization requirements and improve the optimization accuracy and stability.

[0044] Specifically, the first-level optimization includes: In order to enable the neural network to recognize the current optimization priority (light transmittance priority, mechanical strength priority or balanced optimization of both), a set of optimization target weight encodings need to be added to the input layer, and the optimization target weight encoding comes from the optimization target priority set in step S1 . Among them, the input data of the neural network includes the following two parts: the structural parameters of the composite membrane: including key parameters such as membrane thickness, porosity, and material ratio that affect the performance of the composite membrane and the optimization target weight encoding: the optimization priority parameter from S1 , and is added to the input layer in an encoded manner, and the input data format is, for example , where represents the structural parameters of the composite membrane, represents the total number of the structural parameters of the composite membrane, that is, the dimension of the input features, , representing the weight encoding of the optimization objective, is calculated from the contribution degree of the optimization objective itself. It can be seen that the influence degree of the optimization objective is no longer discrete but continuous, enabling the neural network to adjust the optimization direction more smoothly and improve the optimization accuracy.

[0045] During the optimization process, the normalized values of the light transmittance and mechanical strength are calculated to ensure that their dimensions are consistent, so as to have a reasonable numerical range when calculating , and then the weight encoding of the optimization objective is calculated, where ranges from [0, 1], representing the current focus of optimization: when , it means that the light transmittance optimization takes precedence (i.e., takes a larger value); when , it means that the mechanical strength optimization takes precedence (i.e., takes a smaller value); when , it means balanced optimization of both.

[0046] It should be noted that by introducing this continuous variable, the neural network can adjust the optimization strategy more smoothly and no longer rely on discrete binary encoding, thereby improving the optimization accuracy. This weight encoding method can be directly input into the neural network without additional logical judgment, making the optimization process more efficient.

[0047] Specifically, the second-layer optimization includes: the hidden layer of the neural network is responsible for learning complex features in the data and adjusting the optimization direction. In this step, a target sensitivity analysis mechanism is introduced, and according to the influence degrees of the light transmittance and mechanical strength on the optimization result, the activation function of the hidden layer is adaptively adjusted to improve the optimization accuracy and stability.

[0048] The target sensitivity is used to measure the influence degree of the optimization objective on the optimization result. The derivative of the composite film structure parameters can be calculated using the optimization objective set G to represent the influence degree of the change of the composite film structure parameters on the target optimization.

[0049] Furthermore, to facilitate the control of the optimization process, a target sensitivity matrix is constructed:

[0050]

[0051] Among them, represents the sensitivity of the parameter to the light transmittance ; represents the sensitivity of the parameter to the mechanical strength . According to the target sensitivity matrix Based on the calculation results, adjust the activation function of the hidden layer to enable the neural network to better adapt to the current optimization requirements. The specific strategy is as follows: If the light transmittance is large (i.e., the light transmittance has a greater impact on the optimization result), use the ReLU activation function to improve the stability of the light transmittance optimization; if the mechanical strength is large (i.e., the mechanical strength has a greater impact on the optimization result), use the Leaky ReLU activation function to enhance the robustness of the optimization; if the sensitivities of both are comparable, use the Sigmoid activation function to make the optimization process more balanced. This mechanism enables the neural network to automatically adapt to different optimization goals and improves the reliability of the optimization results.

[0052] It should be noted that in the present invention, by adding optimization target encoding to the input layer, the neural network can adjust the optimization strategy according to different optimization requirements; adaptively adjust the activation function of the hidden layer to make the optimization process more stable and reduce the computational cost; and through target sensitivity analysis, ensure the applicability of the optimization results under different composite film structures.

[0053] In the embodiment of the present invention, although a two-stage optimization structure is constructed in the above process, where the first layer (input layer) uses additional weight encoding as the neural network input, and the second layer (hidden layer) adaptively adjusts the activation function using target sensitivity analysis to adapt to different optimization requirements, however, the neural network is usually affected by the following problems during the optimization process: For example, the competition relationship between optimization goals: there is a mutual restriction relationship between the light transmittance and the mechanical strength, and simply optimizing one index may lead to the deterioration of the other index; due to factors such as data noise, measurement errors, and parameter uncertainties, the prediction results of the neural network may gradually deviate from the optimal solution, resulting in difficult model convergence or optimization failure; and under different optimization modes (light transmittance priority, strength priority, balanced optimization), the influence weights of errors on the final optimization results are different, and using a fixed loss function cannot take into account all situations.

[0054] In traditional neural network optimization methods, the loss function is often fixed, such as the mean square error (MSE) or cross-entropy. However, in multi-objective optimization tasks, a single loss function is difficult to effectively handle the balance problem between objectives. For example: Only using MSE may result in good optimization effect of the light transmittance, but excessive loss of mechanical strength; only using the Huber loss may improve the robustness, but it cannot adjust the error weight according to the optimization mode. Therefore, the present invention introduces a dynamic error compensation mechanism, and the specific steps of S3 include,

[0055] During the training process, the light transmittance error and the mechanical strength error are calculated separately to determine the focus of the current optimization. Among them, the light transmittance error can be measured by the difference between the predicted value of the neural network and the true value measured experimentally. The magnitude of this error determines the accuracy of the light transmittance optimization; the mechanical strength error can also be calculated by comparing the predicted value with the true value, and the magnitude of the error reflects the reliability of the optimization algorithm for predicting the mechanical strength.

[0056] To ensure the correct optimization direction, it is necessary to determine the error ratio between the two, that is: if the light transmittance error is much larger than the mechanical strength error, the optimization weight of the light transmittance needs to be increased to reduce the deviation of the light transmittance; if the mechanical strength error is much larger than the light transmittance error, the optimization strategy needs to be adjusted to make the optimization focus shift to the mechanical strength; if the two errors are close, an equilibrium optimization strategy should be adopted to reduce both errors simultaneously.

[0057] In addition, under different optimization modes, the relative importance of the errors will also be different: in the light transmittance priority mode, the mechanical strength error is allowed to increase appropriately, but the light transmittance error must be strictly controlled; in the mechanical strength priority mode, the light transmittance error can increase appropriately, but the mechanical strength error must be reduced to the lowest level; in the equilibrium optimization mode, both errors need to be strictly controlled to ensure the stability of the overall performance.

[0058] Furthermore, to further optimize the error compensation, a dynamic error compensation factor is calculated to dynamically adjust the optimization weights of the light transmittance error and the mechanical strength error:

[0059]

[0060] Among them, is a normalization factor, which is used to avoid the denominator being zero and enhance the numerical stability, is the optimization target weight coding of the light transmittance, is the optimization target weight coding of the mechanical strength;

[0061] It can be seen that calculates the relative difference between the weight codings of the two optimization targets (light transmittance and mechanical strength). The current weighting method usually uses fixed weights, and after introducing the dynamic error compensation factor this invention hopes to make the weight factor dynamically adjustable to meet the needs of different optimization targets. Therefore, the weight factor is expressed as:

[0062] When →0, that is, the optimization weights of the two are close, and the weight distribution remains unchanged from the original weights; when When it approaches 1, the weight allocation will tilt towards the direction with smaller errors to prevent a single large error from affecting the overall optimization. This method can ensure the dynamic balance between optimization objectives and make the error distribution more reasonable.

[0063] It can be seen that when the light transmittance error is large, the error compensation factor will automatically increase the optimization weight of the light transmittance, making the neural network pay more attention to the light transmittance optimization; when the mechanical strength error is large, the error compensation factor will automatically increase the optimization weight of the mechanical strength, making the optimization process pay more attention to the mechanical strength; when the errors of both are similar, the error compensation factor will be adjusted to an equilibrium state, making the optimization weights of both equal.

[0064] Optionally, the error compensation factor can be calculated based on the optimization objective weight and the current error ratio, enabling it to dynamically adapt to different optimization requirements. During the calculation process, the initial weight set for the optimization objective can be referred to and weighted average combined with the current error magnitude to ensure the rationality of the optimization strategy.

[0065] Furthermore, in traditional neural network optimization, the role of the loss function is to measure the deviation between the predicted value and the true value and guide the neural network to adjust the weights to reduce the error. To adapt to different optimization modes, the loss function needs to be dynamically adjusted: in the light transmittance priority mode, the loss function will increase the weight of the light transmittance error, making the neural network pay more attention to the optimization effect of the light transmittance; in the mechanical strength priority mode, the loss function will increase the weight of the mechanical strength error, making the optimization process more inclined to improve the mechanical strength; in the balanced optimization mode, the loss function will balance the light transmittance error and the mechanical strength error, making the optimization weights of both remain relatively stable, which can be expressed as the sum of the light transmittance error multiplied by the dynamic weight factor of the light transmittance error and the mechanical strength error multiplied by the dynamic weight factor of the mechanical strength error.

[0066] This method can effectively improve the flexibility of the optimization process, enabling the neural network to obtain the optimal optimization effect under different optimization modes.

[0067] During the actual optimization process, the optimization requirements for the light transmittance and the mechanical strength will be dynamically adjusted with changes in material properties, manufacturing processes, and environmental factors. If only relying on the initially set optimization weights for training, it may lead to: the optimization strategy being rigid and unable to adjust the optimization direction according to the target requirements, the error weight adjustment lagging and unable to adapt to the change in the optimization direction in a timely manner, resulting in a slower convergence speed, and it may also cause the performance of the final composite film not to reach the global optimum due to the bias towards a single optimization target.

[0068] Therefore, in the embodiments of the present invention, step S4 adopts two consecutive rounds of optimization iteration. In each round of optimization process, not only the error of the optimization target itself is reduced, but also the target weight is dynamically corrected using the error change trend to ensure that the optimization target can be adaptively adjusted according to the actual situation, thereby improving the comprehensive performance of the composite film.

[0069] Specifically, before performing two consecutive rounds of optimization iteration, the following prerequisites need to be ensured: an initial optimization target weight encoding has been established, and the target weight encoding has been obtained through the calculation of the optimization target contribution degree and added as an input feature to the neural network optimization system; the first-layer optimization structure: additional weight encoding in the neural network input layer to ensure that the optimization target weight can affect the entire optimization process; the second-layer optimization structure: automatically adjust the activation function of the hidden layer of the target sensitivity to ensure that different optimization targets can adjust the optimization direction under different activation methods; control the optimization weights of the light transmittance and mechanical strength errors through the error compensation factor β, and establish a loss function suitable for different optimization modes.

[0070] On this basis, as Figure 4 shown, the first round of optimization iteration: calculate the initial error change trend, including:

[0071] In the first round of optimization process, the system needs to perform the following steps: read the initial weight encoding of the optimization target set in S1; calculate the error compensation factor β for controlling the ratio of the optimization weights of the light transmittance and mechanical strength; use the weight encoding as part of the input data to enter the optimization system.

[0072] Perform the first round of optimization training: adopt the two-stage optimization structure of S2, and perform forward propagation calculation on the input data. Calculate the current light transmittance error and mechanical strength error; use the loss function set in S3 to update the network weights and complete the error backpropagation (BP training).

[0073] Record the error of the current round and calculate the error decline rate. That is, if the light transmittance error drops rapidly, it means that the current optimization strategy is effective for the light transmittance optimization, and the light transmittance weight can be appropriately reduced; if the mechanical strength error drops slowly, it means that the mechanical strength optimization is difficult, and its optimization weight needs to be increased to strengthen the optimization intensity.

[0074] Calculate the adjusted weight encoding: if the error decline rate of the light transmittance is greater than that of the mechanical strength error, reduce the light transmittance error and increase the mechanical strength error to make the optimization process tend to mechanical strength optimization; if the error decline rate of the mechanical strength is greater than that of the light transmittance error, reduce the mechanical strength error and increase the light transmittance error to strengthen the light transmittance optimization; if the error decline rates of both are close: keep the weight unchanged and continue with balanced optimization. Calculate the new optimization target weight encoding and update the input data.

[0075] Second-round optimization iteration: The target adaptive switching includes:

[0076] First, perform optimization training based on the adjusted weights: Use the updated optimization target weights for the second-round optimization iteration; during the training process, the error compensation factor β continues to dynamically adjust the optimization strategy to ensure that the error does not overly bias towards a certain optimization target.

[0077] Record the error after the second-round optimization and calculate the error convergence situation: If the error continues to decrease, it indicates that the optimization direction is correct, and the current weight strategy is maintained; if the error decrease slows down, it indicates that the optimization may enter a local optimum, and the optimization strategy needs to be further adjusted. Set an error convergence threshold. If the error change after two rounds of optimization is less than the set threshold, it is considered that the optimization converges, and the weight adjustment ends. Otherwise, continue to perform a new optimization iteration until the error stabilizes.

[0078] Through the above iterative operations, the optimization target can be adaptively adjusted to avoid the solidification of the optimization target, ensure that the light transmittance and mechanical strength of the composite film can reach the optimum simultaneously; improve the optimization accuracy, make the optimization process more accurate through the calculation of the error change trend, reduce ineffective training, and improve the optimization efficiency; prevent the local optimum trap, avoid the optimization strategy from prematurely converging to a suboptimal solution through the dynamic adjustment of the target weight; adapt to different optimization requirements, and can automatically adjust the optimization strategy in both the light transmittance priority mode, the mechanical strength priority mode, and the balanced optimization mode, improving the overall performance of the composite film.

[0079] It should be noted that in step S4, through two consecutive rounds of optimization iterations, the target weight is adjusted in combination with the error change trend, and the weight coding is corrected, enabling the optimization target to be adaptively adjusted according to the optimization requirements. This method can improve the optimization effect of the light transmittance and mechanical strength of the composite film, make the final optimization result more stable and reliable, and at the same time improve the optimization efficiency and ensure the correct final optimization direction.

[0080] In summary, the method for optimizing the performance parameters of a composite film based on a neural network according to the embodiments of the present invention is elucidated. It realizes the collaborative optimization of the light transmittance and mechanical strength of the composite film through an optimization target set, a two-stage optimization structure, a dynamic error compensation mechanism, and two consecutive rounds of optimization iterations. Compared with traditional methods, the present invention can dynamically adjust the priority of the optimization target according to the optimization requirements, make the optimization process more adaptable, and improve the stability and accuracy of the final optimization result.

[0081] The dual - level optimization structure of the present invention makes the optimization process more targeted. The first level enables the network to identify different optimization requirements by attaching optimization - objective weight coding to the input layer of the neural network, and adjusts the optimization strategy accordingly; the second level adaptively adjusts the activation function of the hidden layer based on the target sensitivity to improve the adaptability of the neural network to different optimization objectives. In addition, the introduction of the dynamic error compensation mechanism ensures that the errors of light transmittance and mechanical strength are evenly adjusted during the optimization process, avoiding excessive bias towards a certain optimization objective and resulting in a decline in overall performance. Two consecutive rounds of optimization iterations further improve the self - adaptability of the optimization strategy. By adjusting the target weight coding according to the error change trend, the optimization process becomes more intelligent and avoids falling into local optimal solutions. The present invention can achieve higher optimization accuracy and adaptability in optimizing the performance parameters of the composite film, and improve the comprehensive performance of the composite film.

[0082] In summary, the method for optimizing the performance parameters of a composite film based on a neural network according to the embodiments of the present invention has been elucidated. After considering the specification and practicing the disclosed embodiments herein, those skilled in the art will readily conceive of other embodiments of the present invention. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include known common knowledge or conventional technical means in the technical field not disclosed by the present invention.

[0083] It should be understood that the present invention is not limited to the exact structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present invention is only limited by the appended claims.

Claims

1. A method for optimizing composite membrane performance parameters based on neural network, characterized in that: include: Set a set of optimization targets to dynamically adjust the optimization priority of the targets; Construct a two-level optimization structure, the first layer uses the neural network input layer to add weight encoding, and the second layer automatically adjusts the hidden layer activation function based on the target sensitivity to adapt to different optimization requirements; A dynamic error compensation mechanism is used to jointly analyze the target errors and adjust the neural network loss function to control the errors in different optimization modes; Through two consecutive rounds of optimization iterations, the target weight is adjusted, and the weight coding is corrected according to the error change trend, so as to realize adaptive switching of the optimization target and improve the comprehensive performance of the composite membrane; The optimization priorities of the dynamic control objectives include: Setting optimization priority parameters , controls the weights of transmittance and mechanical strength during the optimization process; Adaptive target weight adjustment based on experimental data; By optimizing the target code, the target weight configuration used in the current optimization process is identified; The two-stage optimization structure includes: First-layer optimization: By adding optimization target weight encoding to the neural network input layer, the network can understand the current optimization priority and distinguish different optimization requirements; Second-layer optimization: Adopting the target sensitivity mechanism, according to the influence of the optimization target on the result, the hidden layer activation function is adaptively adjusted to enable the network to adapt to different optimization requirements; The second layer optimization includes: The optimization target set G is used to derive the composite membrane structure parameters to express the influence of the changes in different composite membrane structure parameters on the target optimization, that is, the target sensitivity; Establish a target sensitivity matrix based on target sensitivity; According to the calculation results of the target sensitivity matrix, the activation function of the hidden layer is adjusted; The optimization target set is , the input data of the neural network includes the following two parts: structural parameters of the composite membrane: including membrane thickness, porosity, material ratio, key parameters affecting the performance of the composite membrane and optimization target weight coding: optimization priority parameters , and added to the input layer in an encoded manner. The input data format is ,in, represents the structural parameters of the composite membrane, represents the total number of structural parameters of the composite film, that is, the dimension of the input feature. , represents the weight encoding of the optimization target, which is calculated by the contribution of the optimization target itself; During the optimization process, the calculation and Normalized values ​​of to ensure that their dimensions are consistent so that When it has a reasonable numerical range, then calculate the optimized target weight encoding ,in, The value range is [0,1], indicating the focus of the current optimization.

2. The method for optimizing composite membrane performance parameters based on neural network according to claim 1, characterized in that: The objectives include light transmittance and mechanical strength.

3. The method for optimizing composite membrane performance parameters based on neural network according to claim 2, characterized in that: in, is the transmittance target, For mechanical strength purposes.

4. The method for optimizing composite membrane performance parameters based on neural network according to claim 1, characterized in that: The dynamic error compensation mechanism is used to jointly analyze the target errors respectively, including: Calculate the transmittance error and mechanical strength error respectively; Ensure the optimization direction is correct according to the error ratio of the transmittance error and the mechanical strength error; The dynamic error compensation factor is calculated to dynamically adjust the optimization weights of transmittance error and mechanical strength error.

5. The method for optimizing composite membrane performance parameters based on neural network according to claim 4, characterized in that: The transmittance error is measured by the difference between the predicted value of the neural network and the actual value measured experimentally; The mechanical strength error is uniformly calculated by comparing the predicted value with the actual value.

6. The method for optimizing composite membrane performance parameters based on neural network according to claim 1, characterized in that: The adjustment of the neural network loss function and the control of the error in different optimization modes include: The optimization modes include a transmittance priority mode, a mechanical strength priority mode and a balanced optimization mode; The loss function is dynamically adjusted in different optimization modes, including: In the transmittance priority mode, the loss function increases the weight of the transmittance error; In the mechanical strength priority mode, the loss function increases the weight of the mechanical strength error; In the balanced optimization mode, the loss function balances the transmittance error and the mechanical strength error so that the optimization weights of the two remain stable.

7. The method for optimizing composite membrane performance parameters based on neural network according to claim 1, characterized in that: The two consecutive rounds of optimization iterations include: In the first round of optimization, the optimization target weight code is read, the error compensation factor β is calculated, and the optimization target weight is dynamically adjusted based on the error change trend to ensure the balance of the optimization strategy; In the second round of optimization, the optimization target weights adjusted in the first round are used for training, and the error compensation factor β is continuously and dynamically adjusted to optimize the error balance and calculate the error convergence. If the error change slows down, the weight is further adjusted, otherwise the optimization convergence is judged to ensure that the final optimization direction is correct.

Images