Insulating brick and method of making same

Abstract

The application discloses a heat-preservation building brick and a manufacturing method thereof, and belongs to the field of heat-preservation building bricks. The heat-preservation building brick and the manufacturing method thereof comprise a brick body, the brick body is composed of two building brick frames, the inner side of the building brick frame is provided with a bearing plate, the inner side of the bearing plate is provided with a gap, the two ends of the bearing plate extend to the two sides of the building brick frame and are fixedly connected with edge sealing plates, the edge sealing plates are parallel to the building brick frames, and the inner sides of the edge sealing plates are in contact with the surfaces of the building brick frames. The bearing plate and the edge sealing plate are designed, so that an effective heat insulation space is formed in the brick body, the parallel arrangement and the mutual contact of the edge sealing plates and the building brick frames ensure the compactness and the sealing property of the structure, the heat bridge effect is reduced, the manufacturing process of the heat-preservation building brick is relatively simple, mass production can be realized through reasonable mold design and pouring process, and the multi-level protection design is convenient for later maintenance and maintenance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of thermal insulation bricks, in particular to a thermal insulation brick and a manufacturing method thereof. BACKGROUND

[0002] Thermal insulation bricks are building materials that combine building enclosure and thermal insulation functions, mainly used for building enclosure thermal insulation. Thermal insulation bricks are building materials that integrate block shells made of high-performance commercial concrete hollow blocks and thermal insulation materials such as foamed commercial concrete or polystyrene foam plastic filling the cavities, through production processes. According to different materials and processes, thermal insulation bricks can be divided into various types, such as self-insulation blocks and composite thermal insulation blocks. Among them, self-insulation blocks refer to wall structures that meet the requirements of national and local current energy-saving building standards without the need for internal and external wall insulation technology.

[0003] For example, Chinese patent application No. 202321700548.8, entitled "Thermal insulation brick", includes a main block, a thermal insulation layer installed on one side of the outer wall of the main block, a protective layer installed on one side of the outer wall of the thermal insulation layer, a mounting mechanism on the thermal insulation layer, and a connecting mechanism between the main block and the protective layer. The connecting mechanism includes a limiting block, a connecting rib, and two limiting strips. The thermal insulation brick can be fixed and limited by the connecting rib, which can prevent the main block, the thermal insulation layer, and the protective layer from separating and causing a decrease in thermal insulation effect. The thermal insulation brick connects the main block and the protective layer together through the reinforcing mechanism to prevent the thermal insulation layer from separating from the main block and the protective layer, making the main block, the thermal insulation layer, and the protective layer more integrated, which can effectively prevent or avoid the separation or damage of the thermal insulation layer.

[0004] However, the existing bricks are mainly made of a whole concrete structure poured and connected, and the connected brick structure conducts and causes temperature overflow during use. When the connected brick structure of the internal and external environment is exposed to different temperature environments, heat is transferred between the brick structures, causing temperature gradient and continuous slow loss of indoor temperature.

[0005] Secondly, the design cross-sectional size of the masonry is too small, and its bearing capacity is insufficient to withstand the predetermined load, resulting in insufficient strength. At the same time, the openings or slots opened on the surface of the brick will weaken the cross-section of the wall and reduce its overall strength. SUMMARY

[0006] The present application aims to solve the problem of the slow indoor temperature dissipation caused by the temperature gradient of the brick in contact with the internal and external environment in the prior art, and provides a heat preservation brick and a manufacturing method thereof.

[0007] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme:

[0008] The heat preservation brick and the manufacturing method thereof comprise a brick body, the brick body is composed of two brick frames, the inner side of the brick frame is provided with a bearing plate, the inner side of the bearing plate has a gap, the two ends of the bearing plate extend to the two sides of the brick frame and are fixedly connected with an edge sealing plate, the edge sealing plate is parallel to the brick frame, the inner side of the edge sealing plate is in contact with the surface of the brick frame, and the materials of the edge sealing plate and the bearing plate are heat insulation resin materials.

[0009] As a preferred technical scheme of the present application, the outer side of the bearing plate is fixedly connected with a reinforcing rib, the side away from the bearing plate of the reinforcing rib extends to the inside of the brick frame, the inner side of the reinforcing rib is fixedly connected with an attached rod located in the inside of the brick frame, and the brick frame is wrapped on the surface of the reinforcing rib and the attached rod by pouring and solidification.

[0010] As a preferred technical scheme of the present application, the outer side of the bearing plate is fixedly connected with a support frame, the side away from the bearing plate of the support frame is in contact with the inner side of the brick frame, and the material of the support frame is resin material.

[0011] As a preferred technical scheme of the present application, the inner side of the bearing plate is provided with a connecting frame, and the inside of the connecting frame is filled with a heat preservation material.

[0012] As a preferred technical scheme of the present application, the heat preservation material comprises a nanometer coiled material filled in the inside of the connecting frame, and both sides of the nanometer coiled material are provided with heat preservation cotton boards located in the inside of the connecting frame.

[0013] As a preferred technical scheme of the present application, the four corners in the inside of the connecting frame are provided with connecting rods, the nanometer coiled material is wrapped on the surface of the connecting rods, both ends of the connecting rods sequentially penetrate the heat preservation cotton boards, the connecting frame, the bearing plate and the support frame and extend to the inside of the support frame, and both ends of the connecting rods are threadedly connected with nuts located in the inside of the support frame.

[0014] As a preferred technical scheme of the present application, the outer side of the brick frame is provided with an insulation layer, the outer side of the insulation layer is provided with a protective layer, the insulation layer comprises a waterproof coating coated on the outer side of the brick frame, the surface of the waterproof coating is adhered with an outer layer plate, and the protective layer comprises a plastic film fixedly connected on the surface of the outer layer plate, and the surface of the plastic film is coated with a sunscreen coating.

[0015] The heat preservation brick and the manufacturing method thereof comprise the following steps:

[0016] S1: According to the design size of the heat preservation brick, prepare the corresponding mold, including the mold of the brick frame and the mold of the supporting plate, edge sealing plate and other components, use heat insulation resin material to injection mold the supporting plate and edge sealing plate and the reinforcing ribs and attachment rods on both sides in the mold, ensure that the inner side of the supporting plate has enough space for subsequent filling of heat preservation material;

[0017] S2: After the supporting plate and the edge sealing plate are formed, necessary trimming and inspection are carried out to ensure that the size and shape meet the design requirements. When the supporting plate is processed, fix the support frame to the outer side of the supporting plate and move the supporting plate together with the edge sealing plate to the inside of the brick frame mold, ensure that the edge sealing plate is parallel to the brick frame mold and tightly contacts the inner surface of the mold, pour the pouring material (such as concrete or other suitable materials) of the brick frame into the mold, so that the pouring material completely fills the gap between the support frame and the mold, and at the same time the material fully wraps the reinforcing ribs and attachment rods, wait for the pouring material to solidify and form the brick frame;

[0018] S3: Install the connecting frame on the inner side of the supporting plate, and fill the nano coiled material and the heat preservation cotton board in the connecting frame. Use the connecting rod to penetrate the heat preservation cotton board, the connecting frame, the supporting plate and the support frame, and install the nut in the support frame to fix the connecting rod and the heat preservation material.

[0019] S4: Apply waterproof coating on the outer side of the brick frame to form an insulation layer. After the waterproof coating is dry, stick the outer layer plate on the surface of the waterproof coating and attach the plastic film on the surface of the outer layer plate. At the same time, apply sunscreen paint on the surface of the plastic film to form a protective layer, thereby completing the production and manufacturing of the heat preservation brick.

[0020] As a preferred technical solution of the present application, in step S3 of the heat preservation brick and its manufacturing method, in order to ensure accurate filling and fixing of the heat preservation material, an intelligent assembly algorithm is proposed. The specific implementation process is as follows:

[0021] I. Data collection and preprocessing

[0022] 1. Sensor arrangement

[0023] Visual sensor: high-resolution industrial camera is arranged at key positions of the assembly line to capture real-time images of the heat preservation material during assembly; the collected image data is represented by variable I k , where k represents the serial number of the collected image;

[0024] Force sensor: force sensors are arranged at the contact parts of the connecting rod, nano coiled material, heat preservation cotton board and other components to monitor the force of each component during assembly; the data of the force sensor is represented by variable F i , where i represents the number of the force sensor;

[0025] Temperature sensor: Temperature sensors are arranged in the assembly environment to monitor the effect of environmental temperature on the filling and fixation of the thermal insulation material; the temperature data is represented as variable T j , where j represents the number of temperature sensors;

[0026] 2. Data collection and labeling

[0027] Image data collection: Image data during the assembly process of the thermal insulation material is collected in real time through visual sensors, and the collection frequency of the image data is f I , and the collected data set is represented as where N I is the total number of image data;

[0028] Force and temperature data collection: Data from force sensors and temperature sensors are collected synchronously, and the force sensor data set is represented as The temperature sensor data set is represented as where N F and N T are the total number of force and temperature data, respectively;

[0029] Data labeling: The collected image and sensor data are labeled, and the labeling content includes the specific position P k , size S k , filling tightness D k , force state F i and temperature state T j of the components;

[0030] 3. Data preprocessing

[0031] Image preprocessing: The collected image data I k is processed through image enhancement, size scaling, and noise filtering in sequence, and the processed image data is represented as

[0032] Data normalization: Force data F i and temperature data T j are normalized to eliminate dimensional differences; the normalization formula is as follows:

[0033] Image data normalization:

[0034]

[0035] where I k (x,y) is the pixel value of the original image I k at position (x,y), I min and I max are the minimum and maximum values of the image pixels, respectively, For normalized image data at pixel value at position (x, y);

[0036] Normalization of force sensor data:

[0037]

[0038] wherein, is the value of the normalized force sensor data, F i is the measured value of the original force sensor at a certain time, F min is the minimum value of the force sensor data, F max is the maximum value of the force sensor data;

[0039] Normalization of temperature sensor data:

[0040]

[0041] wherein, is the value of the normalized temperature sensor data, T j is the measured value of the original temperature sensor at a certain time, T min is the minimum value of the temperature sensor data, T max is the maximum value of the temperature sensor data;

[0042] II. Automatically designing deep learning models using NAS

[0043] 1. Search space definition

[0044] Convolutional layer configuration: Define the parameters of the convolutional layer, including the size of the convolutional kernel, the stride, the padding method, and the number of convolutional layers.

[0045] The output formula of the convolutional layer is:

[0046]

[0047] wherein, K c is the size of the convolutional kernel, S c is the stride, P c is the padding method, N c is the number of convolutional layers, W l is the weight matrix of the convolutional kernel, b l is the bias term, O c,l is the output representation of each convolutional layer, l is the index of the convolutional layer, and m and n are variables used to index the elements inside the convolutional kernel or the pooling window.

[0048] Pooling layer configuration: Define the parameters of the pooling layer, including the size of the pooling window K p , the stride S p , and the number of pooling layers N p ; the output representation of the pooling layer is Op,m ;

[0049] The output formula of the max pooling layer is:

[0050]

[0051] where O p,m (x,y) is the output value corresponding to position (x,y) in the pooling layer m, max denotes taking the maximum value, i and j are the indices within the pooling window, and Kp is the size of the pooling window;

[0052] Fully connected layer configuration: define the parameters of the fully connected layer, including the number of nodes N fc , the activation function σ fc , and the number of layers L fc of the fully connected layer; the output of the fully connected layer is represented as O fc,n , where n is the index of the fully connected layer;

[0053] The output formula of the fully connected layer is:

[0054]

[0055] where O fc,n n represents the value of the nth output node in the fully connected layer, σ fc is the activation function used by the fully connected layer, ∑ p represents the weighted sum of all inputs, O p,m (p) represents the value of the mth output node in the previous layer at index p, W fc,n is the weight matrix of the fully connected layer, and b fc,n is the bias term;

[0056] 2. Search algorithm selection

[0057] Reinforcement learning-based NAS: In this step, a reinforcement learning-based NAS algorithm is used; the parameters of the policy network are θ, the parameters of the agent network are φ, the action sequence used is a t , and the state is s t , then the policy network updates the parameters by maximizing the following objective:

[0058] Reinforcement learning policy update formula:

[0059]

[0060] where θ t+1 is the updated policy parameter, θ t is the policy parameter at the current time, ∇ is the policy gradient, α is the learning rate, R t is the reward value, and π θis a strategy controlled by parameter θ;

[0061] Evolutionary algorithm NAS: In the evolutionary algorithm, each individual in the population represents a model architecture; the fitness function f(x) of the population contains the trade-off between accuracy and complexity, and the fitness function is defined as:

[0062] Fitness function formula:

[0063] f(x) = Accuracy(x) - λ·Complexity(x)

[0064] Where f(x) is the value of the fitness function, λ is the trade-off coefficient, Accuracy(x) represents the accuracy of the model, and Complexity(x) is the complexity of the model;

[0065] 3. Search process

[0066] Training model: In the search process, use the collected data set Train each model architecture generated by NAS, output the predicted insulation material positioning result

[0067] Evaluate the model: Calculate the accuracy Accuracy(x) of each architecture using the validation data set, and select the best model architecture x according to the fitness function * ;

[0068] Model selection: Finally, select the model architecture x with the maximum fitness function value f(x) * As the final architecture, and prepare for further optimization;

[0069] III. Accelerate architecture search combined with Bayesian optimization

[0070] 1. Introduction of Bayesian optimization

[0071] Surrogate model selection: Use Gaussian process (GP) as the surrogate model of Bayesian optimization, and the kernel function of the surrogate model is represented as k(x,x'), which models the performance distribution of the NAS search space

[0072] Gaussian process kernel function formula:

[0073]

[0074] Where k(x,x') is the kernel function in the Gaussian process, is the kernel function amplitude, l is the kernel function scale, x and x' are different architecture parameters, and exp is the exponential function;

[0075] 2. Optimization process

[0076] Sampling and inference: use GP to infer the distribution of the current optimal architecture parameters, and use the expected improvement (EI) strategy to select the next set of architecture parameters x to be evaluated next ;

[0077] Expected improvement (EI) strategy formula:

[0078]

[0079] Where μ(x) is the predicted mean of the GP model, σ(x) is the predicted variance, τ is the current best value, Φ and φ are the cumulative distribution function and probability density function of the standard normal distribution respectively, and EI(x) is the value of the expected improvement function at point x.

[0080] Update the surrogate model: after each evaluation, feed the new performance data y new to the GP model to update the surrogate model Narrow the search space step by step and approach the optimal architecture;

[0081] Four, model training and deployment

[0082] 1. Model training

[0083] Data input: use the preprocessed dataset as input data for the model, the goal is to predict the accurate position P of the insulation material k ;

[0084] Training process: use the final selected optimal model architecture x * to perform multiple rounds of iterative training, and adjust the model weights θ through the gradient descent algorithm;

[0085] Gradient update formula:

[0086]

[0087] Where θ t+1 is the updated parameter, θ t is the current parameter, α is the learning rate, is the gradient with respect to the parameter θ, is the loss function, P k is the actual position, is the predicted position;

[0088] 2. Model verification

[0089] Verification set test: use the verification set data to evaluate the performance of the model and calculate the accuracy Accuracy(x * ) and recall Recall(x * ) of the model;

[0090] Accuracy formula:

[0091]

[0092] where Accuracy(x * ) is the accuracy, TP, TN, FP, FN are true positive, true negative, false positive, false negative numbers respectively;

[0093] Error analysis: on the predicted results and the actual results P k , calculate the error mean E m and the standard deviation σ e to locate the systematic error of the model;

[0094] Error mean formula:

[0095]

[0096] where E m is the error mean, N I is the sample size, is the summation symbol, P k is the true value, is the predicted value;

[0097] 3. Model deployment

[0098] Model integration: integrate the trained model into the intelligent assembly system to process real-time image data and sensor data output the assembly decision of the insulation material;

[0099] System testing: test the stability of the model x * in the real assembly environment to ensure that the model can respond in real time and make accurate decisions;

[0100] Five, intelligent assembly and feedback optimization

[0101] 1. Real-time monitoring

[0102] Assembly monitoring: real-time capture of image data I real and sensor data F real , T real , input into the deep learning model x * , output the predicted insulation material positioning

[0103] where I real represents the new image data set after preprocessing, F real represents the force sensor data set after normalization processing, T realThis represents the new temperature sensor dataset after normalization.

[0104] Dynamic adjustment: based on the model's prediction results Dynamically adjust assembly parameter A adj (The position of the connecting rod and the filling density of the insulation cotton board) ensure assembly accuracy;

[0105] 2. Feedback Learning

[0106] Data collection: Continuously collect new assembly data Especially data under different temperature and force conditions, to enhance the model's adaptability;

[0107] Online learning: The model learns and optimizes itself online based on new data, gradually updating the model parameters θ through incremental learning. new This will further improve the model's performance.

[0108] Learn and update formulas online:

[0109]

[0110] Where, θ t+1 These are the updated model parameters, θ t These are the current model parameters, and η is the learning rate. It is the gradient with respect to the parameter θ. It is the loss function, P new For the target location of the new dataset, To predict the location.

[0111] Compared with the prior art, the present invention provides thermal insulation bricks and a method for manufacturing the same, which has the following beneficial effects:

[0112] 1. The thermal insulation brick and its manufacturing method, through the design of the bearing plate and the edge sealing plate, create an effective thermal insulation space inside the brick. At the same time, the parallel arrangement and mutual contact between the edge sealing plate and the brick frame ensure the tightness and sealing of the structure, reducing the thermal bridging effect. The manufacturing process of the thermal insulation brick is relatively simple, and mass production can be achieved through reasonable mold design and casting process. In addition, its multi-layered protective design also facilitates later maintenance and upkeep.

[0113] 2. The insulated brick and its manufacturing method, by designing a load-bearing plate and connecting frame inside the brick, wherein the connecting frame is filled with nano-roll material and insulation cotton board, these high-efficiency insulation materials can significantly reduce heat transfer, improve the thermal insulation performance of the brick, and thus reduce the energy consumption of the building.

[0114] 3、The thermal insulation brick and its manufacturing method, by setting reinforcing ribs and supporting frames outside the bearing plate, and wrapping the brick frame around the surface of the reinforcing ribs and the attached rods through pouring and solidification, this structure not only enhances the overall strength of the brick, but also improves its deformation resistance, ensuring the stability of the brick during use.

[0115] 4、The thermal insulation brick and its manufacturing method, by connecting the connecting rods designed inside the connecting frame, not only enhances the fixity of the nanometer coiled material and the thermal insulation cotton board, but also further enhances the stability of the entire structure by penetrating the supporting frame and extending to its inside.

[0116] 5、The thermal insulation brick and its manufacturing method, by setting an insulation layer and a protective layer outside the brick frame, where the insulation layer includes a waterproof coating and an outer plate, effectively preventing the damage of moisture penetration to the internal thermal insulation material; the protective layer provides additional protection through plastic film and sunscreen paint, prolonging the service life of the brick.

[0117] 6、The thermal insulation brick and its manufacturing method, by using thermal insulation resin material as the main material of the bearing plate, edge sealing plate and supporting frame, these materials not only have good thermal insulation performance, but also have sufficient strength and durability.

[0118] 7、By introducing intelligent assembly algorithm based on deep learning, the system can monitor and accurately locate the position of thermal insulation materials and their fixed parts in real time, ensuring high precision and consistency of assembly.

[0119] 8、Through real-time monitoring and feedback mechanism of intelligent algorithm, combined with the data of visual sensor and force sensor, the system can dynamically adjust the filling parameters to ensure the tight filling and uniform distribution of thermal insulation materials, thereby improving the consistency and stability in the assembly process. BRIEF DESCRIPTION OF DRAWINGS

[0120] Figure 1 is a structural schematic diagram of the present application;

[0121] Figure 2 is a top view structural schematic diagram of the present application;

[0122] Figure 3 is a three-dimensional structural separation schematic diagram of the present application;

[0123] Figure 4 is a three-dimensional schematic diagram of the local structure of the present application;

[0124] Figure 5 is a structural schematic diagram of the bearing plate of the present application;

[0125] Figure 6 is a structural schematic diagram of part A in the present application; Figure 2

[0126] ​Figure 7 Flowchart of the neural architecture search and the Bayesian optimization algorithm in the present application.

[0127] In the figure: 1, brick body; 2, bricklaying frame; 3, bearing plate; 4, edge plate; 5, reinforcing bar; 6, attachment rod; 7, support frame; 8, connecting frame; 9, nanometer coiled material; 10, thermal insulation cotton board; 11, connecting rod; 12, nut; 13, insulation layer; 14, protective layer; 15, waterproof coating; 16, outer layer plate; 17, plastic film; 18, sunscreen paint; 19, connecting hole; 20, positioning rod; 21, rock wool board. DETAILED DESCRIPTION

[0128] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0129] Example 1:

[0130] Reference Figures 1-6The utility model provides a kind of thermal insulation building brick and its manufacturing method, including brick body 1, brick body 1 is made of two building brick frames 2, the inner side of building brick frame 2 is provided with bearing plate 3, and the inner side of bearing plate 3 exists gap, and the both ends of bearing plate 3 extend to the both sides of building brick frame 2 respectively and are fixedly connected with edge plate 4, and edge plate 4 is parallel with building brick frame 2, and the inner side of edge plate 4 is in contact with the surface of building brick frame 2, and the material of edge plate 4 and bearing plate 3 is heat insulation resin material, and the outer side of bearing plate 3 is fixedly connected with reinforcing rib 5, and the side of reinforcing rib 5 away from bearing plate 3 extends to the inside of building brick frame 2, and the inner side of reinforcing rib 5 is fixedly connected with attached rod 6 located in the inside of building brick frame 2, and building brick frame 2 is wrapped on the surface of reinforcing rib 5 and attached rod 6 by pouring and solidification, and the outer side of bearing plate 3 is fixedly connected with support frame 7, and the side of support frame 7 away from bearing plate 3 is in contact with the inner side of building brick frame 2, and the material of support frame 7 is resin material, and the inner side of bearing plate 3 is provided with connecting frame 8, and the inside of connecting frame 8 is filled with thermal insulation material, and thermal insulation material includes nanometer coiled material 9 filled in the inside of connecting frame 8, and the both sides of nanometer coiled material 9 are provided with thermal insulation cotton board 10 located in the inside of connecting frame 8, and the four corners in the inside of connecting frame 8 are provided with connecting rod 11, and nanometer coiled material 9 is wrapped on the surface of connecting rod 11, and the both ends of connecting rod 11 are in sequence penetrated through thermal insulation cotton board 10, connecting frame 8, bearing plate 3 and support frame 7 and extend to the inside of support frame 7, and the both ends of connecting rod 11 are screw-threadedly connected with nut 12 located in the inside of support frame 7, and the outer side of building brick frame 2 is provided with insulating layer 13, and the outer side of insulating layer 13 is provided with protective layer 14, and insulating layer 13 includes waterproof coating 15 coated on the outer side of building brick frame 2, and the surface of waterproof coating 15 is adhered with outer layer plate 16, and protective layer 14 includes plastic film 17 fixedly connected on the surface of outer layer plate 16, and the surface of plastic film 17 is coated with sunscreen paint 18.

[0131] The utility model provides a kind of thermal insulation building brick and its manufacturing method, including the following steps:

[0132] S1: according to the design size of thermal insulation building brick, prepare corresponding mould, including the mould of building brick frame 2 and the mould of bearing plate 3, edge plate 4 and other components, using heat insulation resin material, bearing plate 3 and edge plate 4 and both sides reinforcing rib 5 and attached rod 6 are injection molded in mould, ensure that the inner side of bearing plate 3 has enough gap, so that subsequent filling thermal insulation material is filled;

[0133] S2: After the carrier plate 3 and the edge plate 4 are formed, necessary trimming and inspection are performed to ensure that the size and shape meet the design requirements. When the carrier plate 3 is processed, the support frame 7 is fixedly connected to the outer side of the carrier plate 3 and the carrier plate 3 is moved together with the edge plate 4 to the inside of the bricklaying frame 2 mold, ensuring that the edge plate 4 is parallel to the bricklaying frame 2 mold and tightly contacts the mold surface. The pouring material such as concrete or other suitable material of the bricklaying frame 2 is poured into the mold, so that the pouring material completely fills the gap between the support frame 7 and the mold, and at the same time, the material fully wraps the reinforcing bars 5 and the attachment rods 6. Wait for the pouring material to solidify and form the bricklaying frame 2.

[0134] S3: Install the connecting frame 8 on the inner side of the carrier plate 3, and fill the nanometer coiled material 9 and the thermal insulation cotton board 10 inside the connecting frame 8. Use the connecting rod 11 to penetrate the thermal insulation cotton board 10, the connecting frame 8, the carrier plate 3 and the support frame 7, and install the nut 12 inside the support frame 7 to fix the connecting rod 11 and the thermal insulation material.

[0135] S4: Apply a waterproof coating 15 on the outer side of the bricklaying frame 2 to form an insulation layer 13. After the waterproof coating 15 dries, stick the outer layer plate 16 on the surface of the waterproof coating 15 and attach the plastic film 17 to the surface of the outer layer plate 16. At the same time, apply a sunscreen coating 18 on the surface of the plastic film 17 to form a protective layer 14, thereby completing the production of the thermal insulation block.

[0136] In step S3, in order to ensure accurate filling and fixing of the thermal insulation material, an intelligent assembly algorithm is proposed. The specific implementation process is as follows:

[0137] I. Data collection and preprocessing

[0138] For example, the image data is I1(100,150) = 200, indicating that the pixel value at the 100th row and the 150th column is 200. The force measured by the force sensor is F1 = 150 N, and the temperature measured by the temperature sensor is T1 = 25℃, with a data range of -20 to 5℃.

[0139] Image normalization:

[0140]

[0141] II. Using NAS to automatically design deep learning models

[0142] Convolutional layer configuration

[0143] Assume the number of convolutional layers N c = 2, the convolution kernel size K c = 3 × 3, the stride S c = 1, and the padding mode is "same"

[0144] Assume the weight matrix W1 of the convolution kernel is:

[0145]

[0146] Convolution operation result:

[0147]

[0148] Pooling layer configuration

[0149] Assume the number of pooling layers N p = 1, pooling window size K p = 2x2, stride S p = 2. Max pooling result:

[0150] O p,1 (50,75)

[0151] = max{O c,1 (100,150), O c,1 (100,151), O c,1 (101,150), O c,1 (101,151)} Fully connected layer configuration

[0152] Assume the number of nodes in the fully connected layer N fc = 256, and the activation function is ReLU.

[0153] Output of the fully connected layer:

[0154]

[0155] Search algorithm selection

[0156] Using reinforcement learning policy update:

[0157] θ t+1 = θ t + 0.01x0.5xReward = θ t + 0.005xReward

[0158] Fitness function using evolutionary algorithm:

[0159] f(x) = 0.9 - 0.01x0.2 = 0.898

[0160] III. Accelerate architecture search combined with Bayesian optimization

[0161] Introduction of Bayesian optimization

[0162] Using Gaussian process kernel function calculation:

[0163] Kernel function is:

[0164]

[0165] Optimization process

[0166] Select the next set of architectures using the EI strategy:

[0167] The EI strategy is:

[0168]

[0169] Four, model training and deployment

[0170] Model training

[0171] Assume the learning rate of gradient descent is α = 0.001, and the loss function is mean square error L.

[0172] Training update:

[0173]

[0174] Model verification

[0175] Accuracy calculation:

[0176] Assume TP = 90, TN = 80, FP = 10, FN = 20, and the accuracy is:

[0177]

[0178] Five, intelligent assembly and feedback optimization

[0179] Real-time prediction of assembly position:

[0180] Assume the predicted position is:

[0181]

[0182] Feedback learning:

[0183]

[0184] Through the optimization of the insulation material filling and fixing system, the following specific numerical results are obtained:

[0185] Positioning error: 2.24 pixels, indicating that the system can accurately position the insulation material.

[0186] Filling error: 0.02, indicating that the filling tightness basically meets the expected standard.

[0187] Force error: 2N, indicating that the system maintains good mechanical stability during assembly.

[0188] Temperature error: 0.2℃, indicating that the system performs stably in temperature control.

[0189] Assembly efficiency is improved: assembly time is shortened by 10%, indicating that the production efficiency of the system is improved.

[0190] Embodiment 2:

[0191] Referring to Figure 1 and Figure 6 , the heat preservation brick and the manufacturing method thereof are basically the same as embodiment 1, and further, the connecting holes 19 are arranged on the two sides of the top and the bottom of the brick frame 2, the positioning rods 20 are inserted into the connecting holes 19, the rock wool board 21 is filled in the support frame 7, the protruding part is arranged on the inner side of the rock wool board 21, and the protruding part away from the rock wool board 21 extends to the inner side of the connecting rod 11.

[0192] By arranging the connecting holes 19 and the positioning rods 20, the connecting efficiency of the brick body 1 can be improved, and the inclination of the brick body 1 is prevented, the rock wool board 21 can further improve the heat preservation effect, and the protruding part on the inner side of the rock wool board 21 can extend to the inner side of the connecting rod 11, so that the connecting rod 11 after installation can effectively fix the rock wool board 21 and prevent the rock wool board 21 from separating from the support frame 7.

[0193] Specifically, during the working / using of the heat preservation brick, the load-bearing plate 3 and the edge sealing plate 4 are made of heat insulation resin material, which has good heat insulation performance and can significantly reduce the heat transfer through the brick body 1, thereby maintaining the stability of the indoor temperature, the gap on the inner side of the load-bearing plate 3 forms a heat insulation space, which blocks the path of heat transfer through direct conduction, and further enhances the heat preservation effect, the nano coiled material 9 and the heat preservation cotton board 10 are filled in the connecting frame 8 on the inner side of the load-bearing plate 3, which have excellent heat insulation performance and can effectively prevent heat transfer, the combination of the high-performance heat insulation property of the nano coiled material 9 and the good heat preservation effect of the heat preservation cotton board 10 makes the heat preservation performance of the entire heat preservation brick more superior, the reinforcing rib 5 and the attached rod 6 not only enhance the structural strength of the brick frame 2, but also form a stable support structure through the extension and wrapping inside the brick frame 2, and such a structure support helps to maintain the overall shape and stability of the heat preservation brick, thereby avoiding the decrease of the heat preservation performance caused by deformation.

[0194] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can make equivalent replacement or change according to the technical solution and the inventive concept of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. Thermal insulation block, comprising a block body (1), characterized in that, The brick body (1) is composed of two bricklaying frames (2), the inner side of the bricklaying frame (2) is provided with a bearing plate (3), the inner side of the bearing plate (3) has a gap, the two ends of the bearing plate (3) extend to the two sides of the bricklaying frame (2) respectively and are fixedly connected with the edge sealing plates (4), the edge sealing plates (4) are parallel to the bricklaying frames (2), the inner side of the edge sealing plate (4) is in contact with the surface of the bricklaying frame (2), and the materials of the edge sealing plate (4) and the bearing plate (3) are heat insulation resin materials; The outer side of the bearing plate (3) is fixedly connected with the reinforcing ribs (5), one side of the reinforcing rib (5) away from the bearing plate (3) extends to the inside of the bricklaying frame (2), and the inner side of the reinforcing rib (5) is fixedly connected with the attachment rods (6) located in the inside of the bricklaying frame (2); and the bricklaying frame (2) is wrapped on the surfaces of the reinforcing rib (5) and the attachment rod (6) by pouring and solidifying. The outer side of the bearing plate (3) is fixedly connected with the support frame (7), one side of the support frame (7) away from the bearing plate (3) is in contact with the inner side of the bricklaying frame (2), and the material of the support frame (7) is a resin material; The inner side of the bearing plate (3) is provided with a connecting frame (8), and the inside of the connecting frame (8) is filled with a heat preservation material; The heat preservation material comprises nanometer coiled material (9) filled in the inside of the connecting frame (8), and both sides of the nanometer coiled material (9) are provided with heat preservation cotton boards (10) located in the inside of the connecting frame (8).

2. The insulating block according to claim 1, characterized in that The four corners in the inside of the connecting frame (8) are provided with connecting rods (11), the nanometer coiled material (9) is wrapped on the surface of the connecting rod (11), both ends of the connecting rod (11) sequentially penetrate the heat preservation cotton board (10), the connecting frame (8), the bearing plate (3) and the support frame (7) and extend to the inside of the support frame (7), and both ends of the connecting rod (11) are screwedly connected with the nuts (12) located in the inside of the support frame (7).

3. The insulating block according to claim 1, characterized in that The outer side of the bricklaying frame (2) is provided with an insulation layer (13), the outer side of the insulation layer (13) is provided with a protective layer (14), the insulation layer (13) comprises a waterproof coating (15) coated on the outer side of the bricklaying frame (2), the surface of the waterproof coating (15) is adhered with an outer layer plate (16), and the protective layer (14) comprises a plastic film (17) fixedly connected to the surface of the outer layer plate (16).

4. The method for manufacturing thermal insulation bricks according to any one of the preceding claims, characterized in that, The method comprises the following steps: S1, according to the design size of the heat preservation brick, corresponding molds are prepared, including the mold of the bricklaying frame (2) and the mold of the bearing plate (3) and the edge sealing plate (4), heat insulation resin material is used, the bearing plate (3) and the edge sealing plate (4) and the reinforcing ribs (5) and the attachment rods (6) on both sides are injection molded in the mold, and it is ensured that the inner side of the bearing plate (3) has enough gaps for subsequent filling of the heat preservation material; S2, after the bearing plate (3) and the edge plate (4) are formed, necessary trimming and inspection are performed to ensure that the size and shape meet the design requirements, when the bearing plate (3) is processed, the support frame (7) is fixedly connected to the outer side of the bearing plate (3) and the bearing plate (3) is moved to the inside of the bricklaying frame (2) mold together with the edge plate (4), ensuring that the edge plate (4) is parallel to the bricklaying frame (2) mold and the inner side is in close contact with the mold surface, the pouring material of the bricklaying frame (2) is poured into the mold, so that the pouring material completely fills the gap between the support frame (7) and the mold, and at the same time the material fully wraps the reinforcing bars (5) and the attached rods (6), and the pouring material is waited to solidify and form the bricklaying frame (2); S3, a connecting frame (8) is installed on the inner side of the bearing plate (3), and nano coiled material (9) and thermal insulation cotton board (10) are filled in the connecting frame (8), a connecting rod (11) is used to penetrate the thermal insulation cotton board (10), the connecting frame (8), the bearing plate (3) and the support frame (7), and a nut (12) is installed in the support frame (7) to fix the connecting rod (11) and the thermal insulation material; S4, a waterproof coating (15) is coated on the outer side of the bricklaying frame (2) to form an insulation layer (13), after the waterproof coating (15) is dried, an outer plate (16) is adhered to the surface of the waterproof coating (15) and a plastic film (17) is attached to the surface of the outer plate (16), and a sunscreen coating (18) is coated on the surface of the plastic film (17) to form a protective layer (14), thereby completing the production and manufacturing of the thermal insulation block; In step S3, further comprising: using an intelligent assembly algorithm to ensure accurate filling and fixing of the thermal insulation material, and the specific implementation process is as follows: I. Data collection and preprocessing 1. Sensor arrangement: Visual sensor: high-resolution industrial cameras are arranged at key positions of the assembly line to capture real-time images of the thermal insulation material during assembly; the collected image data is represented as a variable wherein represents the serial number of the collected image Force sensor: force sensors are arranged at the contact positions of the connecting rod, nanometer coiled material, and thermal insulation cotton board, for monitoring the force conditions of each component during assembly; the data of the force sensor is represented as a variable wherein represents the number of the force sensor; Temperature sensor: Temperature sensors are placed in the assembly environment to monitor the effect of the ambient temperature on the insulation material filling and fixation; the temperature data are represented as a variable wherein represents the number of temperature sensors; 2. Data collection and labeling: Image data acquisition: image data in the process of assembling the thermal insulation material is collected in real time by the visual sensor, the frequency of collecting the image data is , the collected data set is represented as , wherein is the total number of image data; Force and temperature data collection: The data from the force sensor and the temperature sensor are collected simultaneously, the force sensor data set is denoted as , and the temperature sensor data set is denoted as , where are the total number of force and temperature data, respectively; Data labeling: label the collected image and sensor data, including the specific position, size, filling density, stress state and temperature state of the components ;​​​​ 3. Data preprocessing: Image preprocessing: the collected image data is processed through image enhancement, size scaling, and noise filtering in sequence, and the processed image data is represented as ; Data normalization: force data and temperature data were normalized to eliminate dimensional differences; the normalization formula is as follows: normalizing the image data: ​ in, Original image In position Pixel value at that location, These are the minimum and maximum values ​​of the image pixels, respectively. For normalized image data at location Pixel values; Normalizing force sensor data: ​ wherein, is the value of the normalized force sensor data, is the measured value of the raw force sensor at a certain time, is the minimum value of the force sensor data, is the maximum value of the force sensor data; Normalizing temperature sensor data: ​ wherein, is a value of the normalized temperature sensor data, is a measured value of the raw temperature sensor at a certain time, is a minimum value of the temperature sensor data, is a maximum value of the temperature sensor data; II. Using NAS to automatically design a deep learning model 1. Search space definition: Convolutional layer configuration: define the parameters of the convolutional layer, including the convolution kernel size, step, padding method and the number of convolutional layers; The output formula of the convolutional layer is: ； where c represents a convolution layer, is a convolution kernel size, is a stride, is a padding mode, is a number of convolution layers, is a weight matrix of a convolution kernel, is a bias term, is an output representation of each convolution layer, is an index of a convolution layer, m and n are variables for indexing elements inside a convolution kernel or a pooling window. Pooling layer configuration: define parameters of the pooling layer, including the pooling window size , the stride , and the number of pooling layers ; the output of the pooling layer is denoted as ; The output formula of the max pooling layer is: ； where p represents a pooling layer, This is the output value in the pooling layer m corresponding to the position (x, y), denotes taking the maximum value, i and j are indices within the pooling window, and Kp is the size of the pooling window; Fully connected layer configuration: define parameters of the fully connected layers, including the number of nodes , activation functions , and the number of layers of the fully connected layers ; the output of the fully connected layers is represented as , where n is the index of the fully connected layer; The output formula of the fully connected layer is: ； wherein, represents a fully connected layer, denotes the value of the n-th output node in the fully connected layer, is an activation function used by the fully connected layer, denotes the weighted sum over all inputs, denotes the value of the m-th output node in the previous layer at index p, is a weight matrix of the fully connected layer, is a bias term; 2. Search algorithm selection: Reinforcement Learning based NAS: In this step, a reinforcement learning based NAS algorithm is used; the parameters of the policy network are , the parameters of the agent network are , the action sequence used is , the state is , then the policy network updates the parameters by maximizing the following objective: Reinforcement learning strategy update formula: ； wherein, is the updated policy parameter, is the current time policy parameter, is the policy gradient, is the learning rate, is the reward value, is a policy controlled by parameters . Evolution algorithm NAS: in the evolution algorithm, each individual in the population represents a model architecture; the fitness function f(x) of the population includes the trade-off between accuracy and complexity, and the fitness function is defined as: Fitness function formula: ； wherein, is a value of the fitness function, is a weighting factor, denotes the accuracy of the model, is the complexity of the model; 3. Search process: Training model: In the search process, using the collected dataset Training each model architecture generated by the NAS, outputting a predicted insulation material positioning result ; Evaluate model: Calculate accuracy for each architecture using validation dataset and select the best model architecture according to fitness function ; Model selection: eventually, the model with the best fitness function value is chosen The largest model architecture As the final architecture and ready for further optimization; III. Combining Bayesian optimization to accelerate architecture search 1. Introduction of Bayesian optimization: Surrogate model selection: Gaussian process (GP) is used as the surrogate model of Bayesian optimization, and the kernel function of the surrogate model is represented as , modeling the performance distribution of the NAS search space ; Gaussian process kernel function formula: ； wherein, is a kernel function in a Gaussian process, is a kernel function amplitude, is a kernel function scale, and are different architecture parameters, is an exponential function; 2. Optimization process: Sampling and inference: use GP to infer the distribution of the current optimal architecture parameters and use the expected improvement (EI) strategy to select the next set of architecture parameters to be evaluated ; Expected improvement (EI) strategy formula: ； wherein is the predicted mean of the GP model, is the predicted variance, is the current best value, and are the cumulative distribution function and the probability density function of the standard normal distribution, respectively, is the value of the expected improvement function at point x; Update the agent model: after each evaluation, new performance data Feedback to the GP model, update the agent model , gradually narrow the search space, and approach the optimal architecture; IV. Model training and deployment 1. Model training: Data input: using pre-processed dataset As input data for the model, the goal is to predict the exact position of the insulation material ; Training process: using the final selected optimal model architecture Multiple rounds of iterative training are performed, adjusting the model weights through a gradient descent algorithm ; Gradient update formula: ​ wherein, is the updated parameter, is the current parameter, is the learning rate, is the gradient with respect to the parameter θ, is the loss function, is the actual position, is the predicted position; 2. Model verification: Validation set test: Using validation set data Evaluate the performance of the model, calculate the accuracy of the model And recall rate ; Accuracy formula: ​ wherein is the accuracy rate, TP, TN, FP, FN are the true positive, true negative, false positive, false negative numbers, respectively; Error analysis: on the predicted results vs. actual results Error analysis was performed to calculate the mean error and standard deviation to locate systematic errors in the model. Error mean formula: ​ wherein, is the mean error, is the number of samples, is the summation symbol, is the true value, is the predicted value; 3. Model deployment: Model integration: integrate the trained model into the intelligent assembly system to process real-time image data and sensor data , output the assembly decision of the thermal insulation material; System testing: Testing the model in a real assembly environment. The stability of the model ensures that it can respond in real time and make accurate decisions. V. Intelligent assembly and feedback optimization 1. Real-time monitoring: Assembly monitoring: real-time capture of image data during assembly and sensor data into a deep learning model outputs a predicted insulation positioning ; wherein, represents a new image data set after pre-processing, represents a force sensor data set after normalization processing, represents a new temperature sensor data set after normalization processing; Dynamic adjustment: according to the prediction results of the model , dynamically adjust the assembly parameters , ensure assembly accuracy; 2. Feedback learning: Data collection: new assembly data is continuously collected to enhance the adaptability of the model; Online learning: the model learns and optimizes itself online according to new data, and gradually updates the model parameters through incremental learning , further improving the performance of the model; Online learning update formula: ​ wherein, is the updated model parameter, is the current model parameter, is the learning rate, is the gradient with respect to the parameter θ, is the loss function, is the target position for the new data set, is the predicted position.

Images