Wide-working-condition and multi-dimensional force characterization flexible film piezoelectric sensor and signal analysis method

Abstract

The invention relates to the technical field of flexible piezoelectric sensors, and discloses a wide-working-condition multi-dimensional force representation flexible film piezoelectric sensor and a signal analysis method, after a pvdf piezoelectric film is cut in a paper-cut mode, regionalized electrode coating and covering are carried out, and multi-channel signal output is achieved. Therefore, characterization of multi-dimensional force can be realized on one film sensor, and meanwhile, the tension and conformal capability of the sensor are enhanced through the paper-cut design, so that the attachment requirement of a complex surface is met. A multi-layer composite structure is combined and comprises a flexible polymer protection layer and a flexible connection layer, and high sensitivity, excellent tensile property and mechanical durability are achieved through hot pressing and local ultrasonic welding. The sensing capability is enhanced, and the defect of poor tensile property of a traditional PVDF piezoelectric sensor is overcome. The preparation method is simple in process, low in cost and suitable for large-scale industrial production. In addition, the invention provides a deep learning analysis method for multi-channel signals, and classification and nonlinear fitting can be realized.

Description

Technical Field

[0001] The present invention relates to the field of sensor technology, and more specifically to a flexible thin film piezoelectric sensor that has been proven to be applicable to a variety of complex working conditions and has multi-dimensional force characterization such as base vibration, normal touch and tangential stretching, as well as a signal analysis method based on nonlinear waveform characteristics. The sensor can be applied to robot embodied perception, automotive intelligent cockpit and structural status monitoring and other fields. Specifically, it relates to a wide working condition, multi-dimensional force characterization flexible thin film piezoelectric sensor and a signal analysis method. Background Art

[0002] The unique advantage of piezoelectric sensors in converting mechanical energy directly into electrical signals due to their nonlinearity is widely evaluated in the fields of mechanical testing, health structure monitoring, and smart medical care. In recent years, with the rapid development of flexible electronic technology, the potential application of piezoelectric sensors in wearable devices, soft robots, and complex morphology detection has received increasing attention. At present, the materials of piezoelectric sensors mainly include inorganic materials (such as piezoelectric ceramics and piezoelectric crystals) and organic materials (such as polyvinylidene fluoride PVDF). Piezoelectric ceramics (such as PZT) have high sensitivity and stability and are suitable for high-performance applications, but are usually brittle and difficult to adapt to large deformation scenarios. Organic materials such as PVDF have good flexibility and are easy to make into thin films, but do not have excellent stretchability, which limits their application in flexible and wearable devices. In addition, traditional PVDF piezoelectric film sensors can usually only detect signals at a single point or a single channel, and can only reflect relatively single contact information, which is difficult to meet the current high-performance tactile sensors for the simultaneous characterization of multi-dimensional forces.

[0003] Multi-dimensional force sensors are currently an important tool for achieving accurate measurements in complex mechanical environments. Compared with traditional single-dimensional force sensors, multi-dimensional force sensors can simultaneously occupy force gradients in multiple directions in an important structure of a sensor, so that the magnitude, direction and distribution of the force can be comprehensively characterized. At present, although multi-dimensional force sensors have shown their application value in many fields, their environmental applicability still faces severe challenges, especially due to the limitations of device design, the performance and stability in complex environments are difficult to meet actual needs. Existing multi-dimensional force sensors are usually based on rigid materials and fixed structural designs, which leads to insufficient adaptability when bending, attaching, dynamically deforming and under extreme working conditions (such as high temperature, low temperature, heavy load, underwater and corrosive environment). In addition, the current manufacturing of multi-dimensional force sensors is relatively complex and the device cost is high, which further limits the large-scale promotion of high-performance tactile sensing solutions. At the same time, the analysis of multi-channel signals is also a bottleneck in current technology: existing analysis methods mostly rely on traditional algorithms, only consider the amplitude of the sensor signal as the judgment standard, cannot fully reflect the comprehensive contact information, ignore the nonlinear characteristics such as waveform and spectrum, and have application problems for target detection and feature fitting in complex mechanical environments.

[0004] Therefore, under the premise of considering the manufacturing cost and applicability of multiple complex working conditions, it is important to develop a piezoelectric sensor based on low-cost piezoelectric materials with multi-dimensional force characterization capabilities, and develop corresponding signal analysis methods to achieve comprehensive mechanical behavior reflection by processing the nonlinear characteristics of the waveform, which has important research value and application potential. Summary of the invention

[0005] In order to overcome the technical defects of the above-mentioned flexible piezoelectric sensors, the present invention provides a wide-operating-condition, multi-dimensional force characterization flexible piezoelectric film sensor and a signal analysis method. The key to the present invention is to perform customized paper-cut cutting on the complete PVDF piezoelectric film, while enhancing the sensor's stretching ability and conformal attachment ability, to achieve multi-channel sensing with independent output through regionalized electrode coating, and to achieve multi-dimensional force characterization based on the difference in multi-channel signals under different characteristic forces. At the same time, the present invention designs a multi-layer composite structure for packaging, and the preparation method provided has a simple operating procedure, low process cost, and can be mass-produced. For the above-mentioned flexible piezoelectric film sensor with multi-dimensional force characterization capability, the present invention also provides a signal analysis method based on the LSTM model, which can achieve accurate target detection by processing the nonlinear characteristics of the waveform.

[0006] In the first aspect, the present application provides a flexible piezoelectric film sensor with wide working conditions and multi-dimensional force characterization, which adopts the following technical solution:

[0007] A wide-operating-condition, multi-dimensional force characterization flexible piezoelectric film sensor, comprising an upper flexible polymer protective layer, a partially hollowed-out upper flexible connection layer, a regionally separated upper electrode layer, a paper-cut piezoelectric film layer, a regionally separated lower electrode layer, a partially hollowed-out lower flexible connection layer, a lower flexible polymer protective layer, and multiple groups of enameled wires for multi-channel signal output, which are sequentially arranged;

[0008] The upper flexible polymer protective layer and the lower flexible polymer protective layer are both made of high temperature resistant polymer films.

[0009] By adopting the above technical solution, the upper flexible polymer protective layer and the lower flexible polymer protective layer adopt high temperature resistant polymer film, so that the sensor can work normally in a relatively high temperature environment, avoiding the degradation, deformation and even damage of the protective layer due to the temperature rise, thereby significantly improving the overall stability and durability of the sensor and extending its service life. It can also effectively isolate external dust, water vapor and other substances that may corrode the internal structure, prevent these impurities from entering the sensor and affecting the performance of key components such as the electrode layer and the piezoelectric film layer, and ensure that the sensor can work stably and accurately for a long time and maintain good sensing performance.

[0010] The piezoelectric film layer is cut by paper cutting, and the cutting patterns include straight line cutting marks, fractal cutting marks and corresponding evolutionary patterns confirmed after force-electric coupling analysis, which form different deformation effects under the action of multi-dimensional forces.

[0011] By adopting the above technical solution, the straight shear marks, fractal shear marks and corresponding evolutionary patterns confirmed after the electromechanical coupling analysis form a paper-cut area that can be deformed in a complex manner, which changes the structural morphology of the piezoelectric film layer, so that it produces different in-plane and out-of-plane deformation effects when subjected to different types of external forces. For example, when subjected to normal pressure, the piezoelectric film produces a slight deformation in the normal direction; when subjected to in-plane stretching, the hinge structure formed by the paper-cut will experience in-plane elastic deformation, out-of-plane buckling, and torsion as the degree of stretching increases; when subjected to shear force and torsion force, the hinge structure formed by the paper-cut will also produce deformation characteristics different from the above behaviors. Since the sensing mechanism of piezoelectric materials is to generate electric charges when deformed by force, under the action of the above different characteristic forces, the paper-cut piezoelectric film will generate surface charges in completely different areas, and the accumulation of surface charges is not a completely linear process, which will cause the electrical signal output to present different nonlinear waveforms. This special cutting method also gives the piezoelectric film layer better deformability, so that it can maintain its stable performance and work normally when facing bonding surfaces of different shapes or undergoing large deformations such as bending and stretching. It meets the requirements of mechanical flexibility well and facilitates its conformal attachment to the surfaces of various irregular shapes for measurement and monitoring, such as attaching it to aerospace parts with complex surfaces to monitor their structural stress changes.

[0012] Optionally, the regionalized divided upper electrode layer and the regionalized divided lower electrode layer are conductive silver paste or PEDOT:PSS applied in the shear division area of ​​the piezoelectric film layer; the lower electrode layer and the upper electrode layer have the same coating area; the upper electrode layer and the lower electrode layer are not connected.

[0013] By adopting the above technical solution, conductive silver paste or PEDOT:PSS is used as the electrode layer material. Whether it is the conductive silver paste with its excellent conductive properties of metallic silver, or the good conductivity of PEDOT:PSS, an organic conductive polymer, it can ensure that the electrical signal generated in the piezoelectric film layer can be efficiently and stably conducted. The piezoelectric film cut by paper-cutting will generate surface charges in completely different areas under the action of different characteristic forces. Therefore, these areas are selected for the coating of the electrode layer materials that are not interconnected, and each is connected to the enameled wire for signal output, thereby ensuring that the characteristics of the multi-dimensional force are fully reflected under the joint action of multi-channel signals. In addition, the conductive silver paste and PEDOT:PSS can maintain a relatively stable connection state as the piezoelectric film layer is deformed under the action of different forces, avoiding problems such as signal transmission interruption or instability due to the loose combination of the electrode and the film, ensuring that the sensor can still work normally when experiencing various complex deformation conditions, and adapting to application scenarios with high flexibility requirements such as flexible wearable devices and bendable electronic devices. The lower electrode layer and the upper electrode layer have the same coating area. This design ensures that the upper and lower electrodes have consistency and correspondence in the electrical signal collection of the same area of ​​the piezoelectric film layer, and can accurately obtain the changes in electrical energy caused by external forces in the area. At the same time, the upper electrode layer is not connected to the lower electrode layer, which effectively prevents possible short circuits between the electrodes and avoids signal interference caused by abnormal conductive paths between the electrodes, so that the electrical signals collected from the upper and lower electrode layers can independently and accurately reflect the changes in physical quantities in the corresponding areas, further improving the accuracy and reliability of sensor signal collection, and helping to more accurately analyze and judge the changes in the state of the external environment or object monitored by the sensor.

[0014] The partially hollowed-out upper flexible connection layer and the partially hollowed-out lower flexible connection layer are hot-melt adhesive films, and the hollowed-out area is arranged in the cutting area of ​​the above-mentioned paper-cutting piezoelectric film.

[0015] By adopting the above technical scheme, the upper flexible connection layer and the lower flexible connection layer are made of hot-melt adhesive film and the corresponding paper-cut area of ​​the piezoelectric film layer is partially hollowed out. Firstly, on the basis of ensuring that each layer can be well connected and fixed, unnecessary material accumulation is reduced, making the entire sensor structure lighter, more flexible and easier to fit; secondly, when the piezoelectric film layer is greatly deformed due to external force, the hollow design provides it with sufficient strain space, and will not hinder the normal deformation of the film layer due to the limitation of the connection layer, ensuring that the sensor can still output signals stably under large deformation conditions, and can be applied to some scenarios that need to withstand frequent large deformations, such as wearable devices for signal collection and monitoring under conditions of frequent bending and stretching during human movement; finally, the hollowed-out flexible connection layer makes the piezoelectric film cut by paper-cutting touch the flexible polymer protective layers on both sides at different parts when subjected to forces of different dimensions, thereby generating electrical signal output corresponding to the collision behavior, further enhancing the multi-dimensional force characterization capability of the sensor. The hot-melt adhesive film itself has a certain viscosity and good electrical insulation properties. When connecting the various layers of the structure, it can not only stabilize the relative position between the layers to avoid inter-layer displacement affecting the signal transmission path, but also prevent the abnormal conductive contact between the layers from interfering with the normal output of the signal, ensuring that the multi-channel electrical signal generated from the piezoelectric film layer can be smoothly transmitted to the enameled wire through the electrode layer, connecting layer, etc. and output to the outside, thereby ensuring the stability and reliability of the sensor's multi-channel signal transmission.

[0016] The enameled wire has an inner conductor made of copper material and an outer insulating coating of polyurethane paint, and two enameled wires outputting the same sensing channel are wound around each other.

[0017] By adopting the above technical solution, the two wires intertwined with each other can reduce the electromagnetic interference and crosstalk effects of the sensor during use. For example, when the wire moves in a magnetic field, an induced current will be generated, thereby affecting the accuracy of the sensor. After being intertwined, the two enameled wires of the same sensing channel will generate the same current when moving in a magnetic field. Due to the differential principle of the piezoelectric sensor, the influence of the magnetic field is avoided, thereby enhancing the use effect in a magnetic field environment.

[0018] Optionally, the high temperature resistant polymer film includes a thermoplastic polyurethane (TPU) film, a polyetheramide elastomer (PEBA) film or a polyurethane elastomer (PU) film, and the upper flexible polymer protective layer and the lower flexible polymer protective layer are made of the same or different materials; the thickness range of the upper flexible polymer protective layer and the lower flexible polymer protective layer are both 10-300 microns.

[0019] By adopting the above technical solutions, the high temperature resistant polymer film includes thermoplastic polyurethane (TPU) film, polyetheramide elastomer (PEBA) film or polyurethane elastomer (PU) film, and the upper and lower flexible polymer protective layers can be made of the same or different materials. This enables the sensor to flexibly select the appropriate material combination according to the specific use requirements and environmental characteristics. For example, if the sensor needs to work in an environment with certain chemical corrosion, a polyetheramide elastomer (PEBA) film with relatively better chemical resistance can be selected; while if more emphasis is placed on flexibility and cost factors, a thermoplastic polyurethane (TPU) film may be a better choice. Different materials have their own unique performance advantages. By selecting the same or different materials to form the upper and lower protective layers, it is possible to complement each other and optimize the overall comprehensive performance of the sensor. For example, the upper layer uses a material with good wear resistance to resist external friction, and the lower layer uses a material with better moisture resistance to prevent water vapor from invading from the bottom, thereby comprehensively improving the sensor's ability to resist adverse external factors, ensuring the normal operation of key internal structures, and extending the effective working time of the sensor. The thickness range of the upper flexible polymer protective layer and the lower flexible polymer protective layer is limited to 10-300 microns, and this thickness range design is very clever. The thinner lower limit (10 microns) can ensure that the protective layer will not excessively increase the overall rigidity of the sensor due to excessive thickness, affecting its mechanical flexibility, so that the sensor can still maintain good conformal adhesion when attached to surfaces of various shapes, and adapt to scenarios such as wearable devices fitting the curved surface of the human body that require highly flexible deformation. The relatively thick upper limit (300 microns) ensures that there is enough material to provide reliable protection for the internal structure, block possible physical damage, chemical erosion, and intrusion of water vapor, dust, etc. from the outside world, maintain the stable performance of the sensor, and play an effective protective role in resisting certain degrees of impact, scratches, and chemical contact in industrial environments.

[0020] Optionally, the piezoelectric film layer is a polarized PVDF film with a thickness ranging from 14 to 100 microns.

[0021] By adopting the above technical solution, polarized PVDF film is selected as the piezoelectric film layer because PVDF itself has a good piezoelectric effect after polarization treatment. It can effectively convert mechanical energy into electrical energy, or convert electrical energy into mechanical energy. When the sensor is subjected to external pressure, vibration and other mechanical effects, it can accurately generate corresponding electrical signals and has a keen response to changes in external physical quantities, thereby ensuring that the sensor has a high detection sensitivity. It can play an excellent sensing role in many application scenarios such as acoustic sensors for sound detection and pressure sensors for sensing small pressure changes. PVDF film has excellent chemical stability, can withstand the erosion of a variety of chemical substances, and can maintain its own stable performance in a more complex chemical environment. This means that sensors using this film can be used in some scenarios where they may be exposed to chemical substances such as acids and alkalis. For example, when monitoring the status of equipment in a chemical production workshop, there is no need to worry about the film failing due to chemical corrosion, ensuring that the sensor works reliably for a long time. PVDF film also exhibits good biocompatibility, making it particularly suitable for sensor production in the biomedical field. For example, when making implantable medical sensors or wearable health monitoring devices, contact with human tissue will not cause adverse reactions, and it can safely and stably monitor human physiological signals, such as heart rate, blood pressure, and slight pressure changes caused by muscle activity, to provide accurate data support for medical diagnosis and health management. The thickness range is set at 14-100 microns. The thinner lower limit (14 microns) makes the film more likely to deform when subjected to external forces, and can respond more sensitively to weak external stimuli, which helps to improve the overall sensitivity of the sensor and accurately capture subtle changes in physical quantities. The relatively thick upper limit (100 microns) ensures that the film has a certain structural strength. After multiple deformations and external forces of varying degrees, it can still maintain its integrity and stable performance. At the same time, it will not be too fragile due to being too thin, making it difficult to process and install, thus achieving a good balance between sensitivity and mechanical flexibility, so that it can better adapt to the application needs of wearable devices under frequent human activities and flexible electronic devices in various bending and stretching scenarios.

[0022] Optionally, the hot melt adhesive film is ethylene vinyl acetate (EVA) hot melt adhesive film, polyamide (PA) hot melt adhesive film or TPU hot melt adhesive film, with a thickness ranging from 10 to 50 microns; the lower flexible connecting layer is made of the same material as the upper flexible connecting layer.

[0023] Optionally, the hollow area design of the partially hollowed-out hot-melt adhesive film includes a rectangle, an ellipse, an array rectangle and corresponding evolved patterns.

[0024] In the second aspect, the present application provides a method for preparing a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, which adopts the following technical solution:

[0025] A method for preparing a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, the specific steps are as follows:

[0026] Step 1: Use UV curing technology to fix the PVDF film on the metal sheet, use traditional mechanical cutting or precision laser cutting technology to process it into a specified paper-cut pattern, and wash off the UV glue to obtain the piezoelectric film layer;

[0027] Step 1-1: UV curing of PVDF film: Apply hydrolyzed UV glue on one side of the PVDF film with a brush and bond it to the surface of a 1mm thick smooth metal sheet, and cure it under UV light;

[0028] Step 1-2: Cut into specified paper-cut patterns:

[0029] Traditional mechanical cutting method: clamp the metal sheet of the cured PVDF film obtained in step 1-1 on a die-cutting machine to process the paper-cut pattern;

[0030] Precision laser cutting method: clamp the metal sheet of the cured PVDF film obtained in step 1-1 on a laser cutting machine to process the paper-cut pattern;

[0031] Step 1-3: Degumming PVDF film: spray alcohol gel on the surface of the PVDF film cut in step 1-2, completely immerse it to achieve degumming, and obtain a PVDF film cut like paper-cutting;

[0032] Step 2: Apply conductive silver paste on both sides of the piezoelectric film layer to form an electrode layer and connect the wires;

[0033] Step 2-1: Prepare electrode layer: First, fix the piezoelectric film prepared in step 1 on the surface of the substrate, and evenly apply the stretchable conductive silver paste by scraping with a wet film preparation device; dry in a ventilated environment within the range of 40-80 degrees Celsius; and complete the preparation of electrode layers on both sides by the same operation;

[0034] Step 2-2: Preliminary fixing of the wire: Use conductive silver paste to connect one end of the enameled wire to the designated area on the surface of the piezoelectric film prepared in step 2-1, and fix it in a ventilated environment at 40-80 degrees Celsius;

[0035] Step 3: Use thermal lamination technology to press the multi-layer materials together, and form a large deformation flexible piezoelectric sensor that can be conformally attached after local ultrasonic welding:

[0036] Step 3-1: Preparation of protective layer: Using the above-mentioned cutting method, process 10-300 micron high temperature resistant polymer film. When using mechanical processing technology, the blade pressure is changed to 20-100N; when using laser cutting method, the laser power is changed to 0.5-3W;

[0037] Step 3-2: Preparation of the connection layer: Use the above-mentioned cutting technology to process a 10-50 micron hot melt adhesive film. When using mechanical processing technology, the blade pressure is set to 10-100N; when using laser cutting technology, the laser power is set to 0.5-1W;

[0038] Step 3-3: Constant temperature pressing: stack the polymer protective layer, flexible connection layer, and piezoelectric film prepared in step 2, step 3-1, and step 3-2 in order in an aluminum fixture, and place the fixture as a whole in a hot press; set the hot press temperature to 80-100 degrees Celsius, the pressure to 0.6-2.0 MPa, and the pressing time to 2-4 hours;

[0039] Step 3-4: Local ultrasonic welding: Use an ultrasonic welding gun, select a circular flat surface on the welding head, and perform spot welding for 0.5 seconds at the required position.

[0040] By adopting the above technical solution, the PVDF film is first fixed on the metal sheet through UV curing technology, which provides a stable operating basis for subsequent cutting, ensures that the film will not be easily displaced during the processing, and helps to improve the processing accuracy.

[0041] Traditional mechanical cutting and precision laser cutting are provided to process specified paper-cut patterns. Mechanical cutting is suitable for some relatively regular and simple patterns, with low cost and high efficiency; laser cutting has excellent processing capabilities for complex and fine patterns, and can accurately produce straight line cut marks, fractal cut marks and corresponding evolutionary patterns, etc., to meet different design requirements, ensure that the piezoelectric film layer finally has an ideal structural form, and help improve the sensitivity and flexibility of the sensor.

[0042] The degumming method of spraying alcohol gel can quickly and effectively separate the PVDF film from the metal sheet. In addition, the alcohol gel is volatile and no residual impurities will affect the performance of the film, making the entire preparation process coherent and efficient, which helps to improve production efficiency and ensure product quality.

[0043] The electrode layer is prepared by applying the stretchable conductive silver paste by the doctor blade method using a wet film preparation device. The doctor blade method can make the conductive silver paste evenly distributed on both sides of the piezoelectric film layer to form an electrode layer with uniform thickness and good conductivity. The uniform electrode layer can ensure that the electrical signal generated in the piezoelectric film layer can be fully and stably collected and transmitted, avoiding problems such as unstable signal transmission and signal loss caused by local defects or uneven thickness of the electrode layer, thereby improving the accuracy and reliability of sensor signal acquisition.

[0044] One end of the enameled wire is fixed to a designated area on the surface of the piezoelectric film by conductive silver paste in a ventilated environment at 40-80 degrees Celsius. This method utilizes the property of conductive silver paste that solidifies at an appropriate temperature and forms a good conductive path. The ventilated environment helps the solvent to evaporate, making the connection between the wire and the electrode layer firm and the conductive performance reliable. This minimizes signal interference and interruption caused by poor connection, ensuring that the sensor can smoothly output the collected signal.

[0045] During the preparation of the protective layer and the connecting layer, the parameter ranges such as blade pressure and laser power are clearly defined for the mechanical processing technology and laser cutting technology respectively. This allows the accurate processing of high-temperature resistant polymer films and hot-melt adhesive films that meet the thickness requirements according to actual conditions, ensuring that these materials can better play their respective protective and connection functions in the subsequent hot lamination process. At the same time, it ensures processing accuracy, reduces material waste, and improves the controllability and standardization of production.

[0046] Through constant temperature lamination, the temperature, pressure and lamination time of the hot press are precisely set, so that the polymer protective layer, the flexible connection layer and the piezoelectric film layer can be tightly and fully bonded. The appropriate temperature causes the hot melt adhesive film to soften and play a bonding role, and the appropriate pressure and time ensure that the layers are fully in contact and fused, avoiding problems such as bubbles between layers and loose fitting, improving the stability and durability of the overall structure of the sensor, and enabling it to maintain good performance under different use environments and stress conditions.

[0047] Local ultrasonic welding is adopted, and a welding head of a specific shape is selected to perform short-time spot welding operations at the required positions. This precise welding method effectively strengthens the connection of key parts without affecting the overall flexibility and structural integrity of the sensor. For example, spot welding is performed at important parts such as the connection between the wire and the electrode layer, which can not only enhance the firmness of the connection, but also prevent the sensor from becoming hard and brittle due to large-area welding, thereby losing flexibility and conformal adhesion ability. This further improves the manufacturing quality and performance stability of the sensor, so that it can still work normally and accurately transmit signals under large deformation conditions.

[0048] Optionally, the method further comprises:

[0049] According to the thickness of the PVDF film, the temperature, pressure and pressing time of hot pressing are adjusted in a positive correlation; the thicker the PVDF film, the higher the temperature of hot pressing, the greater the pressure of hot pressing, and the longer the pressing time of hot pressing; the thinner the PVDF film, the lower the temperature of hot pressing, the smaller the pressure of hot pressing, and the shorter the pressing time of hot pressing.

[0050] Optionally, the method further comprises:

[0051] According to the thickness of the piezoelectric film layer, the temperature, pressure and pressing time of the hot pressing are adjusted in a positive correlation; the thicker the piezoelectric film layer, the higher the temperature of the hot pressing, the greater the pressure of the hot pressing, and the longer the pressing time of the hot pressing; the thinner the piezoelectric film layer, the lower the temperature of the hot pressing, the smaller the pressure of the hot pressing, and the shorter the pressing time of the hot pressing.

[0052] In a third aspect, the present application provides a signal analysis method for a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, which adopts the following technical solution:

[0053] A signal analysis method for a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, the specific steps are as follows:

[0054] Step 1: Data preprocessing. First, collect multi-channel piezoelectric signal data of flexible thin film piezoelectric sensors, and ensure that the sampling frequency is high enough to capture signal details. De-noise the collected signals, and use statistical methods to detect and remove outliers. Split the continuous signal data into fixed-length time series samples, each containing the same number of time steps. Label the samples according to task requirements.

[0055] Step 2: Model construction. Choose a suitable LSTM model architecture, such as single-layer LSTM, multi-layer LSTM, or bidirectional LSTM. The specific choice depends on the complexity of the task and the characteristics of the data. Determine the key parameters of the LSTM model, such as the input dimension (i.e., time step and number of features), hidden layer dimension, and output dimension. Use a deep learning framework (such as Matlab, TensorFlow, or PyTorch) to build the LSTM model and define the loss function and optimizer. For feature recognition tasks, you can choose the binary cross entropy loss function; for nonlinear fitting tasks, you can choose the mean square error loss function.

[0056] Step 3: Model training. Divide the preprocessed dataset into training set, validation set, and test set, usually in a ratio of 70%, 15%, and 15%. Use the training set to train the LSTM model, and use the validation set to evaluate the model performance after each training cycle, adjusting hyperparameters such as learning rate and batch size to optimize model performance. Monitor the loss function and evaluation indicators during the training process to ensure that the model can achieve good performance on both the training set and the validation set.

[0057] Step 4: Model application. Apply the trained LSTM model to the actual multi-channel piezoelectric signal analysis task. For feature recognition tasks, input new signal data, and the model will output the predicted result of the feature signal; for nonlinear fitting tasks, the model will output the predicted value of the signal. Analyze the prediction results of the model, such as counting the frequency of occurrence of feature signals or evaluating the accuracy of nonlinear fitting. Based on the analysis results, further optimize the model or adjust the application strategy.

[0058] In summary, the present application includes at least one of the following beneficial technical effects:

[0059] The present invention improves the sensitivity and tensile properties of the PVDF piezoelectric film through paper-cut cutting, and can perform customized pattern design as needed to achieve conformal attachment to complex surfaces such as human wear and automobile structure monitoring.

[0060] The present invention realizes multi-dimensional force characterization of piezoelectric sensors according to multi-channel output signals by designing regionalized electrode coating, thereby reducing the design and manufacturing costs of multi-dimensional force sensors and promoting the application of piezoelectric sensors in the fields of robot dexterous hands, automobile smart cockpits, etc.

[0061] The piezoelectric sensor prepared by the present invention is obtained by hot pressing of multiple layers of materials. The thickness of each layer can be selected according to needs. The overall thickness can be as low as 0.2 mm, which can adapt to narrow spaces and meet lightweight requirements.

[0062] The paper cutting process proposed in the present invention is simple and can be quickly produced by conventional methods such as laser cutting and mechanical cutting, which greatly reduces the manufacturing cost; and the electrode is prepared by scraping technology, which does not require a complicated manufacturing process; hot pressing combined with ultrasonic welding greatly prolongs the service life of the device. As an industrialized, low-cost flexible strain sensor, it has great application prospects.

[0063] The present invention combines deep learning signal analysis methods to realize automatic feature extraction of complex sensor signals according to the nonlinear waveform characteristics of the sensor, improves the signal detection capability in weak signal and high noise environments, supports real-time processing and classification of dynamic data, and is suitable for multi-scenario application needs, such as complex functions such as human behavior recognition, structural health monitoring, and smart device interaction, to achieve efficient and accurate signal perception and interpretation. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic structural diagram of an embodiment of the present invention.

[0065] Figure 2 This is a planar pattern of each layer structure of an embodiment of the present invention.

[0066] Figure 3 This is a physical diagram of an embodiment of the present invention.

[0067] Figure 4 This is a diagram illustrating a high temperature resistance test of an embodiment of the present invention.

[0068] Figure 5 This is a diagram illustrating a bump resistance test of an embodiment of the present invention.

[0069] Figure 6 This is a diagram illustrating a heavy object impact resistance test of an embodiment of the present invention.

[0070] Figure 7 This is a diagram illustrating the test of an embodiment of the present invention under a tensile environment.

[0071] Figure 8 This is a diagram illustrating the friction resistance of an embodiment of the present invention.

[0072] Fig. 9 This is a diagram illustrating a water resistance test of an embodiment of the present invention.

[0073] Fig.10 This is an illustration of a milk corrosion resistance test of an embodiment of the present invention.

[0074] Fig.11 It is an illustration of the bending resistance test of an embodiment of the present invention.

[0075] Fig.12 This is a pasted image used in facial action recognition according to an embodiment of the present invention.

[0076] Fig.13 This is a signal display diagram of an embodiment of the present invention applied in facial action recognition.

[0077] Figure numerals: 1. upper flexible polymer protective layer; 2. upper flexible connecting layer; 3. upper electrode layer; 4. piezoelectric film layer; 5. lower electrode layer; 6. lower flexible connecting layer; 7. lower flexible polymer protective layer. DETAILED DESCRIPTION

[0078] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings.

[0079] In the description of this specification, the description with reference to the terms "certain embodiments", "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0080] The embodiment of the present application discloses a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization.

[0081] Example 1

[0082] Reference Figure 1 , Figure 2 and Figure 3 , a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization includes:

[0083] From top to bottom, it consists of an upper flexible polymer protective layer 1, an upper flexible connection layer 2, an upper electrode layer 3, a piezoelectric film layer 4, a lower electrode layer 5, a lower flexible connection layer 6 and a lower flexible polymer protective layer 7.

[0084] The thickness of the upper flexible polymer protective layer 1 and the lower flexible polymer protective layer 7 is 0.09 mm, and the material is thermoplastic polyurethane (TPU).

[0085] The thickness of the upper flexible connection layer 2 and the lower flexible connection layer 6 is 0.02 mm, and the material is low-melting-point thermoplastic polyurethane (TPU).

[0086] The piezoelectric film layer 4 is made of PVDF material and has a thickness of 0.027 mm.

[0087] In this example, the paper-cutting pattern is cut by conventional mechanical cutting. The cut marks are straight line cut marks. The upper electrode layer 3 and the lower electrode layer 5 are coated with cured conductive silver paste, and the thickness after complete curing is 0.005 mm.

[0088] In this example, the kirigami scheme is used to improve the stretchability and sensitivity of the sensor.

[0089] This example mainly verifies the applicability of the sensor in multiple working conditions, including high temperature environment, bumpy environment, heavy object impact, stretching scenario, friction durability, water resistance, corrosion resistance, and bending resistance.

[0090] Reference Figure 4 , sensor sample S1 was placed in an environment of 85℃ for 96 hours, and then placed in air for 12 hours. The sensor appearance was intact, without separation, damage or wire breakage. The sensor still maintained its sensing capability. The signal root mean square error was -0.45%.

[0091] Reference Figure 5 After the sensor sample S2 experienced 160,000 vibration cycles on the exciter, the sensor appearance was intact, without separation, damage, or wire breakage. The sensor still maintained its sensing capability. The signal root mean square error was -4.78%.

[0092] Reference Figure 6, sensor sample S3 uses a 6kg barbell as the impact weight and impacts the simulated seat sponge pad 10 times. The sensor is pasted and fixed. The vertical height is 1.1m. The sensor appearance is intact, without separation, damage or wire breakage. The sensor still maintains its sensing ability. The signal root mean square error is -3.87%.

[0093] Reference Figure 7 The sensor sample S4 was stretched in a static machine at a speed of 50 mm / min, and the measured ultimate destructive force was 40 N and the ultimate strain was 50%.

[0094] Reference Figure 8 , sensor sample S5 was rubbed 102,000 times at a frequency of 2.1Hz in a self-made rotating friction test bench. The sensor's appearance was intact, with no separation, damage, or wire breakage. The sensor still maintained its sensing capability. The signal root mean square error was 4.01%.

[0095] Reference Fig. 9 , sensor sample S6 was immersed in clean water. After standing and drying, the sensor looked intact, without separation, damage or wire breakage. The sensor still maintained its sensing capability. The signal root mean square error was 0.45%.

[0096] Reference Fig.10 , sensor sample S7 was immersed in milk. After standing and drying, the sensor looked intact, without separation, damage or wire breakage. The sensor still maintained its sensing capability. The signal RMS error was 0.046%.

[0097] Reference Fig.11 , sensor sample S8 was bent at a frequency of 1.6Hz for 50,000 bending cycles. The sensor was intact without separation, damage or wire breakage. The sensor still maintained its sensing capability. The signal RMS error was 3.19%.

[0098] Example 2

[0099] A flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization comprises:

[0100] From top to bottom, it consists of an upper flexible polymer protective layer 1, an upper flexible connection layer 2, an upper electrode layer 3, a piezoelectric film layer 4, a lower electrode layer 5, a lower flexible connection layer 6 and a lower flexible polymer protective layer 7.

[0101] The thickness of the upper flexible polymer protective layer 1 and the lower flexible polymer protective layer 7 is 0.01 mm, and the material is polyetheramide elastomer (PEBA).

[0102] The thickness of the upper flexible connection layer 2 and the lower flexible connection layer 6 is 0.01 mm, and the material is ethylene vinyl acetate (EVA).

[0103] The piezoelectric film layer 4 is made of PVDF material and has a thickness of 0.014 mm.

[0104] In this example, the paper-cutting pattern is cut by conventional mechanical cutting. The cut marks are straight line cut marks. The upper electrode layer 3 and the lower electrode layer 5 are coated with cured conductive silver paste, and the thickness after complete curing is 0.005 mm.

[0105] Example 3

[0106] A flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization comprises:

[0107] From top to bottom, it consists of an upper flexible polymer protective layer 1, an upper flexible connection layer 2, an upper electrode layer 3, a piezoelectric film layer 4, a lower electrode layer 5, a lower flexible connection layer 6 and a lower flexible polymer protective layer 7.

[0108] The thickness of the upper flexible polymer protective layer 1 and the lower flexible polymer protective layer 7 is 0.3 mm, and the material is polyurethane elastomer (PU).

[0109] The thickness of the upper flexible connection layer 2 and the lower flexible connection layer 6 is 0.05 mm, and the material is polyamide (PA).

[0110] The piezoelectric film layer 4 is made of PVDF material and has a thickness of 0.1 mm.

[0111] In this example, the paper-cutting pattern is cut by traditional mechanical cutting. The cut marks are fractal cut marks. The upper electrode layer 3 and the lower electrode layer 5 are coated with cured conductive silver glue, and the thickness after complete curing is 0.005mm. The electrode area adopts a three-channel design.

[0112] In this example, the prepared wide-operating-condition, multi-dimensional force characterization flexible thin-film piezoelectric sensor is used for silent facial recognition. Fig.12 The measured signal is shown as Fig.13 shown.

[0113] In this example, a two-layer LSTM algorithm is constructed for three-channel signal analysis, as follows:

[0114] The three-channel piezoelectric signal data of the flexible thin film piezoelectric sensor is collected by a data collector at a sampling rate of 1000Hz. The collected signal is low-pass filtered to remove the 50Hz power frequency interference. The three-channel continuous signal data is divided into time series samples with a fixed length of 300, and each sample contains the same number of time steps. The samples are labeled with initial consonants (sh, zh, ch, b, p).

[0115] Select the two-layer LSTM model. Determine the input dimension of the LSTM model (i.e., the time step is 0.3s and the number of features is 5), the hidden layer dimension is 3, and the output dimension is 5. Use Matlab's built-in functions to build the LSTM model and define the loss function and optimizer.

[0116] The preprocessed data set is divided into training set, validation set and test set with a ratio of 70%, 15% and 15%. The LSTM model is trained using the training set until the training success rate reaches 99.9%.

[0117] The trained two-layer LSTM model was applied to the actual silent facial recognition application, and the recognition success rate for 5 types of initial consonants was 93.17%.

[0118] The present application also discloses a method for preparing a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, and the specific steps are as follows:

[0119] Step 1: Apply hydrolyzed UV glue on one side of the PVDF film with a brush and bond it to the surface of a 1mm thick smooth aluminum plate. Use a scraper to remove excess UV glue and irradiate it under a 395nm wavelength UV lamp for 1 minute. Clamp the obtained metal sheet of the cured PVDF film on the die-cutting machine, use a 60-degree blade, set the knife pressure to 10N, and the cutting speed to 10mm / min to process the paper-cut pattern in CAD format. Spray the surface of the cut PVDF film with 75% alcohol gel, immerse it completely for 10 minutes to achieve debonding, use a fine steel wire brush to gently sweep the surface of the PVDF film to remove the residual UV glue residue, and obtain a PVDF film cut like a paper-cut.

[0120] Step 2: First, fix the piezoelectric film obtained in step 1 on the surface of the substrate, and use a wet film preparation device with a coating width of 100 mm and a coating height of 10 μm to evenly apply the stretchable conductive silver paste by scraping. Dry for 2 hours in a ventilated environment at 60 degrees Celsius. Complete the preparation of the electrode layer on both sides in the same way. Use conductive silver paste to connect one end of the 0.1 mm enameled wire to the specified area on the surface of the piezoelectric film, and dry for 1 hour in a ventilated environment at 60 degrees Celsius to complete the fixation.

[0121] Step 3: Use traditional mechanical cutting technology to process 0.09mm TPU film as a protective layer, with a knife pressure of 40N. Cut and process 0.02mm TPU hot melt adhesive film as a connecting layer, with a knife pressure of 15N. Stack the prepared polymer protective layer, flexible connecting layer, and piezoelectric film in order in an aluminum fixture, and place the fixture as a whole in a hot press. Set the hot press temperature to 90 degrees Celsius, the pressure to 0.6MPa, and the pressing time to 3 hours. Use an ultrasonic welding gun with a power of 1800W and a frequency of 15KHz, and select a circular plane with a diameter of 2mm on the surface of the welding head, and perform spot welding for 0.5s at the required position. After completion, the piezoelectric sensor with excellent mechanical properties, high sensitivity, and large deformation and stretching can be prepared by encapsulation.

[0122] The prepared piezoelectric sensor will generate electric charge when subjected to mechanical stress, thereby realizing deformation and pressure perception.

[0123] In addition, for layer structures of different thicknesses, the method also includes:

[0124] According to the thickness of the PVDF film, the temperature, pressure and pressing time of the hot pressing are adjusted in a positive correlation; the thicker the PVDF film, the higher the temperature of the hot pressing, the greater the pressure of the hot pressing, and the longer the pressing time of the hot pressing; the thinner the PVDF film, the lower the temperature of the hot pressing, the smaller the pressure of the hot pressing, and the shorter the pressing time of the hot pressing. When the PVDF film is thicker (such as close to 100 microns), higher temperatures and longer pressing times are required to ensure complete melting of the film and tight adhesion between the layers. At the same time, in order to overcome the resistance caused by the thicker film, the pressure needs to be increased. When the PVDF film is thinner (such as close to 14 microns), the temperature and pressure need to be lowered, and the pressing time needs to be shortened to avoid damage to the film due to overheating or excessive pressure.

[0125] According to the thickness of the piezoelectric film layer 4, the temperature, pressure and pressing time of the hot pressing are adjusted in a positive correlation; the thicker the piezoelectric film layer 4, the higher the temperature of the hot pressing, the greater the pressure of the hot pressing, and the longer the pressing time of the hot pressing; the thinner the piezoelectric film layer 4, the lower the temperature of the hot pressing, the smaller the pressure of the hot pressing, and the shorter the pressing time of the hot pressing.

[0126] This embodiment also discloses a signal analysis method for a flexible thin film piezoelectric sensor with wide operating conditions and multi-dimensional force characterization, and a deep learning method for analyzing the piezoelectric signal of the sensor.

[0127] For example, by using an end-to-end neural network model, key features can be directly extracted from the original signal for tasks such as action recognition, pressure classification, and anomaly detection. Through multimodal models (such as Transformer architecture or multi-branch networks), joint data analysis and information fusion can be achieved to improve the overall signal interpretation ability. By using transfer learning and few-shot learning, the dependence on large-scale labeled data can be reduced, and the model can be quickly generalized and applied across scenarios.

[0128] A signal analysis method for a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, the specific steps are as follows:

[0129] Step 1: Data preprocessing. First, collect multi-channel piezoelectric signal data of flexible thin film piezoelectric sensors, and ensure that the sampling frequency is high enough to capture signal details. De-noise the collected signals, and use statistical methods to detect and remove outliers. Split the continuous signal data into fixed-length time series samples, each containing the same number of time steps. Label the samples according to task requirements.

[0130] Step 2: Model construction. Choose a suitable LSTM model architecture, such as single-layer LSTM, multi-layer LSTM, or bidirectional LSTM. The specific choice depends on the complexity of the task and the characteristics of the data. Determine the key parameters of the LSTM model, such as the input dimension (i.e., time step and number of features), hidden layer dimension, and output dimension. Use a deep learning framework (such as Matlab, TensorFlow, or PyTorch) to build the LSTM model and define the loss function and optimizer. For feature recognition tasks, you can choose the binary cross entropy loss function; for nonlinear fitting tasks, you can choose the mean square error loss function.

[0131] Step 3: Model training. Divide the preprocessed dataset into training set, validation set, and test set, usually in a ratio of 70%, 15%, and 15%. Use the training set to train the LSTM model, and use the validation set to evaluate the model performance after each training cycle, adjusting hyperparameters such as learning rate and batch size to optimize model performance. Monitor the loss function and evaluation indicators during the training process to ensure that the model can achieve good performance on both the training set and the validation set.

[0132] Step 4: Model application. Apply the trained LSTM model to the actual multi-channel piezoelectric signal analysis task. For feature recognition tasks, input new signal data, and the model will output the predicted result of the feature signal; for nonlinear fitting tasks, the model will output the predicted value of the signal. Analyze the prediction results of the model, such as counting the frequency of occurrence of feature signals or evaluating the accuracy of nonlinear fitting. Based on the analysis results, further optimize the model or adjust the application strategy.

[0133] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, a variety of simple modifications can be made to the technical solution of the present invention, and these simple modifications all belong to the protection scope of the present invention.

[0134] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.

[0135] In addition, various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, characterized in that: It comprises an upper flexible polymer protective layer (1), an upper flexible connection layer (2), an upper electrode layer (3), a piezoelectric film layer (4), a lower electrode layer (5), a lower flexible connection layer (6), a lower flexible polymer protective layer (7), and an enameled wire for outputting signals. The upper flexible polymer protective layer (1) and the lower flexible polymer protective layer (7) are both made of high temperature resistant polymer films; The piezoelectric film layer (4) is a film layer cut in a paper-cutting manner, wherein the paper-cutting pattern includes a straight line cutting mark, a fractal cutting mark and a corresponding evolution pattern, forming a paper-cutting area; The upper flexible connection layer (2) and the lower flexible connection layer (6) are hot-melt adhesive films, and are partially hollowed out corresponding to the paper-cut areas on the piezoelectric film layer (4).

2. The conformally attachable large deformation flexible piezoelectric sensor according to claim 1, characterized in that: The high temperature resistant polymer film comprises a thermoplastic polyurethane (TPU) film, a polyetheramide elastomer (PEBA) film or a polyurethane elastomer (PU) film; the upper flexible polymer protective layer (1) and the lower flexible polymer protective layer (7) are made of the same or different materials; the thickness of the upper flexible polymer protective layer (1) and the lower flexible polymer protective layer (7) are both in the range of 10-300 microns.

3. The conformally attachable large deformation flexible piezoelectric sensor according to claim 1, characterized in that: The piezoelectric film layer (4) is made of polarized PVDF film with a thickness ranging from 14 to 100 micrometers.

4. The conformally attachable large deformation flexible piezoelectric sensor according to claim 1, characterized in that: The hot melt adhesive film is an ethylene vinyl acetate (EVA) hot melt adhesive film, a polyamide (PA) hot melt adhesive film or a TPU hot melt adhesive film, and has a thickness ranging from 10 to 50 microns; the lower flexible connection layer (6) is made of the same material as the upper flexible connection layer (2).

5. The conformally attachable large deformation flexible piezoelectric sensor according to claim 1, characterized in that: The upper electrode layer (3) and the lower electrode layer (5) are conductive silver paste or PEDOT:PSS applied in the shear separation area of ​​the piezoelectric film layer (4); the lower electrode layer (5) and the upper electrode layer (3) have the same application area; the upper electrode layer (3) and the lower electrode layer (5) are not connected.

6. The conformally attachable large deformation flexible piezoelectric sensor according to claim 1, characterized in that: Micro airbags are arranged in the hot melt adhesive film, and the micro airbags are spherical or elliptical.

7. A method for preparing a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, characterized in that: The specific steps are as follows: Step 1: fix the PVDF film on the metal sheet using UV curing technology, process it into a specified paper-cut pattern using traditional mechanical cutting or precision laser cutting technology, and wash off the UV glue to obtain a piezoelectric film layer (4); Step 1-1: UV curing of PVDF film: Apply hydrolyzed UV glue on one side of the PVDF film with a brush and bond it to the surface of a 1mm thick smooth metal sheet, and cure it under UV light; Step 1-2: Cut into specified paper-cut patterns: Traditional mechanical cutting method: clamp the metal sheet of the cured PVDF film obtained in step 1-1 on a die-cutting machine to process the paper-cut pattern; Precision laser cutting method: clamp the metal sheet of the cured PVDF film obtained in step 1-1 on a laser cutting machine to process the paper-cut pattern; Step 1-3: Degumming PVDF film: spray alcohol gel on the surface of the PVDF film cut in step 1-2, completely immerse it to achieve degumming, and obtain a PVDF film cut like paper-cutting; Step 2: Apply conductive silver paste on both sides of the piezoelectric film layer (4) to form an electrode layer and connect wires; Step 2-1: Prepare electrode layer: First, fix the piezoelectric film prepared in step 1 on the surface of the substrate, and evenly apply the stretchable conductive silver paste by scraping with a wet film preparation device; dry in a ventilated environment within the range of 40-80 degrees Celsius; and complete the preparation of electrode layers on both sides by the same operation; Step 2-2: Preliminary fixing of the wire: Use conductive silver paste to connect one end of the enameled wire to the designated area on the surface of the piezoelectric film prepared in step 2-1, and fix it in a ventilated environment at 40-80 degrees Celsius; Step 3: Use thermal lamination technology to press the multi-layer materials together, and then form a large deformation flexible piezoelectric sensor that can be conformally attached after local ultrasonic welding: Step 3-1: Preparation of protective layer: Using the above-mentioned cutting method, process 10-300 micron high temperature resistant polymer film. When using mechanical processing technology, the blade pressure is changed to 20-100N; when using laser cutting method, the laser power is changed to 0.5-3W; Step 3-2: Preparation of the connection layer: Use the above-mentioned cutting technology to process a 10-50 micron hot melt adhesive film. When using mechanical processing technology, the blade pressure is set to 10-100N; when using laser cutting technology, the laser power is set to 0.5-1W; Step 3-3: Constant temperature pressing: stack the polymer protective layer, flexible connection layer, and piezoelectric film prepared in step 2, step 3-1, and step 3-2 in order in an aluminum fixture, and place the fixture as a whole in a hot press; set the hot press temperature to 80-100 degrees Celsius, the pressure to 0.6-2.0 MPa, and the pressing time to 2-4 hours; Step 3-4: Local ultrasonic welding: Use an ultrasonic welding gun, select a circular flat surface on the welding head, and perform spot welding for 0.5 seconds at the required position.

8. The method for preparing a conformally attachable large deformation flexible piezoelectric sensor according to claim 7, characterized in that: The method also includes: According to the thickness of the PVDF film, the temperature, pressure and pressing time of hot pressing are adjusted in a positive correlation; the thicker the PVDF film, the higher the temperature of hot pressing, the greater the pressure of hot pressing, and the longer the pressing time of hot pressing; the thinner the PVDF film, the lower the temperature of hot pressing, the smaller the pressure of hot pressing, and the shorter the pressing time of hot pressing.

9. The method for preparing a conformally attachable large deformation flexible piezoelectric sensor according to claim 7, characterized in that: The method also includes: According to the thickness of the piezoelectric film layer (4), the temperature, pressure and pressing time of the hot pressing are adjusted in a positive correlation; the thicker the piezoelectric film layer (4), the higher the temperature of the hot pressing, the greater the pressure of the hot pressing, and the longer the pressing time of the hot pressing; the thinner the piezoelectric film layer (4), the lower the temperature of the hot pressing, the smaller the pressure of the hot pressing, and the shorter the pressing time of the hot pressing.

10. A signal analysis method for a flexible thin film piezoelectric sensor with wide working conditions and multi-dimensional force characterization, the specific steps are as follows: Step 1: Data preprocessing: First, collect multi-channel piezoelectric signal data of flexible thin film piezoelectric sensors, and ensure that the sampling frequency is high enough to capture signal details; denoise the collected signals, and use statistical methods to detect and remove outliers; split the continuous signal data into fixed-length time series samples, each containing the same number of time steps; and label the samples according to task requirements; Step 2: Model construction: Select a suitable LSTM model architecture, such as single-layer LSTM, multi-layer LSTM or bidirectional LSTM. The specific choice depends on the complexity of the task and the characteristics of the data; determine the key parameters of the LSTM model, such as the input dimension, hidden layer dimension, and output dimension. The input dimension of the LSTM model is the time step and the number of features; use the deep learning framework to build the LSTM model and define the loss function and optimizer; for feature recognition tasks, select the binary cross entropy loss function; for nonlinear fitting tasks, select the mean square error loss function; Step 3: Model training: Divide the preprocessed dataset into training set, validation set, and test set with a ratio of 70%, 15%, or 15%; use the training set to train the LSTM model, and use the validation set to evaluate the model performance after each training cycle, and adjust the learning rate and batch size to optimize the model performance; Monitor the loss function and evaluation metrics during training to ensure that the model achieves good performance on both the training set and the validation set; Step 4: Model application: Apply the trained LSTM model to the actual multi-channel piezoelectric signal analysis task; for the feature recognition task, input new signal data, and the model will output the prediction result of the feature signal; for the nonlinear fitting task, the model will output the predicted value of the signal; analyze the prediction results of the model, such as counting the frequency of occurrence of feature signals, or evaluating the accuracy of nonlinear fitting; Based on the analysis results, optimize the model or adjust the application strategy.

Images