Temperature strain dual-mode flexible sensor and preparation method thereof

Abstract

This invention relates to a dual-modal flexible temperature-strain sensor and its fabrication method. The sensor achieves efficient detection of temperature and strain signals by combining direct ink writing and screen printing processes. The sensor features a unique design, comprising a rigid substrate and a flexible substrate, housing a thermal resistance temperature sensor and a piezoresistive strain sensor respectively, ensuring effective separation and accurate measurement of temperature and strain signals. The use of flexible materials and a FA-BP neural network algorithm based on an improved firefly algorithm enhances stability and measurement accuracy under varying environments. Furthermore, the sensor can still operate normally at high temperatures, outperforming common flexible thin-film materials on the market. This sensor demonstrates significant application potential in smart wearable devices, health monitoring, and industrial automation, providing a new solution for achieving high-precision and high-stability temperature and strain monitoring.

Description

Technical Field

[0001] This invention belongs to the field of health monitoring sensor technology, and relates to a temperature strain dual-modal flexible sensor and its fabrication method. Background Technology

[0002] Against the backdrop of rapid advancements in flexible electronics technology, flexible sensors have become key components in fields such as the Internet of Things, wearable technology, and industrial automation. These sensors can monitor a variety of physical parameters in real time and maintain stable operation in diverse environments and configurations, significantly expanding their application scope.

[0003] However, traditional strain sensors perform poorly under high-temperature conditions, and the strain signals they measure are often coupled with temperature signals, making it difficult to meet the requirements of specific fields. Therefore, it is crucial to develop a strain sensor that can maintain high accuracy and sensitivity in high-temperature environments and avoid temperature signal coupling.

[0004] Although some related technologies have been proposed, such as the flexible resistive strain sensor, its fabrication method, and its application described in Chinese invention patent CN202310100867.3, their measurement results are easily affected by ambient temperature fluctuations, leading to signal coupling problems. Chinese invention patent CN202210887892.6 proposes a static calibration device and method for thin-film strain gauges deposited in situ on the surface of turbine engine rotor blades, but its manufacturing process is complex and costly. Chinese invention patent CN202311746126.9 proposes a Brillouin optical time-domain analysis method and system for temperature and strain decoupling, but the complex signal processing algorithms required by this system increase its complexity and computational load. Summary of the Invention

[0005] This invention discloses a temperature-strain dual-modal flexible sensor and its fabrication method. The sensor utilizes conductive materials and combines ink direct writing and screen printing processes to manufacture a temperature-decoupled flexible strain sensor that can independently measure strain and temperature signals. Furthermore, the Firefly Algorithm (FA) is used to optimize the BP neural network algorithm, constructing an FA-BP neural network algorithm to decouple the strain signal at high temperatures, thus overcoming the performance limitations of existing strain sensors in high-temperature environments.

[0006] A temperature-strain dual-modal flexible sensor includes a temperature-strain sensitive region and a signal processing circuit. The temperature-strain sensitive region includes a soft region substrate, on which a rigid region substrate and a soft region electrode are disposed. The rigid region electrode is disposed on the rigid region substrate. The rigid region electrode is a thermal resistance temperature sensor, and the soft region electrode constitutes a piezoresistive strain sensor. Both the rigid region electrode and the soft region electrode are electrically connected to the signal processing circuit. The signal processing circuit is used to acquire the resistance values ​​of the rigid region electrode and the soft region electrode, and to calculate the temperature and strain values ​​based on the resistance values ​​of the rigid region electrode and the soft region electrode.

[0007] Furthermore, the signal processing circuit includes a resistance calculation circuit and a microcontroller that are electrically connected. The resistance calculation circuit is used to collect the resistance values ​​of the rigid region electrode and the soft region electrode. The microcontroller is used to calculate the temperature value based on the collected resistance value of the rigid region electrode and the strain value based on the resistance value of the soft region electrode.

[0008] Furthermore, the microcontroller integrates a FA-BP neural network model for decoupling temperature and strain. This FA-BP neural network model is obtained through the following steps:

[0009] SA1. Place the temperature-strain sensitive area on a tensile testing machine. Stretch the temperature-strain sensitive area at different temperatures using the tensile testing machine. Simultaneously, use the resistance calculation circuit in the signal processing circuit to collect the resistance values ​​of multiple sets of soft region electrodes and rigid region electrodes. Use a data acquisition instrument to measure the true temperature of the temperature-strain dual-modal flexible sensor. Use the tensile testing machine to measure the true strain of the temperature-strain sensitive area to obtain the training set and test set.

[0010] SA2. Using the resistance values ​​of the soft and rigid electrodes in the training set as inputs, and the real temperature and real strain as outputs, construct and train a BP neural network; then validate it using data from the test set.

[0011] SA3. Obtain the optimal weights and thresholds using the firefly algorithm, and assign the optimal weights and thresholds to the BP neural network as its initial weights and thresholds to complete the construction of the FA-BP neural network model. Calculate the hidden layer weights of the FA-BP neural network based on the optimal weights and thresholds to obtain the FA-BP neural network model.

[0012] Furthermore, the SA2 includes the following steps:

[0013] SA201. Determine the BP neural network structure and activation function type as sigmoid function, and set the maximum number of iterations, learning rate, and initial hidden layer weights. Determine the number of neurons in the input layer, the number of neurons in the output layer, the number of hidden layers, and the number of neurons in the hidden layer.

[0014] SA202. Record the resistance values ​​of the soft region electrode and the rigid region electrode as input samples;

[0015] SA203. During the forward propagation process, the predicted temperature and strain values ​​are calculated based on the input samples, excitation function, and hidden layer weights.

[0016] SA204. Calculate the error between the output value and the actual value. If the error is less than the preset threshold or the maximum number of iterations is reached, stop training and save the network weights. Otherwise, use the error to update the hidden layer weights in the BP neural network and iterate until the condition is met.

[0017] Furthermore, the rigid region electrode is made of alumina powder.

[0018] A method for fabricating a temperature strain dual-modal flexible sensor includes the following steps:

[0019] S1. Add alumina powder and adhesion promoter to an organic solvent. The mass of alumina powder is 50wt%-70wt% of the organic solvent, and the mass of adhesion promoter is 1wt%-3wt% of the organic solvent. Disperse the alumina powder evenly to obtain alumina composite ink.

[0020] S2. Alumina composite ink is printed onto a soft area substrate using screen printing technology, and the alumina composite ink is sintered to obtain a rigid area substrate.

[0021] S3. Add nano-conductive ink to an electronic printer to print rigid region electrodes on a rigid region substrate to obtain a rigid electrode layer. Print soft region electrodes on a soft region substrate to obtain a soft electrode layer that is sensitive to strain temperature. Sinter the rigid electrode layer and the soft electrode layer to stabilize the resistance values ​​of the rigid region electrodes and the soft region electrodes.

[0022] S4. Connect the two ends of the rigid region electrode to the signal processing circuit with wires, and connect the two ends of the soft region electrode to the signal processing circuit with wires.

[0023] Furthermore, in step S2, the substrate of the soft region is a polyimide film resistant to 200°C.

[0024] Furthermore, in step S2, the sintering temperature range is 150℃-200℃, and the sintering time is 30min-60min.

[0025] Furthermore, in step S3, the printing speed of the rigid region electrode is 5-15 mm / s, and the air pressure is 100-180 kPa.

[0026] Furthermore, in step S3, the sintering temperature range is 150℃-200℃, and the sintering time is 60min-90min.

[0027] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0028] (1) The silver-based temperature strain dual-mode flexible sensor of the present invention achieves effective separation of strain signal and temperature signal through the design of independent regions in its structure and back-end data processing circuit.

[0029] The sensor employs independent rigid and flexible substrate structures to measure temperature and strain signals respectively. Electrodes on the rigid substrate detect temperature changes. The rigid substrate possesses good bending characteristics but also excellent tensile strength, allowing the electrodes to accurately measure temperature without being affected by strain. Electrodes on the flexible substrate detect strain changes. Since the strain signal measured by the electrodes on the flexible substrate is affected by temperature, the accurate temperature measured by the rigid substrate can be used for correction and compensation when measuring the strain signal, thus obtaining an accurate strain signal.

[0030] While backpropagation (BP) neural networks excel in many fields, their gradient descent method in decoupling control can lead to slow training speeds and a tendency to get trapped in local optima, especially for nonlinear, high-dimensional, or complex systems. The Firefly Algorithm (FA), with its global search capability and adaptability to complex solution spaces, effectively avoids local optima. By introducing the Firefly Algorithm to optimize the initial weights and thresholds of the BP neural network, the convergence speed and global optimization ability of the network are improved, resulting in better performance in decoupling control. The FA-BP neural network algorithm combines the advantages of both, using the Firefly Algorithm to optimize the parameters of the BP neural network, enhancing its generalization ability and robustness in decoupling control tasks. This combination not only accelerates the training process of the neural network but also improves the accuracy and stability of decoupling control.

[0031] The microcontroller in the sensor's signal processing circuit incorporates a built-in FA-BP neural network. This FA-BP neural network algorithm decouples the strain and temperature, solving the problem of traditional sensors being susceptible to data interference in complex environments. Through this hardware and software combined decoupling method, the sensor of this invention not only improves measurement accuracy but also expands its application range in various complex environments.

[0032] (2) The silver-based temperature strain dual-mode flexible sensor of the present invention can be fitted to a variety of surfaces. Utilizing flexible materials such as polyimide silver electrodes, it can easily adapt to various complex surfaces and shapes, achieving accurate measurements. The use of direct ink writing technology also allows for customized patterns and electrode printing on various surfaces. This adaptability enables the sensor to operate stably under varying environmental conditions, reducing measurement errors caused by shape mismatch and broadening its application range.

[0033] The present invention provides a method for fabricating a silver-based temperature strain dual-modal flexible sensor, which combines direct ink writing and screen printing processes. By directly writing conductive ink onto a flexible substrate, the complex steps of traditional photolithography and etching processes are avoided, thereby shortening the production cycle and reducing manufacturing costs. This process not only improves production efficiency but also further reduces overall production costs by reducing material waste and complex process steps.

[0034] Furthermore, in the silver-based temperature strain dual-mode flexible sensor prepared by this invention, both the metal nanoparticles and the alumina slurry are prepared on a high-temperature resistant flexible substrate and undergo high-temperature sintering treatment. Therefore, the sensor has good temperature resistance and can work normally at 200°C, which is significantly superior to common flexible films such as polyurethane, polydimethylsiloxane, and polyethylene terephthalate. Attached Figure Description

[0035] Figure 1 : A fabrication flowchart of a temperature strain dual-modal flexible sensor;

[0036] Figure 2 Exploded view of the structure of a silver-based temperature strain dual-mode flexible sensor;

[0037] Figure 3 Strain test results of a silver-based temperature strain dual-modal flexible sensor;

[0038] Figure 4 Temperature test results of a silver-based temperature strain dual-mode flexible sensor;

[0039] Figure 5 Flowchart for temperature-strain signal extraction from a silver-based dual-mode flexible temperature-strain sensor;

[0040] Figure 6 Schematic diagram of a temperature strain decoupling model based on the FA-BP neural network algorithm;

[0041] Figure 7 Strain test results of a silver-based temperature strain dual-modal flexible sensor at multiple temperatures;

[0042] Figure 8: Structure diagram of the FA-BP neural network algorithm;

[0043] Figure 9 Comparison of strain predictions and test set results using the FA-BP neural network algorithm;

[0044] Figure 10 Comparison chart of the effects of FA-BP neural network algorithm and BP neural network algorithm.

[0045] In the attached figures: 1. Rigid region substrate; 2. Rigid region electrode; 3. Soft region substrate; 4. Soft region electrode; 5. First wire connection; 6. Second wire connection; 7. Third wire connection; 8. Fourth wire connection. Detailed Implementation

[0046] To make the objectives and technical solutions of this invention clearer and easier to understand, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. The specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0047] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more. In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0048] Example 1

[0049] Reference Figure 2 and Figure 5A silver-based temperature-strain dual-modal flexible sensor includes a temperature-strain sensitive region and a signal processing circuit. The signal processing circuit includes a resistance calculation circuit and a microcontroller. The temperature-strain sensitive region includes a rigid substrate 1, a flexible substrate 3, a rigid electrode 2, and a flexible electrode 4. The flexible substrate 3 includes a first region and a second region. The rigid substrate 1 is fixed to the first region. The rigid electrode 2 is printed onto the rigid substrate 1 using a direct ink writing process, and the flexible electrode 4 is printed onto the second region using the same process. The rigid electrode 2 is a thermal resistance temperature sensor, sensitive to temperature; the flexible electrode 4 is a piezoresistive strain sensor, sensitive to strain. Four wire connection points are provided for the rigid electrode 2 and the flexible electrode 4, which are connected to the signal processing circuit via copper wires: a first wire connection point 5 at the first end of the rigid electrode 2, a second wire connection point 6 at the second end of the rigid electrode 2, a third wire connection point 7 at the first end of the flexible electrode 4, and a fourth wire connection point 8 at the second end of the flexible electrode 4.

[0050] At the first wire connection 5 and the second wire connection 6, copper wires are used to lead out the resistance signals and connect them to the signal processing circuit. The signal processing circuit includes a resistance calculation circuit and a microcontroller connected electrically. The resistance calculation circuit collects the resistance value of the rigid region electrode 2, and the microcontroller calculates the temperature value based on the collected resistance value of the rigid region electrode 2. At the third wire connection 7 and the fourth wire connection 8, copper wires are used to lead out the resistance signals and connect them to the signal processing circuit. The resistance calculation circuit in the signal processing circuit collects the resistance value of the flexible region electrode 4, and the microcontroller in the signal processing circuit calculates the strain value based on the collected resistance value of the flexible region electrode 4.

[0051] Preferably, the microcontroller in the signal processing circuit includes a communication module, through which strain values ​​and temperature values ​​are sent to the host computer.

[0052] Rigid electrode 2 on rigid substrate 1 is specifically used to detect temperature changes, while soft electrode 4 on soft substrate 3 is used to detect strain changes. Since the strain signal in the soft region is affected by temperature, the temperature data measured by the rigid electrode can be used to decouple the strain data when measuring the strain signal. Rigid substrate 1 is mainly made of alumina powder and has a rectangular structure. Rigid electrode 2 on rigid substrate 1 is used to detect temperature changes. Rigid substrate 1 has good bending characteristics but good tensile strength; the rigid electrode 2 on it can avoid the influence of strain and only measure temperature changes.

[0053] The flexible substrate 3 is made of polyimide (PI) and has a rectangular structure. The flexible electrode 4 on the flexible substrate 3 is used to detect strain changes. Since the strain signal measured by the flexible electrode 4 is affected by temperature, the temperature data measured by the rigid electrode 2 can be used for corresponding correction and compensation when measuring the strain signal.

[0054] Both the rigid region electrode 2 and the soft region electrode 4 adopt a grid structure. Both the rigid region electrode 2 and the soft region electrode 4 are silver electrodes.

[0055] Reference Figure 8 The microcontroller in the signal processing circuit integrates a FA-BP neural network algorithm for decoupling temperature strain. The training steps of the FA-BP neural network algorithm include:

[0056] SA1. The temperature-strain sensitive region of the silver-based temperature-strain dual-mode flexible sensor is placed on a tensile testing machine. At different temperatures, the temperature-strain sensitive region is stretched using the tensile testing machine. Simultaneously, the resistance values ​​of multiple sets of the flexible region electrode 4 and the rigid region electrode 2 are collected using the resistance calculation circuit in the signal processing circuit. The actual temperature of the sensor is measured by a data acquisition instrument using the thermocouple method, and the actual strain of the temperature-strain sensitive region is measured using the tensile testing machine. Training and testing sets are obtained. The training set includes the resistance values ​​of the flexible region electrode 4 and the rigid region electrode 2, along with the corresponding temperature and strain. The testing set includes the resistance values ​​of the flexible region electrode 4 and the rigid region electrode 2, along with the corresponding temperature and strain. Figure 7 As shown, the strain response of the sensor differs at different temperatures.

[0057] SA2. Construct a BP neural network. Using the resistance values ​​of the soft region electrode 4 and the rigid region electrode 2 from the training set as inputs, and the actual temperature and actual strain as outputs, construct and train the BP neural network. Validate the network using test set data.

[0058] SA3. Optimize the BP neural network using the firefly algorithm to construct the FA-BP neural network model. Utilizing the principle that brighter light attracts more fireflies, the optimal weights and thresholds are obtained using the firefly algorithm. These optimal weights and thresholds are then assigned to the BP neural network as its initial weights and thresholds, completing the construction of the FA-BP neural network model. The hidden layer weights of the FA-BP neural network are calculated based on the optimal weights and thresholds obtained from the firefly algorithm, resulting in the FA-BP neural network model. For example... Figure 10 As shown, the FA-BP neural network model converges faster and performs better than the BP neural network.

[0059] SA4. The FA-BP neural network model is embedded into the microcontroller of the signal processing circuit. The resistance calculation circuit of the signal processing circuit reads the resistance values ​​of the current soft region electrode 4 and the rigid region electrode 2, and inputs them into the FA-BP neural network algorithm of the microcontroller of the signal processing circuit to calculate the current strain value and temperature value of the sensor.

[0060] Step SA1 includes the following steps:

[0061] SA101: Connects the microcontroller in the signal processing circuit of the silver-based temperature strain dual-mode flexible sensor to a host computer, which is loaded with mobile and computer applications.

[0062] SA102: Place the silver-based temperature strain dual-mode flexible sensor on the tensile machine, and connect the silver-based temperature strain dual-mode flexible sensor and the resistance calculation circuit of the signal processing circuit.

[0063] SA103: The host computer transmits data collection commands to the microcontroller while simultaneously stretching the temperature-strain sensitive region at different temperatures using a tensile testing machine. At this time, the data acquisition instrument measures the actual temperature of the sensor using the thermocouple method. The actual temperature is transmitted back to the host computer by the data acquisition instrument. The actual strain of the temperature-strain sensitive region is measured using the tensile testing machine, and the actual strain is transmitted back to the host computer by the tensile testing machine.

[0064] SA104: The microcontroller receives commands and uses the resistance calculation circuit in the signal processing circuit to collect the resistance values ​​of the soft area electrode 4 and the rigid area electrode 2, and then transmits the two collected resistance data back to the host computer.

[0065] SA105: The host computer verifies and stores the received data, which includes the resistance value of the soft region electrode 4, the resistance value of the rigid region electrode 2, the actual temperature value, and the actual strain value. Then, it issues another data collection command to continue data collection.

[0066] SA106: After the data acquisition process is completed, the host computer randomly divides the stored data into a training set and a test set at an 8:2 ratio. The data content includes the resistance value of the soft region electrode 4, the resistance value of the rigid region electrode 2, the true temperature value, and the true strain value. In SA2, referencing... Figure 6 , including

[0067] Next steps:

[0068] SA201: Determine the BP neural network structure and activation function type as sigmoid, set the maximum number of iterations to 1000, and set the learning rate η=0.1 and the initial hidden layer weights. This involves determining the number of neurons in the input layer (m), the number of neurons in the output layer (n), the number of hidden layers (p), and the number of neurons in the hidden layer (k).

[0069] SA202: The resistance values ​​of the soft region electrode 4 and the rigid region electrode 2 are denoted as input sample X.

[0070]

[0071] X is the input matrix of the BP neural network. The resistance calculation circuit in the signal processing circuit reads the resistance value of the soft region electrode 4. The resistance value calculation circuit in the signal processing circuit reads the resistance value of the rigid region electrode 2.

[0072] SA203: During forward propagation, the predicted temperature and strain values ​​are calculated based on the input samples, excitation function, and hidden layer weights.

[0073]

[0074] The input variable matrix for the activation function, These are the input variables for the activation functions of each neuron in the hidden layer of the neural network.

[0075]

[0076] The output value matrix of the activation function. This represents the output value of the activation function for each neuron in the hidden layer of the neural network.

[0077]

[0078] W is the hidden layer weight matrix. Representing the Activation function input variables The first layer connected to the hidden layer Hidden layer weights.

[0079]

[0080] N is the weight matrix of the output layer. Representing the Activation function output variables The first layer connected to the hidden layer Hidden layer weights.

[0081] From the above, we can conclude that:

[0082]

[0083]

[0084] Therefore, the final estimated temperature and strain are output as follows:

[0085]

[0086] This represents the strain value estimated by the trained BP neural network. This represents the temperature value estimated by the trained BP neural network.

[0087] SA204: Calculate the error between the output value and the actual value. If the error is less than a preset threshold or the maximum number of iterations is reached, stop training and save the network weights; otherwise, update the hidden layer weights in the BP neural network using the error, and iterate until the condition is met.

[0088] The core of the temperature strain decoupling model based on BP neural network lies in adjusting the hidden layer weights between each neuron. The predicted value matrix output by the BP neural network... Let Y be the actual value matrix, E be the loss function, and η be the learning rate constant.

[0089] Therefore, it can be calculated that...

[0090]

[0091] in This represents the actual strain value collected in SA103. This represents the actual temperature value collected in SA103.

[0092] right Taking the derivative, we have

[0093]

[0094] The activation function uses the sigmoid function:

[0095]

[0096] Then we can get:

[0097]

[0098]

[0099]

[0100] Therefore, the weights of the K+1 hidden layer :

[0101]

[0102] Using K+1 hidden layer weights The predicted value matrix of the BP neural network output is calculated as follows: The loss function E is less than a pre-set threshold at the (n+1)th iteration. The calculation stops when the BP neural network training is complete, and the hidden layer weights at this point are obtained. .like Figure 9 As shown, the accuracy of the BP neural network algorithm is obtained by using the test set data for verification.

[0103] The SA3 mentioned above includes the following steps:

[0104] SA301: The basic idea of ​​the Firefly Algorithm (FA) is as follows: Within a certain range, there are randomly appearing, brightly glowing fireflies. These individuals are attracted to the light emitted by other fireflies; the brighter the firefly, the greater its attraction. Brighter fireflies randomly "move," while other fireflies "gather" towards them. Therefore, the brightness of a firefly is defined as:

[0105]

[0106] Where I0 is the original luminosity of the firefly's location, which is related to the objective function value; the higher the luminosity, the higher the objective function value. γ is the luminosity absorption coefficient. r represents the position X of any two fireflies i and j at their respective locations. i and X j The distance between them is also called the Cartesian distance.

[0107] The formula for calculating r is as follows:

[0108]

[0109] in , Let i and j be the spatial coordinates of firefly. , The k-th dimension coordinates; k=1, …,d; d is the dimension of the problem.

[0110] SA302: Because the attractiveness of fireflies is directly proportional to their brightness, the attractiveness β(r) is defined.

[0111]

[0112] Where β0 is the attraction when r = 0.

[0113] SA303: Firefly i will move towards other fireflies j that are brighter than itself, and the movement method is as follows.

[0114]

[0115] Let i be the position of firefly i after it moves; α is the step size factor; rand is a random real number uniformly distributed in [0, 1]; the second term on the right side of the equation represents the attraction of the firefly to the light intensity, and the third term is a random perturbation term.

[0116] SA304: Therefore, the equation for updating the position of the brightest firefly is as follows:

[0117]

[0118] in, The location of the brightest firefly; The updated position of the brightest firefly.

[0119] SA305: Sets the input matrix of the firefly algorithm to X from SA202, and the output to predicted temperature and predicted strain, initializing the firefly population for the algorithm. Each individual firefly corresponds to a threshold and weights in the BP neural network. Calculates the firefly fitness. As shown in equation (7):

[0120]

[0121] In the formula: Let p be the actual output value of all p-th training vectors, where p = 1, …, N; The predicted values ​​for all training vectors; N is the number of training vectors;

[0122] SA306: The optimal firefly individual is selected by changing the firefly's position, brightness, and decision radius. Then, the initial weights and thresholds are replaced to complete the construction of the FA-BP neural network.

[0123] This invention provides a flexible temperature and strain dual-mode sensor that is simple to manufacture and has high precision and high sensitivity. At the same time, the microcontroller in the signal processing circuit integrates a temperature and strain decoupling model based on the FA-BP neural network algorithm to calculate temperature and strain data, providing a more reliable monitoring solution.

[0124] This invention also provides a method for fabricating a silver-based temperature strain bimodal flexible sensor, combining two 3D printing processes: screen printing and direct ink writing. In the screen printing process, an alumina paste is prepared, and a uniform and strongly adherent alumina thin film substrate is formed through precise printing and sintering steps. In the direct ink writing process, a highly conductive nano-silver ink is used to construct the structure of the silver-based strain sensor. Precise patterning of the silver conductive circuitry is achieved by precisely controlling the ink spraying and drying process. The entire fabrication process must be carried out under strictly controlled environmental conditions to ensure the sensor's performance and reliability. This integrated process allows the silver-based temperature strain bimodal flexible sensor of this invention to adapt to various complex surfaces and shapes structurally, and to achieve high-precision, high-stability temperature and strain signal detection functionally, providing new possibilities for smart wearable devices, health monitoring, and industrial automation.

[0125] Reference Figure 1 A method for fabricating a silver-based temperature strain dual-mode flexible sensor includes the following steps:

[0126] S1. To prepare an alumina powder composite ink, alumina powder and an adhesion promoter are added to an organic solvent. The mass of the alumina powder is 50wt%-70wt% of the organic solvent, and the mass of the adhesion promoter is 1wt%-3wt% of the organic solvent. The mixture is then dispersed uniformly to obtain the alumina composite ink.

[0127] S2. Alumina composite ink is printed onto a flexible substrate using screen printing. The flexible substrate is made of a flexible polymer such as polyimide (PI). The composite ink is then sintered at high temperature to obtain a rigid substrate with a strain response much smaller than that of the flexible substrate.

[0128] S3. Add nano-silver conductive ink to the electronic printer to create patterned electrodes: Set the printing speed to 5-15 mm / s and the printing pressure to 100-180 kPa. Appropriate printing parameters are crucial; excessively high printing speed or low printing pressure may cause broken lines; excessively low printing speed or high printing pressure may cause ink buildup, resulting in overly thick and overlapping lines. Print rigid region electrodes 2 on the rigid region substrate 1 to obtain a rigid electrode layer that is sensitive to temperature but not strain. Print soft region electrodes 4 on the soft region substrate 3 to obtain a soft electrode layer that is sensitive to strain and temperature. Sinter the rigid and soft electrode layers to stabilize their resistance.

[0129] S4. After sintering and curing, the resistance signal is led out from the first wire connection 5 and the second wire connection 6 with copper wire and connected to the signal processing circuit. The resistance signal is led out from the third wire connection 7 and the fourth wire connection 8 with copper wire and connected to the signal processing circuit. The microcontroller in the signal processing circuit includes a communication module, which is used to send the strain value and temperature value to the host computer.

[0130] The host computer can be a mobile phone, computer, tablet or other terminal.

[0131] In S1, the adhesion promoter is ethyl cellulose.

[0132] In S1, the organic solvent is a mixed solution of anhydrous ethanol, acetone, terpineol, and dibutyl phthalate in proportions of 3 parts, 2 parts, 1 part, and 1 part.

[0133] In S2, the soft region substrate is a polymer film resistant to 200°C.

[0134] In S2, the printing process is screen printing.

[0135] In S2, the sintering temperature range is 150℃-200℃, and the sintering time is 30min-60min.

[0136] In S3, the conductive ink is a metal nano-conductive ink.

[0137] In S3, the method for fabricating patterned electrodes is ink direct writing process.

[0138] In S3, the sintering temperature range is 150℃-200℃, and the sintering time is 60min-90min.

[0139] Example 2

[0140] A method for fabricating a silver-based temperature strain dual-modal flexible sensor includes the following steps:

[0141] S1. To prepare an alumina powder composite ink, alumina powder and an adhesion promoter are added to an organic solvent. The mass of the alumina powder is 50 wt% of the organic solvent, and the mass of the adhesion promoter is 1 wt% of the organic solvent. The solvent is then dispersed uniformly to obtain the alumina composite ink.

[0142] S2. Alumina composite ink is printed on a soft area substrate using screen printing technology, and the composite ink is sintered at high temperature to obtain a rigid area substrate 1 with a strain response much smaller than that of the soft area substrate.

[0143] S3. Add conductive ink to the electronic printer to create patterned electrodes: Set the printing speed to 5 mm / s and the printing air pressure to 100 kPa. Print rigid region electrodes 2 on the rigid region substrate 1 to obtain a rigid electrode layer that is sensitive to temperature but not to strain. Print soft region electrodes 4 on the soft region substrate 3 to obtain a soft electrode layer that is sensitive to strain and temperature. Sinter the rigid electrode layer and the soft electrode layer to stabilize the resistance of the rigid region electrodes and the soft region electrodes.

[0144] In S1, the adhesion promoter is ethyl cellulose.

[0145] In S1, the organic solvent is a mixed solution of anhydrous ethanol, acetone, terpineol, and dibutyl phthalate in proportions of 3 parts, 2 parts, 1 part, and 1 part.

[0146] In S2, the soft region substrate is a polymer film resistant to 200°C.

[0147] In S2, the printing process is screen printing.

[0148] In S2, the sintering temperature range is 150℃ and the sintering time is 30min.

[0149] In S3, the conductive ink is a metal nano-conductive ink.

[0150] In S3, the method for fabricating patterned electrodes is ink direct writing process.

[0151] In S3, the sintering temperature range is 150℃ and the sintering time is 60min.

[0152] S4. After sintering and curing, the resistance signal is led out from the first wire connection 5 and the second wire connection 6 with copper wire and connected to the signal processing circuit. The resistance signal is led out from the third wire connection 7 and the fourth wire connection 8 with copper wire and connected to the signal processing circuit. The microcontroller in the signal processing circuit includes a communication module, which is used to send the strain value and temperature value to the host computer.

[0153] S5. The FA-BP neural network algorithm is used to decouple the strain and temperature. The neural network is trained and simulated in the relevant software environment, and the relevant program is written to achieve the goal.

[0154] like Figure 3 and Figure 4As shown in the corresponding curve, a silver-based temperature strain dual-mode flexible sensor fabricated based on the above steps can obtain the temperature value by directly detecting the change in resistance of the electrodes in the rigid region, while the electrodes on the soft region substrate are used to detect strain changes. Since the strain signal in the soft region is affected by temperature, the temperature data measured by the electrodes in the rigid region can be used for corresponding correction and compensation when measuring the strain signal.

[0155] Example 3

[0156] A method for fabricating a silver-based temperature strain dual-modal flexible sensor includes the following steps:

[0157] S1. To prepare an alumina powder composite ink, alumina powder and an adhesion promoter are added to an organic solvent. The mass of the alumina powder is 70 wt% of the organic solvent, and the mass of the adhesion promoter is 3 wt% of the organic solvent. The solvent is then dispersed uniformly to obtain the alumina composite ink.

[0158] S2. Alumina composite ink is printed on a soft area substrate using screen printing technology, and the composite ink is sintered at high temperature to obtain a rigid area substrate 1 with a strain response much smaller than that of the soft area substrate.

[0159] S3. Add conductive ink to the electronic printer to create patterned electrodes: Set the printing speed to 15mm / s and the printing air pressure to 180kPa. Print rigid region electrodes 2 on the rigid region substrate 1 to obtain a rigid electrode layer that is sensitive to temperature but not to strain. Print soft region electrodes 4 on the soft region substrate 3 to obtain a soft electrode layer that is sensitive to strain and temperature. Sinter the rigid electrode layer and the soft electrode layer to stabilize the resistance of the rigid region electrodes and the soft region electrodes.

[0160] In S1, the adhesion promoter is ethyl cellulose.

[0161] In S1, the organic solvent is a mixed solution of anhydrous ethanol, acetone, terpineol, and dibutyl phthalate in proportions of 3 parts, 2 parts, 1 part, and 1 part.

[0162] In S2, the soft region substrate is a polymer film resistant to 200°C.

[0163] In S2, the printing process is screen printing.

[0164] In S2, the sintering temperature range is 200℃ and the sintering time is 60min.

[0165] In S3, the conductive ink is a metal nano-conductive ink.

[0166] In S3, the method for fabricating patterned electrodes is ink direct writing process.

[0167] In S3, the sintering temperature range is 200℃ and the sintering time is 90min.

[0168] S4. After sintering and curing, the resistance signal is led out from the first wire connection 5 and the second wire connection 6 with copper wire and connected to the signal processing circuit. The resistance signal is led out from the third wire connection 7 and the fourth wire connection 8 with copper wire and connected to the signal processing circuit. The microcontroller in the signal processing circuit includes a communication module, which is used to send the strain value and temperature value to the host computer.

[0169] The host computer can be a mobile phone, computer, tablet or other terminal.

[0170] S5. The FA-BP neural network algorithm is used to decouple the strain and temperature. The neural network is trained and simulated in the relevant software environment, and the relevant program is written to achieve the goal.

[0171] like Figure 3 and Figure 4 As shown in the corresponding curve, a silver-based temperature strain dual-mode flexible sensor fabricated based on the above steps can obtain the temperature value by directly detecting the change in resistance of the electrodes in the rigid region, while the electrodes on the soft region substrate are used to detect strain changes. Since the strain signal in the soft region is affected by temperature, the temperature data measured by the electrodes in the rigid region can be used for corresponding correction and compensation when measuring the strain signal.

[0172] Example 4

[0173] A method for fabricating a silver-based temperature strain dual-modal flexible sensor includes the following steps:

[0174] S1. To prepare an alumina powder composite ink, alumina powder and an adhesion promoter are added to an organic solvent. The mass of the alumina powder is 60 wt% of the organic solvent, and the mass of the adhesion promoter is 2 wt% of the organic solvent. The solvent is then dispersed uniformly to obtain the alumina composite ink.

[0175] S2. Alumina composite ink is printed on a soft area substrate using screen printing technology, and the composite ink is sintered at high temperature to obtain a rigid area substrate 1 with a strain response much smaller than that of the soft area substrate.

[0176] S3. Add conductive ink to the electronic printer to create patterned electrodes: Set the printing speed to 10mm / s and the printing air pressure to 150kPa. Print rigid region electrodes 2 on the rigid region substrate 1 to obtain a rigid electrode layer that is sensitive to temperature but not to strain. Print soft region electrodes 4 on the soft region substrate 3 to obtain a soft electrode layer that is sensitive to strain and temperature. Sinter the rigid electrode layer and the soft electrode layer to stabilize the resistance of the rigid region electrodes and the soft region electrodes.

[0177] In S1, the adhesion promoter is ethyl cellulose.

[0178] In S1, the organic solvent is a mixed solution of anhydrous ethanol, acetone, terpineol, and dibutyl phthalate in proportions of 3 parts, 2 parts, 1 part, and 1 part.

[0179] In S2, the soft region substrate is a polymer film resistant to 200°C.

[0180] In S2, the printing process is screen printing.

[0181] In S2, the sintering temperature range is 180℃ and the sintering time is 50min.

[0182] In S3, the conductive ink is a metal nano-conductive ink.

[0183] In S3, the method for fabricating patterned electrodes is ink direct writing process.

[0184] In S3, the sintering temperature range is 180℃ and the sintering time is 80min.

[0185] S4. After sintering and curing, the resistance signal is led out from the first wire connection 5 and the second wire connection 6 with copper wire and connected to the signal processing circuit. The resistance signal is led out from the third wire connection 7 and the fourth wire connection 8 with copper wire and connected to the signal processing circuit. The microcontroller in the signal processing circuit includes a communication module, which is used to send the strain value and temperature value to the host computer.

[0186] The host computer can be a mobile phone, computer, tablet or other terminal.

[0187] S5. The FA-BP neural network algorithm is used to decouple the strain and temperature. The neural network is trained and simulated in the relevant software environment, and the relevant program is written to achieve the goal.

[0188] like Figure 3 and Figure 4As shown in the corresponding curve, a silver-based temperature strain dual-mode flexible sensor fabricated based on the above steps can obtain the temperature value by directly detecting the change in resistance of the electrodes in the rigid region, while the electrodes on the soft region substrate are used to detect strain changes. Since the strain signal in the soft region is affected by temperature, the temperature data measured by the electrodes in the rigid region can be used for corresponding correction and compensation when measuring the strain signal.

[0189] Example 5

[0190] A method for fabricating a silver-based temperature strain dual-modal flexible sensor includes the following steps:

[0191] S1. To prepare an alumina powder composite ink, alumina powder and an adhesion promoter are added to an organic solvent. The mass of the alumina powder is 50 wt% of the organic solvent, and the mass of the adhesion promoter is 3 wt% of the organic solvent. The solvent is then dispersed uniformly to obtain the alumina composite ink.

[0192] S2. Alumina composite ink is printed on a soft area substrate using screen printing technology, and the composite ink is sintered at high temperature to obtain a rigid area substrate 1 with a strain response much smaller than that of the soft area substrate.

[0193] S3. Add conductive ink to the electronic printer to create patterned electrodes: Set the printing speed to 10mm / s and the printing air pressure to 120kPa. Print rigid region electrodes 2 on the rigid region substrate 1 to obtain a rigid electrode layer that is sensitive to temperature but not to strain. Print soft region electrodes 4 on the soft region substrate 3 to obtain a soft electrode layer that is sensitive to strain and temperature. Sinter the rigid electrode layer and the soft electrode layer to stabilize the resistance of the rigid region electrodes and the soft region electrodes.

[0194] In S1, the adhesion promoter is ethyl cellulose.

[0195] In S1, the organic solvent is a mixed solution of anhydrous ethanol, acetone, terpineol, and dibutyl phthalate in proportions of 3 parts, 2 parts, 1 part, and 1 part.

[0196] In S2, the soft region substrate is a polymer film resistant to 200°C.

[0197] In S2, the printing process is screen printing.

[0198] In S2, the sintering temperature range is 180℃ and the sintering time is 50min.

[0199] In S3, the conductive ink is a metal nano-conductive ink.

[0200] In S3, the method for fabricating patterned electrodes is ink direct writing process.

[0201] In S3, the sintering temperature range is 200℃ and the sintering time is 80min.

[0202] S4. After sintering and curing, the resistance signal is led out from the first wire connection 5 and the second wire connection 6 with copper wire and connected to the signal processing circuit. The resistance signal is led out from the third wire connection 7 and the fourth wire connection 8 with copper wire and connected to the signal processing circuit. The microcontroller in the signal processing circuit includes a communication module, which is used to send the strain value and temperature value to the host computer.

[0203] The host computer can be a mobile phone, computer, tablet or other terminal.

[0204] S5. The FA-BP neural network algorithm is used to decouple the strain and temperature. The neural network is trained and simulated in the relevant software environment, and the relevant program is written to achieve the goal.

[0205] like Figure 3 and Figure 4 As shown in the corresponding curve, a silver-based temperature strain dual-mode flexible sensor fabricated based on the above steps can obtain the temperature value by directly detecting the change in resistance of the electrodes in the rigid region, while the electrodes on the soft region substrate are used to detect strain changes. Since the strain signal in the soft region is affected by temperature, the temperature data measured by the electrodes in the rigid region can be used for corresponding correction and compensation when measuring the strain signal.

[0206] In summary, the present invention aims to provide a flexible temperature strain dual-mode sensor that is simple to manufacture and possesses high precision and high sensitivity, thus providing a more reliable monitoring solution.

[0207] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A temperature strain dual-mode flexible sensor, characterized in that, The device includes a temperature-strain sensitive region and a signal processing circuit. The temperature-strain sensitive region includes a soft region substrate (3), on which a rigid region substrate (1) and a soft region electrode (4) are disposed. On the rigid region substrate (1), a rigid region electrode (2) is disposed. The rigid region electrode (2) is a thermal resistance temperature sensor, and the soft region electrode (4) constitutes a piezoresistive strain sensor. Both the rigid region electrode (2) and the soft region electrode (4) are electrically connected to the signal processing circuit. The signal processing circuit is used to acquire the resistance values ​​of the rigid region electrode (2) and the soft region electrode (4), and to calculate the temperature and strain values ​​based on the resistance values ​​of the rigid region electrode (2) and the soft region electrode (4). The signal processing circuit includes a resistance calculation circuit and a microcontroller connected by electrical connection. The resistance calculation circuit is used to collect the resistance value of the rigid region electrode (2) and the resistance value of the soft region electrode (4). The microcontroller is used to calculate the temperature value based on the collected resistance value of the rigid region electrode (2) and the strain value based on the resistance value of the soft region electrode (4). The rigid region electrode (2) is made of alumina powder; The microcontroller integrates an FA-BP neural network model for decoupling temperature and strain. This FA-BP neural network model is obtained through the following steps: SA1. Place the temperature-strain sensitive area on a tensile machine. At different temperatures, stretch the temperature-strain sensitive area with the tensile machine. At the same time, use the resistance calculation circuit in the signal processing circuit to collect the resistance values ​​of multiple soft area electrodes (4) and rigid area electrodes (2). Use a data acquisition instrument to measure the true temperature of the temperature-strain dual-mode flexible sensor. Use the tensile machine to measure the true strain of the temperature-strain sensitive area to obtain the training set and test set. SA2. Using the resistance values ​​of the soft region electrode (4) and the rigid region electrode (2) in the training set as inputs, and the real temperature and real strain as outputs, construct and train a BP neural network; use the test set data for verification. SA3. Obtain the optimal weights and thresholds using the firefly algorithm, and assign the optimal weights and thresholds to the BP neural network as its initial weights and thresholds to complete the construction of the FA-BP neural network model. Calculate the hidden layer weights of the FA-BP neural network based on the optimal weights and thresholds to obtain the FA-BP neural network model.

2. The temperature strain dual-mode flexible sensor according to claim 1, characterized in that, The SA2 includes the following steps: SA201. Determine the BP neural network structure and activation function type as sigmoid function, and set the maximum number of iterations, learning rate, and initial hidden layer weights. Determine the number of neurons in the input layer, the number of neurons in the output layer, the number of hidden layers, and the number of neurons in the hidden layer. SA202, Record the resistance values ​​of the soft region electrode (4) and the rigid region electrode (2) as input samples; SA203. During the forward propagation process, the predicted temperature and strain values ​​are calculated based on the input samples, excitation function, and hidden layer weights. SA204. Calculate the error between the output value and the actual value. If the error is less than the preset threshold or the maximum number of iterations is reached, stop training and save the network weights. Otherwise, use the error to update the hidden layer weights in the BP neural network and iterate until the condition is met.

3. The method for fabricating a temperature strain dual-mode flexible sensor according to claim 1, characterized in that, Includes the following steps: S1. Add alumina powder and adhesion promoter to an organic solvent. The mass of alumina powder is 50wt%-70wt% of the organic solvent, and the mass of adhesion promoter is 1wt%-3wt% of the organic solvent. Disperse the alumina powder evenly to obtain alumina composite ink. S2. Alumina composite ink is printed on a soft area substrate (3) using screen printing process, and the alumina composite ink is sintered to obtain a rigid area substrate (1). S3. Add nano-conductive ink to an electronic printer, print a rigid region electrode (2) on a rigid region substrate (1) to obtain a rigid electrode layer, and print a soft region electrode (4) on a soft region substrate (3) to obtain a soft electrode layer that is sensitive to strain temperature; sinter the rigid electrode layer and the soft electrode layer to stabilize the resistance values ​​of the rigid region electrode (2) and the soft region electrode (4). S4. Connect the two ends of the rigid region electrode (2) to the signal processing circuit with wires, and connect the two ends of the soft region electrode (4) to the signal processing circuit with wires.

4. The method for fabricating a temperature strain dual-modal flexible sensor according to claim 3, characterized in that, In S1, the organic solvent is a mixed solution of anhydrous ethanol, acetone, terpineol and dibutyl phthalate; the ratio is 3 parts anhydrous ethanol, 2 parts acetone, 1 part terpineol and 1 part dibutyl phthalate.

5. The method for fabricating a temperature strain dual-modal flexible sensor according to claim 3, characterized in that, In step S2, the soft region substrate (3) is a polyimide film resistant to 200°C.

6. The method for fabricating a temperature strain dual-mode flexible sensor according to claim 3, characterized in that, In step S2, the sintering temperature range is 150℃-200℃, and the sintering time is 30min-60min.

7. The method for fabricating a temperature strain dual-mode flexible sensor according to claim 3, characterized in that, In step S3, the printing speed of the rigid region electrode (2) is 5-15 mm / s, and the air pressure is 100-180 kPa.

8. The method for fabricating a temperature strain dual-mode flexible sensor according to claim 3, characterized in that, In step S3, the sintering temperature range is 150℃-200℃, and the sintering time is 60min-90min.

Images