Chip design method and system of silicon piezoresistive pressure sensor

Abstract

The application provides a chip design method and system of a silicon piezoresistive pressure sensor, comprising: constructing a design parameter set of the silicon piezoresistive pressure sensor; obtaining a performance parameter set corresponding to the design parameter set according to a preset simulation software; constructing a numerical design library according to the design parameter set and the performance parameter set; and establishing a bidirectional numerical design model based on a Bi-GRU neural network according to the numerical design library. The method or system can inversely realize active design of structure parameters and piezoresistance through target performance parameters, and can forwardly obtain performance parameters of a pressure chip through design parameters, so that a result can be given without multiple simulation calculations, and the design efficiency is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of pressure sensors, in particular to a chip design method and system of a silicon piezoresistive pressure sensor. BACKGROUND

[0002] The silicon piezoresistive pressure sensor is a sensor that measures pressure by using the piezoresistive effect of single crystal silicon material. Its working principle is that when single crystal silicon material is subjected to external pressure, its internal lattice structure will deform, causing the resistance value to change. By measuring the change in resistance, the size of the pressure acting on the silicon material can be calculated. The silicon piezoresistive pressure sensor has been widely used in industrial oil pressure measurement, industrial transportation pipeline, aerospace ignition and extinction, etc. due to its simple structure, low cost and strong batch manufacturing capability.

[0003] However, due to the physical properties of silicon-based materials, before the pressure sensor chip is prepared, the chip size needs to be selected according to the working environment of the chip and the performance requirements of the chip, and the chip structure needs to be simulated multiple times to meet the working needs. When the working environment and performance requirements of the chip change, the chip structure needs to be simulated and redesigned. For P-type doped pressure sensitive resistors, the TCR is positive, that is, as the temperature rises, the pressure sensitive resistor value increases. Under different doping concentrations, the piezoresistive coefficient decreases with the increase of temperature, but the decrease amplitude is different, so different doping concentrations need to be selected to reduce the influence of temperature on the performance of the chip. This process is time-consuming and laborious. SUMMARY

[0004] The chip design method and system of the silicon piezoresistive pressure sensor provided by the embodiments of the present application can overcome the technical problems existing in the prior art, and can realize the active design of the structure parameters and the piezoresistance through the target performance parameters, and obtain the performance parameters of the pressure chip through the design parameters. The result can be obtained without multiple simulation calculations, which greatly improves the design efficiency.

[0005] To achieve the above-mentioned purpose, the present application provides a chip design method of a silicon piezoresistive pressure sensor, characterized in that it comprises:

[0006] Constructing a design parameter set of a silicon piezoresistive pressure sensor;

[0007] Obtaining a performance parameter set corresponding to the design parameter set according to a preset simulation software;

[0008] Constructing a numerical design library according to the design parameter set and the performance parameter set;

[0009] Establishing a bidirectional numerical design model based on Bi-GRU neural network according to the numerical design library.

[0010] The application also provides a chip design system of a silicon piezoresistive pressure sensor, comprising a database module, a design parameter module, a performance parameter module and a model evaluation module.

[0011] The database module comprises a design parameter set and a performance parameter set; the design parameter set comprises a chip structure form, a chip size, a diaphragm size, a piezoresistive doping concentration, a heavy doping concentration, a chip working temperature and a pressure range; and the performance parameter set comprises a chip output voltage, a sensitivity, a linearity, a natural frequency, a maximum equivalent stress, a thermal sensitivity drift and a thermal zero-point drift.

[0012] The design parameter module is configured to construct a forward numerical design model based on a Bi-GRU neural network according to the database module, and obtain a chip performance parameter result prediction according to a design parameter.

[0013] The performance parameter module is configured to construct a reverse numerical design model based on a Bi-GRU neural network according to the database module, and obtain a chip design parameter result prediction according to a performance parameter.

[0014] The model evaluation module is configured to evaluate the model performance and adjust the parameters or optimize the model structure as needed.

[0015] The chip design method and system of the silicon piezoresistive pressure sensor provided by the application obtain corresponding performance parameters by using simulation software and numerical calculation according to design parameters, construct a numerical design library, and then use a Bi-GRU neural network to mine forward and reverse information of the design parameters and the performance parameters; wherein the first layer of GRU obtains forward information of the input sequence, and the second layer of GRU further extracts reverse features of the input sequence on the basis of the first layer, which can improve the expression ability of the model. This method realizes bidirectional design of the pressure chip design parameters and the performance parameters, that is, the active design of the structure parameters and the piezoresistance is realized in reverse through the target performance parameters, and the performance parameter indicators of the pressure chip can also be obtained in forward through the design parameters. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0017] Figure 1 A chip design method flowchart of a silicon piezoresistive pressure sensor provided by the embodiment of the application;

[0018] Figure 2 A Bi-GRU neural network pressure chip numerical design structure schematic diagram provided for an embodiment of the present application is shown in FIG. 1.

[0019] Figure 3 A Bi-GRU neural network multivariable input multivariable output pressure chip design parameter forward model diagram provided for an embodiment of the present application is shown in FIG. 4.

[0020] Figure 4 A Bi-GRU neural network multivariable input multivariable output pressure chip performance parameter reverse model diagram provided for an embodiment of the present application is shown in FIG. 5. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0022] In addition, the terms “first”, “second”, and the like are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. In the description of the following embodiments, the meaning of “a plurality of” is two and more than two, unless otherwise specifically limited.

[0023] The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described again in some embodiments. To make the technical solutions of the present application clearer, the embodiments of the present application will be described in detail below in connection with the drawings.

[0024] Figure 1 A chip design method flowchart of a silicon piezoresistive pressure sensor provided for an embodiment of the present application is shown in FIG. 6, which comprises the following steps. Figure 1

[0025] S1: Constructing a design parameter set of the silicon piezoresistive pressure sensor.

[0026] In designing the silicon piezoresistive pressure sensor, the following key design parameters need to be considered: chip structure morphology, chip size, diaphragm size, piezoresistive doping concentration, heavy doping concentration, chip working temperature, and pressure range.

[0027] ​Each design parameter has an important influence on the performance of the silicon piezoresistive pressure sensor. The chip structure form involves the geometric shape and internal structure design of the sensor chip, such as the layout of the bridge resistance. The structure form will directly affect the pressure distribution and the conversion efficiency of the electrical signal. The chip size determines the mechanical properties and thermal properties of the sensor. A larger chip may provide a larger contact area and better thermal stability, but may increase the manufacturing cost and complexity. The diaphragm is the part of the sensor that senses pressure changes. The size of the diaphragm affects its sensitivity to pressure and response speed. The piezoresistive effect is achieved by doping impurities in the silicon wafer. The piezoresistive doping concentration determines the rate of resistance change and the sensitivity of the sensor. The heavily doped region is usually used to form a good ohmic contact or adjust the resistance value. The heavily doped concentration affects the long-term stability and temperature coefficient of the sensor. The chip operating temperature range affects the material properties and output stability of the sensor. It is necessary to ensure that the performance of the sensor remains consistent at the expected operating temperature. The pressure range is the range of pressures that the sensor can measure, from the minimum detectable pressure to the maximum pressure. The range defines the application scenarios and market demand for the sensor.

[0028] S2: Obtain a set of performance parameters corresponding to the set of design parameters according to the preset simulation software.

[0029] Using simulation analysis and numerical calculation, such as COMSOL The set of performance parameters corresponding to the set of design parameters can be obtained. The specific steps include:

[0030] Construct a basic model of the pressure sensor according to the chip size and diaphragm size;

[0031] Through preset numerical calculation, predict the performance parameters of the model under different design parameters, such as: by changing the piezoresistive doping concentration, chip operating temperature (-50-400℃), pressure range, etc. Design parameters to obtain model performance parameters.

[0032] Further, in the embodiments of the present application, the set of performance parameters includes: chip output voltage, sensitivity, linearity, natural frequency, maximum equivalent stress, thermal sensitivity drift and thermal zero point drift.

[0033] S3: Construct a numerical design library according to the set of design parameters and the set of performance parameters.

[0034] Constructing a numerical design library is to systematically organize the set of design parameters and the set of performance parameters to facilitate querying, analyzing and optimizing design.

[0035] Further, in order to ensure the consistency and comparability of the data, before constructing the numerical design library, it also includes: parameter standardization of the set of design parameters and the set of performance parameters.

[0036] Standardization usually refers to scaling data to fall into a small specific interval. The purpose of doing so is to remove the unit limit of data and convert it into a pure value without dimension, so that indicators with different units or magnitudes can be compared and weighted.

[0037] The embodiment of the application uniformly converts the data in the design parameter set and the performance parameter set into values between [0~1]. In this way, the influence of the design parameters on the performance parameters can be more effectively identified, and the best design scheme can be found.

[0038] S4: Establish a bidirectional numerical design model based on Bi-GRU neural network according to the numerical design library.

[0039] Figure 2 The Bi-GRU neural network pressure chip numerical design structure diagram provided by the embodiment of the application is shown in Figure 2 , wherein {x1,..,x i ,...,x i ,...,x j ,...,x 14} represent the input at time {t1,...t j ,...,t 14}, h0 represents the initial hidden vector, h7 and h 14 are the outputs of the forward and reverse GRU at the last time respectively.

[0040] Further, the specific steps of constructing the above model include:

[0041] S41: Build Anaconda integrated development environment;

[0042] The embodiment of the application adopts Anaconda integrated development environment, which is a popular Python data science and machine learning distribution, containing many pre-installed libraries and tools.

[0043] S42: Randomly sort and divide the data of the numerical design library;

[0044] In order to improve the generalization ability of the model and ensure the fairness and accuracy of the model evaluation, the data of the numerical design library needs to be randomly sorted and divided, and the specific steps include:

[0045] Load the data of the numerical design library;

[0046] Randomly sort the data of the numerical design library to ensure that the order of the data will not affect the results of model training.

[0047] The data sorted randomly is divided into a training set and a test set according to a preset proportion, and 80% of the data is used as the training set and 20% of the data is used as the test set in the embodiment of the application.

[0048] S43: Import the pytorch learning framework.

[0049] PyTorch is a very powerful tool that can quickly implement and iterate models, and to some extent, improve the efficiency of chip design.

[0050] S44: Build a forward numerical design model based on a Bi-GRU neural network.

[0051] Figure 3 The Bi-GRU neural network multivariate input multivariate output pressure chip design parameter forward model graph provided by the embodiment of the application is shown in FIG. Figure 3 The model is composed of four parts, namely, a design parameter input sequence, a Bi-GRU neural network model graph, seven fully connected layers, and performance parameter output.

[0052] Further, the specific steps for building the above forward model include:

[0053] S441: Set the model parameters, wherein the model parameters include: the number of hidden neurons, the learning rate, the batch size, the epoch value, and the multivariate input sequence length sequence.

[0054] S442: Convert the design parameter data in the training set into a three-dimensional tensor as the input of the first layer GRU forward neuron of the forward numerical design model.

[0055] S443: Reverse the design parameter sequence as the input of the second layer GRU reverse neuron; two layers of GRU can provide richer context semantic information, obtain the forward and reverse information of the design parameters, and improve the expression ability of the model.

[0056] S444: Access 7 layers of fully connected layers at the output layer of the neural network, each layer corresponding to a performance parameter prediction task; decompose the multivariate output into multi-task learning, and each task is a multivariate single output.

[0057] S445: Obtain the prediction results of the seven performance parameters according to the output of the fully connected layer.

[0058] S446: Calculate the loss function, calculate the loss of each output result, and then add the seven loss values to obtain the average as the final loss result.

[0059] S447: Complete the training of the forward numerical design model based on the Bi-GRU neural network.

[0060] S45: build a reverse numerical design model based on a Bi-GRU neural network.

[0061] Figure 4 A Bi-GRU neural network multivariable input multivariable output pressure chip performance parameter reverse model diagram provided for an embodiment of the application is shown in FIG. 1, which is composed of four parts, namely a performance parameter input sequence, a Bi-GRU neural network model diagram, seven fully connected layers, and design parameter outputs. Figure 4

[0062] Further, the specific steps for building the reverse model include:

[0063] S451: set model parameters, wherein the model parameters include: the number of hidden neurons, the learning rate, the batch size, the epoch value, and the input sequence length.

[0064] S452: convert the performance parameter data in the training set into a three-dimensional tensor as the input of the first layer GRU forward neuron of the reverse numerical design model.

[0065] S453: flip the performance parameter sequence as the input of the second layer GRU reverse neuron, through which the forward and reverse information of the performance parameter can be obtained, and the expression ability of the model can be improved.

[0066] S454: connect seven fully connected layers to the output layer of the neural network, each layer corresponding to a prediction task of a design parameter; decompose the multivariable output into multi-task learning, and each task is a multivariable single output.

[0067] S455: obtain the prediction results of the seven design parameters according to the output of the fully connected layer.

[0068] S456: calculate the loss function, calculate the loss of each output result, and then add the seven loss values to obtain the average as the final loss result.

[0069] S457: complete the training of the reverse numerical design model based on the Bi-GRU neural network.

[0070] Further, after completing the building of the Bi-GRU neural network-based bidirectional numerical design model, the model performance needs to be evaluated based on the test set data; and the parameters need to be adjusted or the model structure needs to be optimized according to the needs.

[0071] Specifically, the design parameter data in the test set is used as the input of the model to evaluate the effect of the forward numerical design model; and the performance parameter data in the test set is used as the input of the model to evaluate the effect of the reverse numerical design model.

[0072] ​In the above evaluation process, the model evaluation indexes include: mean absolute error MAE, root mean square error RMSE and mean absolute percentage error MAPE;

[0073]

[0074] Wherein, m is the number of different size chips, x i is the model input data, f(x i ) is the model prediction result, y i is the true value corresponding to the prediction result.

[0075] S46: save the trained bidirectional numerical design model based on Bi-GRU neural network.

[0076] In order to avoid repeated training process of the model, and can further analyze and deploy the model, it is necessary to save the trained bidirectional numerical design model based on Bi-GRU neural network, so as to use in different applications and environments.

[0077] The embodiment of the application also provides a chip design system of a silicon piezoresistive pressure sensor, which comprises a database module, a design parameter module, a performance parameter module and a model evaluation module.

[0078] The database module comprises a design parameter set and a performance parameter set; the design parameter set comprises chip structure morphology, chip size, diaphragm size, piezoresistive doping concentration, heavy doping concentration, chip working temperature and pressure range; the performance parameter set comprises chip output voltage, sensitivity, linearity, natural frequency, maximum equivalent stress, thermal sensitivity drift and thermal zero point drift.

[0079] The design parameter module is used for constructing a forward numerical design model based on Bi-GRU neural network according to the database module, and obtaining chip performance parameter result prediction according to design parameters.

[0080] The performance parameter module is used for constructing a reverse numerical design model based on Bi-GRU neural network according to the database module, and obtaining chip design parameter result prediction according to performance parameters.

[0081] The model evaluation module is used for evaluating model performance and adjusting parameters or optimizing model structure as needed.

[0082] The chip design method and system of the silicon piezoresistive pressure sensor provided by the application obtain corresponding performance parameters by using simulation software and numerical calculation according to design parameters, construct a numerical design library, and then use a Bi-GRU neural network to mine forward and reverse information of the design parameters and the performance parameters; wherein the first layer of GRU obtains forward information of the input sequence, and the second layer of GRU further extracts reverse features of the input sequence on the basis of the first layer, which can improve the expression ability of the model. Through this method, bidirectional design of the pressure chip design parameters and the performance parameters is realized, that is, through the target performance parameters, active design of the structure parameters and the piezoresistance is realized in reverse, and meanwhile, the performance parameter indexes of the pressure chip can also be obtained in forward through the design parameters.

[0083] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope and spirit of the application being indicated by the following claims.

[0084] It should be understood that the application is not limited to the precise construction that has been described above and shown in the accompanying drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application. The scope of the application is limited only by the appended claims.

Claims

1. A chip design method of a silicon piezoresistive pressure sensor, characterized by, include: A design parameter set for a silicon piezoresistive pressure sensor is constructed, which includes: chip structure morphology, chip size, diaphragm size, piezoresistive doping concentration, heavy doping concentration, chip operating temperature, and pressure range. The performance parameter set corresponding to the design parameter set is obtained according to the preset simulation software. The performance parameter set includes: chip output voltage, sensitivity, linearity, natural frequency, maximum equivalent stress, thermal sensitivity drift and thermal zero drift. The design parameter set and the performance parameter set are standardized. A numerical design library is constructed based on the standardized design parameter set and performance parameter set. Set up the Anaconda integrated development environment; load the data from the numerical design library; randomly sort the data from the numerical design library; divide the randomly sorted data into training and testing sets according to a preset ratio; import the PyTorch learning framework; build a forward numerical design model based on a Bi-GRU neural network; build a backward numerical design model based on a Bi-GRU neural network; save the trained bidirectional numerical design model based on a Bi-GRU neural network. Specifically, the performance parameter set corresponding to the design parameter set obtained from the preset simulation software includes: A basic model of the pressure sensor is constructed based on the chip size and the diaphragm size; the performance parameters of the model under different design parameters are predicted through preset numerical calculations.

2. The method of claim 1, wherein, The construction of the forward numerical design model based on the Bi-GRU neural network specifically includes: Set the model parameters, which include: the number of hidden neurons, learning rate, batch size, epoch value, and length of multivariate input sequence; The design parameter data in the training set is transformed into a three-dimensional tensor and used as the input to the first layer of the GRU forward neuron in the forward numerical design model. The design parameter sequence is flipped and used as the input to the second layer of GRU reverse neurons; Seven fully connected layers are connected to the output layer of the neural network, with each layer corresponding to a prediction task for a performance parameter; The prediction results for seven performance parameters are obtained based on the output of the fully connected layer; Calculate the loss function; Complete the training of a forward numerical design model based on a Bi-GRU neural network.

3. The method of claim 1, wherein, The construction of the inverse numerical design model based on the Bi-GRU neural network specifically includes: Set the model parameters, which include: the number of hidden neurons, learning rate, batch size, epoch value, and input sequence length; The performance parameter data in the training set is converted into a three-dimensional tensor and used as the input to the first layer of the GRU forward neuron in the inverse numerical design model. The performance parameter sequence is flipped and used as the input to the second layer of GRU reverse neurons; Seven fully connected layers are connected to the output layer of the neural network, with each layer corresponding to a prediction task with one design parameter. The predicted results for seven design parameters are obtained based on the output of the fully connected layer; Calculate the loss function; Complete the training of the inverse numerical design model based on the Bi-GRU neural network.

4. The method of claim 1, wherein, After completing the construction of the bidirectional numerical design model based on the Bi-GRU neural network, the following steps are also included: The model performance was evaluated based on the test set data. According to the need, the parameters are adjusted or the model structure is optimized.

5. The method of claim 4, wherein, The model performance is evaluated based on the test set data, and specifically includes: The design parameter data in the test set is used as the model input, and the effect of the forward numerical design model is evaluated; The performance parameter data in the test set is used as the model input, and the effect of the reverse numerical design model is evaluated.

6. The method of claim 4, wherein, The model evaluation index includes: mean absolute error (MAE), root mean square error (RMSE) and mean absolute percentage error (MAPE); Wherein, m is the number of different size chips, x i is the model input data, f(x i ) is the model prediction result, y i is the true value corresponding to the prediction result.

7. A chip design system of a silicon piezoresistive pressure sensor employing the chip design method according to any one of claims 1 to 6, characterized by It includes: A database module, a design parameter module, a performance parameter module and a model evaluation module. The database module includes a design parameter set and a performance parameter set; the design parameter set includes: chip structure morphology, chip size, diaphragm size, piezoresistive doping concentration, heavy doping concentration, chip working temperature and pressure range; the performance parameter set includes: chip output voltage, sensitivity, linearity, natural frequency, maximum equivalent stress, thermal sensitivity drift and thermal zero point drift; The design parameter module is used to construct a forward numerical design model based on a Bi-GRU neural network according to the database module, and to obtain chip performance parameter result prediction according to design parameters; the performance parameter module is used to construct a reverse numerical design model based on a Bi-GRU neural network according to the database module, and to obtain chip design parameter result prediction according to performance parameters; The model evaluation module is used to evaluate the model performance and to adjust the parameters or optimize the model structure as needed.

Images