Rapid heat shielding metamaterial based on neural network optimization and design method thereof

Abstract

The invention discloses a rapid heat shielding metamaterial based on neural network optimization and a design method thereof, and belongs to the field of heat shielding metamaterial structure design and preparation. The method comprises the following steps of: setting a comprehensive objective function by taking the heat shielding capability of a heat shielding metamaterial in a steady state and the time for achieving a heat shielding effect in a transient state as a comprehensive optimization objective; establishing a heat shielding metamaterial model, performing uniform orthogonality on the equivalent heat conductivity of the layered metamaterial, and selecting a plurality of groups of simulation test samples; obtaining Mv values under the condition of different equivalent heat conductivities based on steady-state solution; based on transient solution, obtaining tv values under the condition of different equivalent heat conductivities; building an artificial neural network model; inputting the equivalent thermal conductivity and the Mv value and the tv value under the corresponding condition into an artificial neural network model for multi-objective optimization calculation; and designing and establishing a heat shielding metamaterial. The design method is quick and simple, and the heat shielding metamaterial obtained by the method has a good heat shielding effect and can quickly reach a heat shielding state.

Description

Technical Field

[0001] The present invention belongs to the field of heat shielding metamaterial structure design and preparation, and specifically relates to a rapid heat shielding metamaterial based on neural network optimization and a design method thereof. Background Art

[0002] Thermal metamaterials refer to materials that achieve special thermal properties that real materials do not have, such as thermal invisibility, thermal shielding, thermal torsion, and thermal phantom, by designing the macroscopic or microscopic configuration inside the material. Traditional thermal metamaterial design is usually based on theoretical calculations or empirical design, which are slow and difficult to design. Moreover, the designed thermal metamaterials often have problems with thermal insulation or thermal mutation singularities, which greatly increases the difficulty of the actual design and preparation of thermal metamaterials.

[0003] Therefore, there is an urgent need to provide a new thermal metamaterial design method to achieve good heat shielding effect and perform heat shielding quickly. Summary of the invention

[0004] The present invention aims to solve the problems of complex design, long design cycle, many design variables, difficult and high cost of subsequent processing of existing heat shielding metamaterials, and proposes a rapid heat shielding metamaterial and a design method thereof based on neural network optimization.

[0005] The specific technical solutions adopted by the present invention are as follows:

[0006] In a first aspect, the present invention provides a design method for a rapid heat shielding metamaterial based on neural network optimization, the specific steps of which are as follows:

[0007] S1: Taking the thermal shielding ability of the thermal shielding metamaterial in steady state and the time to achieve the thermal shielding effect in transient state as the comprehensive optimization goals, set the comprehensive objective function; the thermal shielding ability in steady state is characterized by the surface integral Mv value of the absolute value of the temperature difference between each point in the shielding area in steady state and the ideal thermal shielding steady state temperature, and the time to achieve the thermal shielding effect in transient state is characterized by the time tv value for the average temperature of the shielding area to rise from the initial temperature to the ideal steady-state thermal shielding effect;

[0008] S2: Establish a heat shielding metamaterial model including multiple layers of concentric rings; take the equivalent thermal conductivity of each layer of concentric rings as a variable and set the variable range of the variable; make the equivalent thermal conductivity of multiple layers of concentric rings uniformly orthogonal and select several groups of simulation test samples;

[0009] S3: inputting the equivalent thermal conductivity of each layer of concentric rings in the several groups of simulation test samples in step S2 into the thermal shielding metamaterial model, and obtaining the Mv value under different equivalent thermal conductivities based on the steady-state solution, so as to evaluate the thermal shielding effect under this condition;

[0010] S4: inputting the equivalent thermal conductivity of each layer of concentric rings in the several groups of simulation test samples in step S2 into the thermal shielding metamaterial model, and obtaining tv values ​​under different equivalent thermal conductivities based on transient solution, so as to evaluate the speed of achieving the thermal shielding effect under this condition;

[0011] S5: Build an artificial neural network model;

[0012] S6: inputting the Mv value obtained in step S3 and the tv value obtained in step S4 into the artificial neural network model to perform multi-objective optimization calculation, and obtaining the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function;

[0013] S7: According to the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function obtained in step S6, a heat shielding metamaterial is designed and established based on the neutral inclusion principle.

[0014] Preferably, in step S1, the Mv value and the tv value are normalized and then added in a certain ratio to obtain a comprehensive optimization objective function.

[0015] Preferably, the heat shielding metamaterial model in step S1 is established using COMSO commercial software.

[0016] Preferably, the variable range of the thermal conductivity variable of each layer of concentric rings in step S2 is set to 2 to 394 W / (m·K).

[0017] Preferably, the heat shielding metamaterial model is composed of two layers of concentric rings; the two-layer heat shielding metamaterial model is established as follows:

[0018] A two-dimensional model is selected in the COMSO commercial software, and a solid heat transfer physical field is established; the solid heat transfer physical field is a square with a side length of 5 m, and concentric circular curves with radii of 1 m, 1.5 m and 2 m are sequentially arranged from the center of the circle outward, which are respectively recorded as the first curve, the second curve and the third curve, and the center of the circle coincides with the center position of the square; the ring formed by the second curve and the third curve is recorded as the first layer of concentric circular rings, and the ring formed by the first curve and the second curve is recorded as the second layer of concentric circular rings;

[0019] The temperature on the left side of the square is set to 0K, the temperature on the right side is set to 100K, and the upper and lower sides are thermal insulation conditions; the interior of the first curve is the heat shielding area, and the thermal conductivity of the heat shielding area is set to 394W / (m·K); the thermal conductivity of the outer area of ​​the third curve is 75-80W / (m·K), the equivalent thermal conductivity of the first layer of concentric rings is recorded as k1, and the equivalent thermal conductivity of the second layer of concentric rings is recorded as k2; the model of the two layers of heat shielding metamaterial is meshed.

[0020] Furthermore, in order to reduce the calculation time in step S4, the calculation stop condition is set as the average temperature of the shielded area reaches 45K, or the actual heat transfer time reaches 400 seconds.

[0021] Further, step S7 is specifically as follows:

[0022] S71: Solve the ratio of the two materials in each layer of concentric rings:

[0023]

[0024] l1+l2=1

[0025] Where: K s is the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function, K1 is the thermal conductivity of material one, in W / (m·K), K2 is the thermal conductivity of material two, in W / (m·K); l1 and l2 are the proportions of material one and material two to be solved in this layer of concentric rings respectively;

[0026] S71: According to the ratio of the two materials in each layer of concentric rings obtained in step S71, the density and specific heat capacity of each layer of concentric rings are obtained to complete the design of the heat shielding metamaterial.

[0027] Preferably, in step S5, the artificial neural network model is built using isight commercial software.

[0028] Preferably, the parameters of the artificial neural network model are as follows: bias 0.2, weight 0.5, number of layers 1, number of nodes 3.

[0029] In a second aspect, the present invention provides a heat shielding metamaterial obtained using the design method described in the first aspect.

[0030] Compared with the current design and preparation method of thermal shielding metamaterials, the present invention has the following beneficial effects:

[0031] (1) The thermal shielding metamaterial design method provided by the present invention introduces an artificial neural network model, takes the thermal shielding ability of the thermal shielding metamaterial in a steady state and the time to achieve the thermal shielding effect in a transient state as the comprehensive optimization target, and quickly optimizes the equivalent thermal conductivity of the concentric rings. This method can be used to design thermal shielding metamaterial structures with different numbers of concentric rings, with strong design flexibility and wide applicability.

[0032] (2) The thermal shielding metamaterial design method provided by the present invention can use the commercial software Comsol for simulation calculation and design. The preparation process is clear, the requirements for the external environment are low, and the requirements for theoretical basis and experimental conditions are extremely low. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 The double-layer heat shielding metamaterial model set in the commercial software COMSO in Example 1;

[0034] Figure 2 The grid division of the double-layer heat shielding metamaterial model in Example 1;

[0035] Figure 3 is the Mv value corresponding to the 1600 groups of simulation test samples selected in Example 1;

[0036] Figure 4 This is the thermal field distribution diagram of the heat shielding metamaterial obtained in Example 1;

[0037] Figure 5 This is a graph showing the change of the Mv value of the heat shielding metamaterial obtained in Example 1 over time;

[0038] Figure 6 When the double-layer heat shielding metamaterial model is set as a uniform material as a whole, the temperature field distribution state and isotherm diagram in the model are in steady state. DETAILED DESCRIPTION

[0039] The present invention is further described and illustrated below in conjunction with the accompanying drawings and specific embodiments. The technical features of each embodiment of the present invention can be combined accordingly without conflicting with each other.

[0040] Example 1

[0041] This embodiment prepares a fast heat shielding heat shielding metamaterial, and the specific preparation method is carried out according to the following steps:

[0042] 1. Determination of design goals and objective functions. The thermal shielding capability of the thermal shielding metamaterial in a steady state and the time to achieve the thermal shielding effect in a transient state are taken as comprehensive optimization goals, and a comprehensive objective function is set.

[0043] In this step, the heat shielding capability and the time to achieve the heat shielding effect are simultaneously selected as multi-objective optimization. The optimization purpose is to ensure that the heat shielding metamaterial has the best possible heat shielding effect while having the shortest possible time to achieve the steady-state heat shielding effect.

[0044] The thermal shielding capacity in steady state is characterized by the surface integral Mv value of the absolute value of the temperature difference between each point in the shielding area in steady state and the ideal thermal shielding in steady state. The smaller the Mv value, the better the steady-state thermal shielding effect. The time to achieve the thermal shielding effect in transient state is characterized by the time tv value when the average temperature of the shielding area rises from the initial temperature to the ideal steady-state thermal shielding effect. The smaller the tv value, the faster the thermal shielding effect can be achieved.

[0045] After the Mv value and the tv value are normalized, they are added in a certain ratio to obtain a comprehensive optimization objective function. In this embodiment, the contribution ratios of the Mv value and the tv value are 0.7 and 0.3 respectively.

[0046] 2. Establish a double-layer heat shielding metamaterial model based on Comsol commercial software.

[0047] Select the 2D model in Comsol and establish the solid heat transfer physics field.

[0048] like Figure 1 As shown, the solid heat transfer physical field is a square with a side length of 5m. Concentric circular curves with radii of 1m, 1.5m and 2m are arranged from the center of the circle to the outside, which are respectively recorded as the first curve, the second curve and the third curve. The center of the circle coincides with the center of the square; the ring formed by the second curve and the third curve is recorded as the first layer of concentric rings, and the ring formed by the first curve and the second curve is recorded as the second layer of concentric rings.

[0049] The temperature on the left side of the square is set to 0K, the temperature on the right side is set to 100K, and the upper and lower sides are thermal insulation conditions; the interior of the first curve is the heat shielding area, and the thermal conductivity of the heat shielding area is set to 394W / (m·K); the thermal conductivity of the area outside the third curve is 75~80W / (m·K), the equivalent thermal conductivity of the first layer of concentric rings is denoted as k1, and the equivalent thermal conductivity of the second layer of concentric rings is denoted as k2.

[0050] Mesh the model of two layers of heat shielding metamaterials, such as Figure 2 shown.

[0051] 3. Selection and sampling of variable design parameters.

[0052] The equivalent thermal conductivity k1 of the first layer of concentric rings and the equivalent thermal conductivity k2 of the second layer of concentric rings are taken as variables, and the variable ranges of the variables are set.

[0053] In this embodiment, in order to expand the design space as much as possible, the thermal conductivity of copper and organic silicon PMDS is referenced, the thermal conductivity of copper is 394W / (m·K), and the thermal conductivity of PDMS is 0.15W / (m·K). Therefore, the variable range of the thermal conductivity variable of each layer of concentric rings is selected to be 2 to 394W / (m·K).

[0054] The equivalent thermal conductivity of the double-layer concentric rings is uniformly orthogonal, and 1600 sets of simulation test samples are selected. The results are shown in Table 1. Figure 3 These are the Mv values ​​corresponding to the 1600 sets of simulation test samples.

[0055] Table 1 Simulation test sample results

[0056] <![CDATA[k1]]> <![CDATA[k2]]> Mv tv <![CDATA[k1]]> <![CDATA[k2]]> Mv tv 12 392 2.872700166 131.2 12 242 3.193709477 131.2 12 382 2.892556457 131.2 12 232 3.216541778 131.2 12 372 2.912645333 131.2 12 222 3.239441966 131.2 12 362 2.932966876 131.2 12 212 3.262366713 131.2 12 352 2.953520572 131.2 12 202 3.285262223 131.2 12 342 2.974305203 131.2 12 192 3.308061451 131.2 12 332 2.995318711 131.2 12 182 3.330680411 131.2 12 322 3.016558041 131.2 12 172 3.353013242 148.4 12 312 3.038018949 131.2 12 162 3.374925468 148.4 12 302 3.059695772 131.2 12 152 3.396244698 148.2 12 292 3.081581147 131.2 2 312 0.586652901 325.2 12 282 3.103665673 131.2 2 302 0.591538602 325.2 12 272 3.125937489 131.2 2 292 0.596507124 325.2 12 262 3.148381762 131.2 2 282 0.601560508 325.2 12 252 3.170980044 131.2 ... ... ... ...

[0057] 4. Calculation of thermal shielding capacity under steady state based on steady state solver.

[0058] The equivalent thermal conductivity k1 of the first layer of concentric rings and the equivalent thermal conductivity k2 of the second layer of concentric rings in 1600 sets of simulation test samples were input into the thermal shielding metamaterial model, and the steady-state solver in the comsol commercial software was used to obtain the Mv values ​​under different equivalent thermal conductivities to evaluate the thermal shielding effect in this case. The data of the thermal shielding effect under steady state are shown in Figure 4 shown.

[0059] 5. Calculation of thermal equilibrium velocity under transient conditions based on the transient solver.

[0060] The equivalent thermal conductivity k1 of the first layer of concentric rings and the equivalent thermal conductivity k2 of the second layer of concentric rings in 1600 sets of simulation test samples were input into the thermal shielding metamaterial model. The transient solver in the COMSOL commercial software was used to obtain the TV values ​​under different equivalent thermal conductivities, which were used to evaluate the speed at which the thermal shielding effect was achieved in this case.

[0061] In transient calculation, in order to save calculation cost, the upper limit of real heat transfer time is set to 400 seconds. If it exceeds 400 seconds and the shielded area has not reached thermal steady state, the calculation is abandoned and the maximum time is kept as 400 seconds. The transient calculation saves the real heat transfer time when the shielded area reaches thermal steady state as the parameter of the speed of reaching thermal steady state.

[0062] 6. Construction of artificial neural network based on isight commercial software.

[0063] The artificial neural network structure parameters are set as follows: bias 0.2, weight 0.5, number of layers 1, and number of nodes 3.

[0064] 7. Multi-objective optimization design based on artificial neural network model.

[0065] The data with tv values ​​exceeding 400 seconds are discarded, and the Mv values ​​obtained in step 4 and the tv values ​​obtained in step 5 are input into the artificial neural network model for multi-objective optimization calculation to obtain the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function.

[0066] After this step is optimized, the equivalent thermal conductivity of the first layer of concentric rings k1 = 11.4 W / (m·K) and the equivalent thermal conductivity of the second layer of concentric rings k2 = 392.3 W / (m·K) are obtained as preferred parameters.

[0067] 8. Based on the above optimization data, design and establish a double-layer heat shielding metamaterial model.

[0068] (1) Solve the ratio of the two materials in each layer of concentric rings:

[0069]

[0070] l1+l2=1

[0071] Where: K s is the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function, K1 is the thermal conductivity of material one, and the unit is W / (m·K), K2 is the thermal conductivity of material two, and the unit is W / (m·K); l1 and l2 are the proportions of material one and material two that need to be solved in this layer of concentric rings, respectively.

[0072] Substitute the equivalent thermal conductivity of the first concentric ring k1 = 11.4 W / (m·K) and the equivalent thermal conductivity of the second concentric ring k2 = 392.3 W / (m·K) into K s , material one is selected as copper, material two is selected as PDMS, and the proportion of copper and PDMS in the first layer of concentric rings and the second layer of concentric rings is solved.

[0073] (2) According to the ratio of the two materials in each layer of concentric rings obtained in step (1), the density and specific heat capacity of each layer of concentric rings are solved to obtain a heat shielding metamaterial.

[0074] This example prepares a heat shielding metamaterial with rapid heat shielding function. After testing, the heat shielding effect of the heat shielding metamaterial in steady state is as follows: Figure 4 , the Mv value is 2.88. The transient Mv changes with time and the center average temperature changes with the real heat transfer time as shown in the following Figure 5 At the same time, in order to more significantly compare the heat shielding effect of the heat shielding metamaterial, a simulation calculation is performed on a uniform material, and the thermal field distribution under steady state is shown in Figure 6 .

[0075] The thermal shielding metamaterial prepared in this embodiment has an Mv value of 2.88 in steady state, which shows excellent steady-state thermal shielding ability compared with the Mv value of 13.5 of the uniform material. At the same time, the thermal shielding metamaterial takes about 125 seconds to reach steady state in transient state, and can quickly achieve thermal steady state effect, with excellent comprehensive effect, achieving the required rapid thermal shielding effect.

[0076] The above-described embodiment is only a preferred solution of the present invention, but it is not intended to limit the present invention. A person skilled in the relevant technical field may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, any technical solution obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present invention.

Claims

1. A design method for rapid heat shielding metamaterial based on neural network optimization, characterized in that: The specific steps are as follows: S1: Taking the thermal shielding ability of the thermal shielding metamaterial in steady state and the time to achieve the thermal shielding effect in transient state as the comprehensive optimization goals, set the comprehensive objective function; the thermal shielding ability in steady state is characterized by the surface integral Mv value of the absolute value of the temperature difference between each point in the shielding area in steady state and the ideal thermal shielding steady state temperature, and the time to achieve the thermal shielding effect in transient state is characterized by the time tv value for the average temperature of the shielding area to rise from the initial temperature to the ideal steady-state thermal shielding effect; S2: Establish a heat shielding metamaterial model including multiple layers of concentric rings; take the equivalent thermal conductivity of each layer of concentric rings as a variable and set the variable range of the variable; make the equivalent thermal conductivity of multiple layers of concentric rings uniformly orthogonal and select several groups of simulation test samples; S3: inputting the equivalent thermal conductivity of each layer of concentric rings in the several groups of simulation test samples in step S2 into the thermal shielding metamaterial model, and obtaining the Mv value under different equivalent thermal conductivities based on the steady-state solution, so as to evaluate the thermal shielding effect under this condition; S4: inputting the equivalent thermal conductivity of each layer of concentric rings in the several groups of simulation test samples in step S2 into the thermal shielding metamaterial model, and obtaining tv values ​​under different equivalent thermal conductivities based on transient solution, so as to evaluate the speed of achieving the thermal shielding effect under this condition; S5: Build an artificial neural network model; S6: inputting the Mv value obtained in step S3 and the tv value obtained in step S4 into the artificial neural network model to perform multi-objective optimization calculation, and obtaining the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function; S7: According to the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function obtained in step S6, a heat shielding metamaterial is designed and established based on the neutral inclusion principle.

2. The design method according to claim 1, characterized in that: In step S1, the Mv value and the tv value are normalized and then added in a certain ratio to obtain a comprehensive optimization objective function.

3. The design method according to claim 1, characterized in that: The thermal shielding metamaterial model described in step S1 is established using COMSO commercial software.

4. The design method according to claim 1, characterized in that: In step S2, the variable range of the thermal conductivity variable of each layer of concentric rings is set to 2 to 394 W / (m·K).

5. The design method according to claim 1, characterized in that: The thermal shielding metamaterial model is composed of two layers of concentric rings. The two-layer thermal shielding metamaterial model is established as follows: A two-dimensional model is selected in the COMSO commercial software, and a solid heat transfer physical field is established; the solid heat transfer physical field is a square with a side length of 5 m, and concentric circular curves with radii of 1 m, 1.5 m and 2 m are sequentially arranged from the center of the circle outward, which are respectively recorded as the first curve, the second curve and the third curve, and the center of the circle coincides with the center position of the square; the ring formed by the second curve and the third curve is recorded as the first layer of concentric circular rings, and the ring formed by the first curve and the second curve is recorded as the second layer of concentric circular rings; The temperature on the left side of the square is set to 0K, the temperature on the right side is set to 100K, and the upper and lower sides are thermal insulation conditions; the interior of the first curve is the heat shielding area, and the thermal conductivity of the heat shielding area is set to 394W / (m·K); the thermal conductivity of the outer area of ​​the third curve is 75-80W / (m·K), the equivalent thermal conductivity of the first layer of concentric rings is recorded as k1, and the equivalent thermal conductivity of the second layer of concentric rings is recorded as k2; the model of the two layers of heat shielding metamaterial is meshed.

6. The design method according to claim 5, characterized in that: In order to reduce the calculation time in step S4, the calculation stop condition is set as the average temperature of the shielded area reaches 45K, or the actual heat transfer time reaches 400 seconds.

7. The design method according to claim 5, characterized in that: Step S7 is specifically as follows: S71: Solve the ratio of the two materials in each layer of concentric rings: l1+l2＝1 Where: K s is the equivalent thermal conductivity of each layer of concentric rings under the optimal condition of the objective function, K1 is the thermal conductivity of material one, in W / (m·K), K2 is the thermal conductivity of material two, in W / (m·K); l1 and l2 are the proportions of material one and material two to be solved in this layer of concentric rings respectively; S71: According to the ratio of the two materials in each layer of concentric rings obtained in step S71, the density and specific heat capacity of each layer of concentric rings are obtained to complete the design of the heat shielding metamaterial.

8. The design method according to claim 1, characterized in that: In step S5, the artificial neural network model is built using isight commercial software.

9. The design method according to claim 1, characterized in that: The parameters of the artificial neural network model are as follows: bias 0.2, weight 0.5, number of layers 1, number of nodes 3.

10. A heat shielding metamaterial obtained by using the design method according to any one of claims 1 to 9.

Images