Method for simultaneously carrying out surface hydrophobization modification on white carbon black and aluminum hydroxide and preparing white carbon black and aluminum hydroxide in composite insulator silicone rubber

Through ball milling process and intelligent algorithm optimization, the dispersion problem of aluminum hydroxide and white carbon black filler in silicone rubber composite insulators is solved, and the high compatibility and dispersion of the filler is achieved, extending the insulator life and improving performance.

CN119019850AInactive Publication Date: 2024-11-26YANGZHOU SHUANGBAO ELECTRIC EQUIP
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411117942.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-26
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The inorganic filler particles of aluminum hydroxide and white carbon black in silicone rubber composite insulators are small, and the surface is poorly compatible with the silicone rubber matrix, which makes it difficult to disperse evenly during the molding process, and is prone to powdering and hole defects, shortening service life.

Method used

The surface modifier composed of ball milling process, silane coupling agent and epoxy resin, is realized through mechanical force chemistry and space volume extrusion effect, and the surface hydrophobization modification of aluminum hydroxide and white carbon black is combined with multimodal Transformer, GNN and adaptive LSTM intelligent algorithms optimize the dispersion and mixing parameters of fillers to ensure uniform dispersion.

Benefits of technology

Improves the uniform dispersion and interface compatibility of aluminum hydroxide and white carbon black in silicone rubber, extends the service life of composite insulators, improves mechanical and electrical performance, simplifies process control, improves production efficiency and accuracy .

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119019850A_ABST
    Figure CN119019850A_ABST
Patent Text Reader

Abstract

The invention provides a method for simultaneously carrying out surface hydrophobization modification on white carbon black and aluminum hydroxide and preparing the white carbon black and aluminum hydroxide in composite insulator silicone rubber, and the technical scheme is as follows: the white carbon black and aluminum hydroxide are used as raw materials, gamma-aminopropyltriethoxysilane and oligomer bisphenol A diglycidyl ether are used as surface modifiers, and a mechanical ball milling method is adopted to prepare the composite insulator silicone rubber. By means of the mechanical force chemical effect in the ball milling process, collision between white carbon black and aluminum hydroxide particles and the particle volume occupying effect, particle surface hydrophobization modification is achieved while respective aggregates of the white carbon black and the aluminum hydroxide particles are opened. Modified white carbon black and aluminum hydroxide are used as fillers to be applied to molding processing of composite insulator silicon rubber, so that the interfacial compatibility between the two fillers and matrix silicon rubber is improved, and the dispersion uniformity of the two fillers in the silicon rubber matrix is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of rubber processing, and in particular relates to a method for simultaneously surface hydrophobizing white carbon black and aluminum hydroxide and preparing the same in composite insulator silicone rubber. Background Art

[0002] Silicone rubber composite insulators are widely used in UHV lines due to their light weight and pollution flashover resistance, and are likely to replace traditional glass and ceramic insulators. However, considering the requirements for electrical properties, mechanical properties and flame retardancy, aluminum hydroxide and white carbon black are often used as inorganic fillers in composite insulator silicone rubber, which play the role of flame retardant and reinforcing agent, and their total filling amount can reach the weight of the silicone rubber matrix. However, due to the small size of inorganic filler particles and poor compatibility between the surface and the silicone rubber matrix, it is difficult to evenly disperse aluminum hydroxide and white carbon black in the silicone rubber matrix during the molding process of silicone rubber composite insulators. This makes silicone rubber composite insulators prone to defects such as powdering and holes under harsh service conditions including high voltage, high humidity and heat, and high ultraviolet radiation, which in turn accelerates the aging of silicone rubber, reduces the service life of the composite insulator shed sheath, and also poses a huge challenge to the daily inspection of the power grid.

[0003] The modification of aluminum hydroxide or white carbon black has always been a problem of concern in rubber and even polymer processing. For example, in the manufacture of low rolling resistance tires, how to achieve hydrophobic modification, open the self-agglomeration of white carbon black particles, increase the compatibility of white carbon black with the rubber matrix, and improve the uniformity of the dispersion of white carbon black in the matrix has always been one of the difficulties that the industry needs to overcome. In view of the surface modification of aluminum hydroxide and white carbon black and the problem of self-particle agglomeration, and considering that these two fillers are the main filler system in the composite insulator silicone rubber, the present invention designs a method for the simultaneous mechanical and chemical modification of the two fillers, using a ball milling process, with the assistance of a surface modifier composed of a silane coupling agent and an epoxy resin, and using the effect of the space volume squeeze between the two particles, the crushing of the filler particle agglomerates, the surface modification hydrophobization and the improvement of particle dispersion are achieved; further, the modified mixed filler is used as the filling phase to manufacture a composite insulator silicone rubber in which aluminum hydroxide and white carbon black are uniformly distributed in the matrix. Summary of the invention

[0004] The present invention is mainly devoted to simultaneously solving the dispersion and surface modification of inorganic fillers including aluminum hydroxide and white carbon black with high filling amount in composite insulators, so that the obtained mixed filler is suitable for silicone rubber molding processing, thereby making the aluminum hydroxide and white carbon black uniformly dispersed in the silicone rubber.

[0005] The technical solution of the present invention is: a method for simultaneously surface hydrophobizing white carbon black and aluminum hydroxide and preparing them in composite insulator silicone rubber, comprising the following steps:

[0006] (1) Using fumed silica and aluminum hydroxide ATH as raw materials, weighing the fumed silica and ATH according to the ratio of composite insulator, adding anhydrous ethanol, and ultrasonically dispersing to obtain a suspension;

[0007] (2) transferring the above suspension to a ball mill, adding γ-aminopropyltriethoxysilane KH550 and oligomer bisphenol A diglycidyl ether E-44, and ball milling to obtain a suspension;

[0008] (3) the obtained suspension is centrifuged and dried at 60° C. under vacuum to obtain a surface hydrophobically modified fumed silica and aluminum hydroxide ATH mixed filler;

[0009] (4) kneading and mixing the obtained modified mixed filler with methyl vinyl silicone rubber, and adding other composite insulator silicone rubber processing aids during the kneading process to obtain a mixed rubber;

[0010] (5) The obtained mixed rubber is then subjected to plate vulcanization to obtain vulcanized silicone rubber.

[0011] 4. The preparation method according to claim 1 is characterized in that: in step (4), the mass ratio of methyl vinyl silicone rubber to mixed filler is 4:5; the added rubber additives include vulcanizers, vulcanization accelerators, antioxidants, etc.

[0012] Furthermore, in step (1), the mass ratio of the two fillers, fumed silica and ATH, is 1:4, and the solid content of the filler in ethanol is 0.25 g / ml.

[0013] Furthermore, in step (2), the ball milling is performed using a planetary ball mill, the ball milling speed is 400 rpm, and the ball milling time is 1 to 4 hours; the added modifier KH550 is 0.5% to 5.0% of the mass of the filler particles, and the added modifier E-44 is 1.0% to 3.0% of the mass of the filler particles.

[0014] Furthermore, in step (4), the mass ratio of methyl vinyl silicone rubber to mixed filler is 4:5; and the added rubber additives include vulcanizers, vulcanization accelerators, antioxidants, etc.

[0015] Furthermore, in step (4), an intelligent control system based on data analysis is introduced to optimize parameters such as temperature, time, and pressure during kneading and mixing to ensure uniform dispersion of aluminum hydroxide and white carbon black in silicone rubber; the specific steps are as follows:

[0016] 1. Sensor Arrangement

[0017] Sensors are arranged on the kneading and mixing equipment to monitor parameters such as temperature, pressure, time and amount of additives in real time; the details are as follows:

[0018] Temperature sensor: arranged inside the kneader and mixer to record the temperature T(t) at different time points;

[0019] Pressure sensor: arranged in the working chamber of the kneader and mixer to record the pressure P(t);

[0020] Time recording: automatically record the time t of the entire kneading and mixing process;

[0021] Additive dosage sensor: records the amount of additive used A(t);

[0022] 2. Data Collection

[0023] Data collection under different experimental conditions involves recording input parameters and corresponding output performance indicators:

[0024] Input parameter matrix X:

[0025] Record the temperature T, time t, pressure P and additive amount A under each experimental condition;

[0026]

[0027] Where n is the number of experimental conditions;

[0028] Output result matrix Y:

[0029] Record the filler dispersion D, mechanical properties M and electrical properties E under each experimental condition;

[0030]

[0031] 3. Data Preprocessing

[0032] Data preprocessing includes cleaning and standardization steps:

[0033] 1. Data cleaning: remove outliers and fill in missing data to ensure data integrity and reliability;

[0034] 2. Data standardization: standardize input parameters of different dimensions (such as temperature, time, pressure, and additive dosage) to the same scale to improve the effect of model training;

[0035]

[0036] Among them, x′ is the standardized data, x is the original data, μ is the mean, and σ is the standard deviation;

[0037] 4. Model Building

[0038] 1. Multimodal Transformer

[0039] Multimodal Transformer is used to process the relationship between different modes (such as temperature, time, pressure, and additive amount);

[0040] (1) Input layer:

[0041] The input parameters X include temperature T, time t, pressure P and additive amount A under all experimental conditions;

[0042] X={T,t,P,A}

[0043] (2) Encoder:

[0044] Linear transformation and embedding:

[0045] Perform linear transformation and embedding on the input parameters to generate a high-dimensional representation E

[0046] E=W e X+b e

[0047] Among them, E is a high-dimensional representation, W e is the embedding matrix, b e is the bias term;

[0048] (3) Self-attention mechanism

[0049] Calculate self-attention weights to capture the interrelationships between input parameters;

[0050]

[0051] Among them, Q, K, V are query, key and value matrices respectively, Attention is the self-attention mechanism calculation, softmax is the activation function, d k is the dimension of the key; the self-attention mechanism helps the model focus on the most relevant parts of the input data;

[0052] (4) Decoder

[0053] The decoder further processes the representation processed by the self-attention mechanism and outputs the hidden state h Transformer

[0054] h Transformer =Decoder(Attention(Q,K,V))

[0055] Among them, h Transformer is the hidden state of the Transformer decoder, and Decoder is the decoder function;

[0056] 2. Graph Neural Network (GNN)

[0057] (1) Graph structure representation

[0058] Representation of nodes and edges:

[0059] In GNN, experimental conditions are nodes, and the relationships between nodes are edges; the adjacency matrix A is used to represent the connection relationship between nodes;

[0060]

[0061] Where A is the adjacency matrix, is a set of real numbers, N is the number of nodes;

[0062] (2) Message passing mechanism

[0063] Node status update:

[0064] Through the message passing mechanism, the state of each node is updated according to the information of its neighbor nodes;

[0065]

[0066] in, is the state of node i at the tth iteration, is the state of node i at the t+1th iteration, is the set of neighbor nodes of node i, σ is the activation function; this mechanism helps the model capture the mutual influence between different experimental conditions; W1 and W2 are weight matrices, is the cumulative sum of the state vectors of node i’s neighbor node j at time step t;

[0067] (3) Graph Aggregation

[0068] Aggregate the states of all nodes to get the representation h of the entire graph G ;

[0069]

[0070] Among them, h G is the overall representation of the graph after T iterations, Pooling is the aggregation function, is the state vector of node i after T iterations;

[0071] 3. Adaptive LSTM

[0072] Adaptive LSTM is used to capture the temporal dependencies of time series data during the experiment;

[0073] (1) Input layer

[0074] Input time series data:

[0075] The input of the adaptive LSTM is time series data, including temperature T(t), pressure P(t), time t, and additive amount A(t);

[0076] X t ={T(t), P(t), t, A(t))

[0077] Among them, X t is the input data vector at time step t;

[0078] (2) LSTM unit

[0079] LSTM state update:

[0080] The LSTM unit updates the state through a combination of memory unit and forget gate, input gate, and output gate;

[0081] Forget Gate:

[0082] f t =σ(W f ·[h t-1 , X t ]+b f )

[0083] Among them, f t is the activation value of the forget gate, σ is the Sigmoid activation function, W f is the weight matrix of the forget gate, h t-1 is the hidden state at the previous moment, b f is the bias vector of the forget gate;

[0084] Input Gate:

[0085] i t =σ(W i ·[h t-1 , X t ]+b i )

[0086] Among them, i t is the activation value of the input gate, W i is the weight matrix of the input gate, b i is the bias vector of the input gate;

[0087] Output Gate:

[0088] o t =σ(W o ·[h t-1 , X t ]+b o )

[0089] Among them, t is the activation value of the output gate, W ois the weight matrix of the output gate, b o is the bias vector of the output gate;

[0090] Candidate memory status:

[0091]

[0092] in, is the candidate memory unit state at the current moment, tanh is the hyperbolic tangent activation function, W C is the weight matrix of the candidate memory unit, b C is the bias vector of the candidate memory unit;

[0093] Current memory status:

[0094]

[0095] Among them, C t is the memory cell state at the current moment, f t is the activation value of the forget gate, C t-1 is the memory unit state at the previous moment, i t is the activation value of the input gate;

[0096] Output status:

[0097] h t =o t tanh(C t )

[0098] Among them, h t is the hidden state at the current moment;

[0099] (3) Adaptive Mechanism

[0100] Parameter adaptive update: During the training process, the parameters of LSTM are adaptively updated according to the gradient of the loss function to better fit the time series data;

[0101]

[0102] Among them, θ t is the parameter of the tth iteration, θ t-1 is the parameter of the t-1th iteration, α is the learning rate, L t is the loss function, is the loss function L at the current time t t The gradient of the model parameters θ;

[0103] 4. Comprehensive model architecture

[0104] (1) Multimodal Fusion

[0105] The output hidden states of the multimodal Transformer, GNN, and adaptive LSTM are fused to obtain the comprehensive representation H;

[0106] H=[h Transformer, h G ,h t ]

[0107] Among them, H is a comprehensive representation;

[0108] (2) Fully connected layer

[0109] Input the fused representation H into the fully connected layer and output the prediction result y;

[0110] y=W·H+b

[0111] Among them, y is the predicted performance index, W is the weight matrix, and b is the bias term;

[0112] (3) Output results

[0113] The model finally outputs filler dispersion D, mechanical properties M and electrical properties E;

[0114] y={D,M,E}

[0115] Among them, y is the output prediction result;

[0116] 5. Real-time optimization and adjustment

[0117] 1. Real-time monitoring:

[0118] The trained model is integrated into the production line to monitor the parameters during kneading and mixing in real time;

[0119] 2. Dynamic adjustment:

[0120] According to the prediction results of the model, the kneading and mixing parameters such as temperature, time, pressure, and amount of additives are dynamically adjusted to ensure that the optimized parameters can improve the dispersion of the filler and the mechanical and electrical properties of the final product.

[0121] Furthermore, in step (5), the temperature of the plate vulcanization is 160° C., the vulcanization pressure is 10 MPa, and the vulcanization time is 20 minutes.

[0122] Compared with the prior art, the present invention has the following beneficial effects:

[0123] 1. The present invention adopts a mechanical ball milling method, selects a silane coupling agent and an oligomer to constitute a surface modifier, and utilizes the mechanical chemical effect and the volume crowding effect between different fillers to simultaneously achieve surface hydrophobic modification and agglomerate crushing of two main fillers, aluminum hydroxide and white carbon black, thereby improving the interface compatibility between the filler and the rubber matrix, and improving the dispersion uniformity of the two fillers in the rubber, thereby establishing a molding and processing method for a composite insulator silicone rubber with high compatibility and high dispersion of fillers.

[0124] 2. This method can simultaneously achieve surface hydrophobic modification and agglomerate crushing of two fillers. The process is simple and easy to operate and control.

[0125] 3. The introduction of intelligent algorithms such as multimodal Transformer, GNN and adaptive LSTM effectively solves the problems of uniform dispersion of fillers and optimization of mixing parameters. Multimodal Transformer optimizes the dispersion of fillers by comprehensively processing parameters under different experimental conditions; GNN mines the complex relationship between different conditions and improves the accuracy of parameter configuration; adaptive LSTM captures the dependencies in time series data, ensuring uniform dispersion of fillers and stability of final product performance.

[0126] 4. Through real-time prediction and optimization of parameters, the intelligent algorithm not only improves the production efficiency of composite insulator silicone rubber, but also ensures the mechanical and electrical properties of the final product, significantly improving the accuracy and economic benefits of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0127] Figure 1 Scan the aluminum-silicon element surface of silica and aluminum hydroxide filled silicone rubber;

[0128] in: Figure 1-1 , unmodified particles, Figure 1-2 , separately modified fumed silica particles, Figure 1 -3. Modify ATH particles separately, Figure 1 -4. Mixed modified particles;

[0129] Figure 2 Optimize flow charts for kneading and mixing based on intelligent algorithms. DETAILED DESCRIPTION

[0130] like Figure 1-2 As shown, a method for simultaneously surface hydrophobicizing white carbon black and aluminum hydroxide and applying them in composite insulator silicone rubber of the present invention comprises the following specific steps:

[0131] Using fumed silica and aluminum hydroxide ATH as raw materials, weighing the fumed silica and aluminum hydroxide ATH according to the ratio of composite insulator, adding anhydrous ethanol, and ultrasonically dispersing to obtain suspension 1;

[0132] The suspension 1 is transferred to a ball mill, and γ-aminopropyltriethoxysilane KH550 and oligomer bisphenol A diglycidyl ether E-44 are added, and the suspension 2 is obtained by ball milling;

[0133] The obtained suspension 2 was centrifuged and dried at 60°C in vacuum to obtain a surface hydrophobically modified fumed silica and aluminum hydroxide ATH mixed filler;

[0134] The obtained modified fumed silica and aluminum hydroxide ATH mixed filler are kneaded and mixed with methyl vinyl silicone rubber, and other composite insulator silicone rubber processing aids are added during the kneading process to obtain a mixed rubber;

[0135] The obtained mixed rubber is then subjected to plate vulcanization to obtain vulcanized silicone rubber.

[0136] In the above step (1), the mass ratio of the two fillers, fumed silica and aluminum hydroxide ATH, is 1:4; and the solid content of the filler in ethanol is 0.25 g / ml.

[0137] In the above step (2), a planetary ball mill is used for ball milling, the ball milling speed is 400 rpm, and the ball milling time is 1N4 hours; the KH550 added is 0.5% to 5.0% of the mass of the filler particles, and the E-44 is 1.0% to 3.0% of the mass of the filler particles.

[0138] In the above step (4), the mass ratio of methyl vinyl silicone rubber to mixed filler is 4:5; the added silicone rubber processing aids include vulcanizing agent, vulcanization accelerator and antioxidant.

[0139] Embodiment: 10g of fumed silica and 40g of ATH were weighed and dispersed in 150mL of anhydrous ethanol, and a suspension ① was obtained by ultrasonic dispersion; 2.5g of KH550 and 1.5g of E-44 were weighed and added to the suspension ①, and transferred to a ball mill jar, and the ball mill speed was 400 rpm, and the suspension ② was obtained after ball milling for 2h. The obtained suspension was centrifuged to remove ethanol, and the suspension was transferred to a vacuum drying oven for drying; the obtained product was ground to obtain a mixed filler; methyl vinyl silicone rubber and the mixed filler were put into a kneader in a ratio of 4:5 and fully mixed, and then opened for refining, and a vulcanizing agent was added. Assume the following experimental data:

[0140] Temperature (T): {180, 200, 220} degrees Celsius

[0141] Time (t): {30,45,60} minutes

[0142] Pressure (P): {1.5, 2.0, 2.5} MPa

[0143] Amount of additive (A): {5, 10, 15} parts

[0144] Performance indicators:

[0145] Filler dispersion (D): {0.85, 0.90, 0.88}

[0146] Mechanical properties (M): {75,80,78}MPa

[0147] Electrical properties (E): {9.5, 9.8, 9.6} kV / mm

[0148] 1. Data Collection

[0149] Input parameter matrix X:

[0150]

[0151] Output result matrix Y:

[0152]

[0153] 2. Data Preprocessing

[0154] standardization

[0155] Suppose we calculate the mean and standard deviation of each input parameter:

[0156] Temperature: μ T =200,σ T =20

[0157] Time: μ t =45,σ t =15

[0158] Pressure: μ P =2.0,σ P =0.5

[0159] Amount of additives: μ A =10,σ A =5

[0160] Normalize the input data:

[0161]

[0162] 3. Model building

[0163] (1) Multimodal Transformer

[0164] Input Layer:

[0165]

[0166] Encoder:

[0167] E=W eX′+b e

[0168] Self-Attention Mechanism:

[0169]

[0170] (2) GNN

[0171] Graph structure representation: adjacency matrix A

[0172]

[0173] Message passing mechanism:

[0174]

[0175] (3) Adaptive LSTM

[0176] Input time series data:

[0177] x t ={T(t), P(t), t, A(t)}

[0178] X t ={180,1.5,30,5}

[0179] LSTM state update

[0180] Forget Gate:

[0181] f t =σ(W f ·[h t-1 , X t ]+b f )

[0182] Input Gate:

[0183] i t =σ(W i ·[h t-1 , X t ]+b i )

[0184] Output Gate:

[0185] o t =σ(W o ·[h t-1 , X t ]+b o )

[0186] Candidate memory status:

[0187]

[0188] Current memory status:

[0189]

[0190] Output status:

[0191] h t =o t tanh(C t )

[0192] (4) Comprehensive model architecture

[0193] Multimodal fusion:

[0194] H=[h Transformer ,h G ,h t ]

[0195] Fully connected layer:

[0196] y=W·H+b

[0197] 4. Practical application examples

[0198] Input parameters: temperature T = 210 ° C, time t = 50 minutes, pressure P = 2.2 MPa, additive amount A = 12 parts

[0199] Model predictions:

[0200]

[0201] The prediction results show that the filler dispersion is 0.89, the mechanical properties are 79MPa, and the electrical properties are 9.7kV / mm. The parameters in the kneading and mixing process are optimized to improve the filler dispersion, mechanical properties and electrical properties, and a rubber compound is obtained. The obtained rubber compound is subjected to plate vulcanization (160°C, vulcanization pressure of 10MPa, and vulcanization time of 20 minutes) to obtain a silicone rubber material.

[0202] Figure 1 The elemental scans of Al and Si in silicone rubber prepared by using white carbon black and aluminum hydroxide as filler systems, white carbon black alone modified, aluminum hydroxide alone modified, and white carbon black and aluminum hydroxide modified at the same time as filler systems are given. From the elemental scans, when white carbon black and aluminum hydroxide are both unmodified ( Figure 1-1 ), the silica and aluminum hydroxide in the silicone rubber showed serious agglomeration, which was manifested by the obvious uneven distribution of Al and Si elements in the scanning area; when silica was modified alone ( Figure 1-2 ), the distribution uniformity of silica in silicone rubber is relatively Figure 1-1 The agglomeration of aluminum hydroxide is more serious. This is because the modified silica has better compatibility with the matrix, which weakens the volume repulsion between particles in the matrix. When aluminum hydroxide is modified alone ( Figure 1 -3), the distribution uniformity of aluminum hydroxide was not significantly improved compared with that of unmodified, while the uneven distribution of white carbon black was slightly improved due to the volume crowding effect; when both white carbon black and aluminum hydroxide were modified at the same time ( Figure 1 -4), compared with the cases of no modification or single modification, the distribution uniformity of white carbon black and aluminum hydroxide in the matrix has been effectively improved. This is due to the fact that during the simultaneous modification process, under the action of mechanical ball milling and the crowding out of heterogeneous particles, the two fillers are agglomerated and opened. At the same time, under the action of the modifier and the mechanical force of ball milling, the particles achieve surface hydrophobic modification and have better compatibility with the rubber matrix. The use of fillers in different modified combinations has different effects on the performance of silicone rubber. When both white carbon black and aluminum hydroxide are not modified, the filler has poor reinforcement performance on the matrix, resulting in low tensile strength; the interaction between the particles and the matrix is ​​low, and the elongation at break is reflected in the elasticity of the silicone rubber cross-linked network. The silicone rubber obtained by filling the mixed filler with the modified mixture shows good tensile strength and elongation at break, reflecting the improved compatibility between the filler and the matrix and the improved dispersion uniformity, so that the silicone rubber takes into account both strength and elasticity.

[0203] Table 1 shows the mechanical properties of silica and aluminum hydroxide filled silicone rubber:

[0204] Table 1

[0205]

[0206] In summary, the method of simultaneously hydrophobizing the surface of white carbon black and aluminum hydroxide and using them in composite insulator silicone rubber of the present invention, the technical scheme includes: using white carbon black and aluminum hydroxide as raw materials, using γ-aminopropyl triethoxysilane and oligomer bisphenol A diglycidyl ether as surface modifiers, using the method of mechanical ball milling, utilizing the mechanical force chemical action and the collision between white carbon black and aluminum hydroxide particles and the particle volume crowding effect during the ball milling process, while opening the respective agglomerates of white carbon black and aluminum hydroxide particles, the particle surface hydrophobizing modification is achieved. Using modified white carbon black and aluminum hydroxide as fillers, they are applied to the molding process of composite insulator silicone rubber, which improves the interface compatibility of the two fillers with the matrix silicone rubber and improves the uniformity of their dispersion in the silicone rubber matrix.

Claims

1. A method for simultaneously surface hydrophobizing white carbon black and aluminum hydroxide and preparing them in composite insulator silicone rubber, characterized in that: The following steps are involved: (1) Using fumed silica and aluminum hydroxide ATH as raw materials, weighing the fumed silica and aluminum hydroxide ATH according to the ratio of composite insulator, adding anhydrous ethanol, and ultrasonically dispersing to obtain a suspension 1; (2) The suspension 1 is transferred to a ball mill, γ-aminopropyltriethoxysilane KH550 and oligomer bisphenol A diglycidyl ether E-44 are added, and suspension 2 is obtained by ball milling; (3) The obtained suspension 2 is centrifuged and dried at 60° C. in a vacuum to obtain a surface hydrophobically modified fumed silica and aluminum hydroxide ATH mixed filler; (4) kneading and mixing the obtained modified fumed silica and aluminum hydroxide ATH mixed filler with methyl vinyl silicone rubber, and adding other composite insulator silicone rubber processing aids during the kneading process to obtain a mixed rubber; (5) The obtained mixed rubber is then subjected to plate vulcanization to obtain vulcanized silicone rubber.

2. The preparation method according to claim 1, characterized in that: In step (1), the mass ratio of the two fillers, fumed silica and aluminum hydroxide ATH, is 1:4; and the solid content of the filler in ethanol is 0.25 g / ml.

3. The preparation method according to claim 1, characterized in that: In step (2), the ball milling is performed using a planetary ball mill, the ball milling speed is 400 rpm, and the ball milling time is 1 to 4 hours; the KH550 added is 0.5% to 5.0% of the mass of the filler particles, and the E-44 is 1.0% to 3.0% of the mass of the filler particles.

4. The preparation method according to claim 1, characterized in that: In step (4), the mass ratio of methyl vinyl silicone rubber to mixed filler is 4:5; the added silicone rubber processing aids include vulcanizing agent, vulcanization accelerator and antioxidant.

5. The preparation method according to claim 1, characterized in that: In step (4), an intelligent control system based on data analysis is introduced to optimize the temperature, time and pressure parameters during kneading and mixing to ensure uniform dispersion of aluminum hydroxide and white carbon black in silicone rubber; the specific steps are as follows:

1. Sensor Arrangement Sensors are arranged on the kneading and mixing equipment to monitor the temperature, pressure, time and additive dosage parameters in real time; the details are as follows: Temperature sensor: arranged inside the kneader and mixer to record the temperature T(t) at different time points; Pressure sensor: arranged in the working chamber of the kneader and mixer to record the pressure P(t); Time recording: automatically record the time t of the entire kneading and mixing process; Additive dosage sensor: records the amount of additive used A(t); 2. Data Collection Data collection under different experimental conditions involves recording input parameters and corresponding output performance indicators: Input parameter matrix X: Record the temperature T, time t, pressure P and additive amount A under each experimental condition; Where n is the number of experimental conditions; Output result matrix Y: Record the filler dispersion D, mechanical properties M and electrical properties E under each experimental condition; 3. Data Preprocessing Data preprocessing includes cleaning and standardization steps:

1. Data cleaning: remove outliers and fill in missing data to ensure data integrity and reliability; 2. Data standardization: standardize input parameters of different dimensions: temperature, time, pressure, and additive dosage to the same scale to improve the effect of model training; Among them, x′ is the standardized data, x is the original data, μ is the mean, and σ is the standard deviation; 4. Model Building 1. Multimodal Transformer Multimodal Transformer is used to process the relationship between different modes: temperature, time, pressure, and additive amount; (1) Input layer: The input parameters X include temperature T, time t, pressure P and additive amount A under all experimental conditions; X={T,t,P,A} (2) Encoder: Linear transformation and embedding: Perform linear transformation and embedding on the input parameters to generate a high-dimensional representation E E=W e X+b e Among them, E is a high-dimensional representation, W e is the embedding matrix, b e is the bias term; (3) Self-attention mechanism Calculate self-attention weights to capture the interrelationships between input parameters; Among them, Q, K, V are query, key and value matrices respectively, Attention is the self-attention mechanism calculation, softmax is the activation function, d k is the dimension of the key; the self-attention mechanism helps the model focus on the most relevant parts of the input data; (4) Decoder The decoder further processes the representation processed by the self-attention mechanism and outputs the hidden state h Transformer : h Transformer =Decoder(Attention(Q,K,V)) Among them, h Transformer is the hidden state of the Transformer decoder, and Decoder is the decoder function; 2. Graph Neural Network GNN (1) Graph structure representation Representation of nodes and edges: In GNN, experimental conditions are nodes and the relationships between nodes are edges; the adjacency matrix A is used to represent the connection relationship between nodes; Where A is the adjacency matrix, is a set of real numbers, N is the number of nodes; (2) Message passing mechanism Node status update: Through the message passing mechanism, the status of each node is updated according to the information of its neighbor nodes; in, is the state of node i at the tth iteration, is the state of node i at the t+1th iteration, is the set of neighbor nodes of node i, σ is the activation function; this mechanism helps the model capture the mutual influence between different experimental conditions; W1 and W2 are weight matrices, is the cumulative sum of the state vectors of node i’s neighbor node j at time step t; (3) Graph Aggregation Aggregate the states of all nodes to get the representation h of the entire graph G ; Among them, h G is the overall representation of the graph after T iterations, Pooling is the aggregation function, is the state vector of node i after T iterations; 3. Adaptive LSTM Adaptive LSTM is used to capture the temporal dependencies of time series data during the experiment; (1) Input layer Input time series data: The input of the adaptive LSTM is time series data, including temperature T(t), pressure P(t), time t and additive amount A(t); X t ={T(t),P(t),t,A(t)) Among them, X t is the input data vector at time step t; (2) LSTM unit LSTM state update: The LSTM unit updates the state through a combination of memory unit and forget gate, input gate, and output gate; Forget Gate: f t =σ(W f ·[h t-1 ,X t ]+b f ) Among them, f t is the activation value of the forget gate, σ is the Sigmoid activation function, W f is the weight matrix of the forget gate, h t-1 is the hidden state at the previous moment, b f is the bias vector of the forget gate; Input Gate: i t =σ(W i ·[h t-1 ,X t ]+b i ) Among them, i t is the activation value of the input gate, W i is the weight matrix of the input gate, b i is the bias vector of the input gate; Output Gate: the t =σ(W o ·[h t-1 ,X t ]+b o ) Among them, t is the activation value of the output gate, W o is the weight matrix of the output gate, b o is the bias vector of the output gate; Candidate memory status: in, is the candidate memory unit state at the current moment, tanh is the hyperbolic tangent activation function, W C is the weight matrix of the candidate memory unit, b C is the bias vector of the candidate memory unit; Current memory status: Among them, C t is the current state of the memory unit, ft is the activation value of the forget gate, C t-1 is the memory unit state at the previous moment, i t is the activation value of the input gate; Output status: h t =o t ·tanh(C t ) Among them, h t is the hidden state at the current moment; (3) Adaptive Mechanism Parameter adaptive update: During the training process, the parameters of LSTM are adaptively updated according to the gradient of the loss function to better fit the time series data; Among them, θ t is the parameter of the tth iteration, θ t-1 is the parameter of the t-1th iteration, α is the learning rate, L t is the loss function, is the loss function L at the current time t t The gradient of the model parameters θ; 4. Comprehensive model architecture (1) Multimodal Fusion The output hidden states of the multimodal Transformer, GNN, and adaptive LSTM are fused to obtain the comprehensive representation H; H=[h Transformer ,h G ,h t ] Among them, H is a comprehensive representation; (2) Fully connected layer Input the fused representation H into the fully connected layer and output the prediction result y; y=W·H+b Among them, y is the predicted performance index, W is the weight matrix, and b is the bias term; (3) Output results The model finally outputs filler dispersion D, mechanical properties M and electrical properties E; y={D,M,E} Among them, y is the output prediction result; 5. Real-time optimization and adjustment 1. Real-time monitoring: The trained model is integrated into the production line to monitor the parameters during kneading and mixing in real time; 2. Dynamic adjustment: According to the prediction results of the model, the temperature, time, pressure and additive dosage parameters of kneading and mixing are dynamically adjusted to ensure that the optimized parameters can improve the dispersion of the filler and the mechanical and electrical properties of the final product.

6. The preparation method according to claim 1, characterized in that: In step (5), the temperature of the plate vulcanization is 160° C., the vulcanization pressure is 10 MPa, and the vulcanization time is 20 minutes.

Citation Information

Patent Citations

  • Preparation method of silicone rubber composite insulating material with excellent mechanical properties

    CN111944318A

  • Modification method for white carbon black

    CN112011092A

  • Silicone rubber material with good hydrophobicity and flame retardance and preparation method thereof

    CN114031945A

  • Silicone rubber material as well as preparation method and application thereof

    CN115386233A

  • Multi-mode smelting method based on the classification system of molten iron

    US20230368021A1