CN113569527B - Current source injection model building method based on machine learning - Google Patents
Current source injection model building method based on machine learning Download PDFInfo
- Publication number
- CN113569527B CN113569527B CN202110850762.0A CN202110850762A CN113569527B CN 113569527 B CN113569527 B CN 113569527B CN 202110850762 A CN202110850762 A CN 202110850762A CN 113569527 B CN113569527 B CN 113569527B
- Authority
- CN
- China
- Prior art keywords
- injection
- machine learning
- model
- current source
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/39—Circuit design at the physical level
- G06F30/398—Design verification or optimisation, e.g. using design rule check [DRC], layout versus schematics [LVS] or finite element methods [FEM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Geometry (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for establishing a current source injection model based on machine learning, which specifically comprises the following steps: obtaining a data set required by machine learning through 3D TCAD modeling simulation; giving a main body form of an injection current injection model based on physical and mathematical transformation of the device; selecting time information t to obtain a training set required by machine learning; selecting an f (r) machine learning model; training and optimizing based on the generated training set to obtain parameters corresponding to the f (r) model; and recovering a current source injection model based on the obtained f (r) parameter. The method for establishing the current source model is based on a machine learning model and an algorithm, and optimizes and models the core expression in the main body model through the data set obtained by 3D TCAD simulation, so that the method can cope with the physical characteristics under the advanced process.
Description
Technical Field
The invention belongs to the technical field of semiconductors and integrated circuits, relates to an irradiation effect simulation technology and an anti-radiation reinforcement technology in an aerospace electronic or nuclear explosion environment, and particularly relates to a single event effect evaluation technology.
Background
The single event effect refers to that high-energy charged particles in an irradiation environment can generate energy deposition when passing through a sensitive area of an electronic device so as to generate a large number of electron-hole pairs, and the electron-hole pairs are respectively collected by a corresponding n area and a p area in a drifting process, so that instantaneous pulse current is generated, and the logic state of a sensitive node of the device is influenced, wherein a phenomenon that the level of the node of the device is inverted by mistake is called a single event upset effect (Single Event Upset, SEU). In the evaluation of the single event upset effect, a method of injecting a current source is generally adopted. How to characterize the instantaneous pulse current with a current source is therefore important for evaluating the susceptibility of the semiconductor device to single event upset effects.
The current source injection model commonly used is the double exponential model proposed in the g.c. messenger, document "Collection of charge on junction nodes from ion tracks," IEEE Trans.Nucl.Sci., vol.NS-29, no.6, pp.2024-2031, dec.1982, as follows:
where Q is the amount of charge collected, τ α Is the time constant of the fall of the junction current, τ β Rise time constant of junction current, τ α And τ β T is a time variable depending on the process parameters. The model may evaluate the single event upset threshold used to evaluate the device, but the model cannot evaluate the impact of particles on surrounding devices, and is not suitable for advanced process use in some cases.
CN102982216a discloses a method for creating a current source model based on injection distance, which is based on a current source model based on injection diffusion in one dimension, and assumes that all charges are collected by the same sensitive node, unlike actual charges are collected by multiple sensitive nodes.
CN111079366a discloses a method for establishing a current source model facing charge sharing, which is based on a two-dimensional diffusion idea, introduces an injection distance and a reference distance, and solves the problem of charge sharing through the combined action of the injection distance and the reference distance.
Under the nano process, the physical characteristics of the semiconductor device become complex, even new physical characteristics appear, and the existing model is difficult to meet the requirements.
Disclosure of Invention
The invention aims to solve the problem that the existing current source injection model cannot cope with the physical characteristics under the advanced process, and provides a method for establishing the current source injection model based on machine learning.
The technical scheme of the invention is as follows: a method for establishing a current source injection model based on machine learning comprises the following specific steps:
s1, obtaining a data set required by machine learning through 3D TCAD modeling simulation;
s2, giving out a main body form of an injection current injection model based on device physical and mathematical transformation;
s3, selecting time information t to obtain a training set required by machine learning;
s4, selecting an f (r) machine learning model;
s5, training and optimizing based on the training set generated in the step S3 to obtain parameters corresponding to the f (r) model;
s6, recovering a specific expression of I (r, t) based on the f (r) parameter obtained in the step S5, wherein I (r, t) is the established current source injection model.
Further, the data set in step S1 is I (r, t) data obtained by 3D TCAD modeling simulation using the injection distance as an input condition, where r is the injection distance (i.e., the distance between the ion point and the collection point), and t is a time variable.
Further, the body form of the injection current injection model described in step S2 is expressed as:
wherein Q is L =10 (LET), LET is the linear transmission energy (as input parameter, depending on the irradiation environment),D n,p is the diffusivity of the carrier, D n Represents the diffusivity of electrons, D p Representing the diffusivity of holes, T is a fixed amount of time, r is the injection distance (i.e., the distance between the ion point and the collection point), I (r) is the injection current for an injection distance r, and f (r) is a functional form that machine learning needs to obtain.
Further, the training set in step S3 is I (r, t) data with the specified time variable being a fixed time variable and the corresponding input condition variable.
Further, the restored expression of I (r, t) in step S6:
the invention has the beneficial effects that: the method for establishing the current source model is based on a machine learning model and an algorithm, and optimizes and models the core expression in the main body model through the data set obtained by 3D TCAD simulation, so that the method can cope with the physical characteristics under the advanced process.
Drawings
FIG. 1 is a flow chart of the setup of the present invention.
Fig. 2 shows a combination of a current source injection model and an NMOS circuit according to an embodiment of the present invention, wherein the arrow indicates the current direction.
Fig. 3 shows a combination of a current source injection model and a PMOS circuit according to an embodiment of the present invention, wherein the arrow indicates the current direction.
FIG. 4 shows a 6T SRAM cell and a current source injection model according to an embodiment of the present invention.
Detailed Description
The invention will be further described with reference to the accompanying drawings and specific examples.
1. The injection current injection model with time variable is modeled as:
wherein Q is L ＝10(LET)，D n,p Is the diffusivity of the carrier, D n Represents the diffusivity of electrons, D p The diffusion rate of holes is represented, r is the diffusion distance, and related data such as "semiconductor physics" can be referred to.
2. Fixed time parameter t=t:
3. the relevant parameters of f (r) are obtained through machine learning training.
4. T +.t is used as the test set.
5. And finally, the obtained I (r, t) expression is the established current source injection model.
The flow of the invention is shown in figure 1, and the specific steps are as follows:
s1, obtaining a data set required by machine learning through 3D TCAD modeling simulation;
s2, giving a current source injection model main body form based on device physical and mathematical transformation;
s3, selecting time information t to obtain a training set required by machine learning;
s4, selecting an f (r) machine learning model;
s5, training and optimizing based on the training set generated in the step S3 to obtain parameters corresponding to the f (r) model;
s6, recovering the specific expression of the I (r, t) based on the f (r) parameter obtained in the step S5.
Based on the current source injection model described above, fig. 2 shows the current source injection model combined with an NMOS circuit, wherein the arrow is the current direction, and fig. 3 shows the current source injection model combined with a PMOS circuit, wherein the arrow is the current direction.
As shown in fig. 2, the current source injection model and the NMOS circuit combined injection model comprise two parts: an NMOS transistor and current source model, wherein D, S, B, G are the drain, source, base and gate of the NMOS transistor, respectively; the connection relation is as follows: the current source model is connected across the drain and the base of the NMOS transistor, and the current flows from the drain to the base of the transistor.
As shown in fig. 3, the current source injection model and PMOS circuit combined injection model comprises two parts: as shown in fig. 3, the PMOS transistor and current source model, wherein D, S, B, G are the drain, source, base, and gate of the PMOS transistor, respectively; the connection relation is as follows: the current source model is connected across the drain and the base of the PMOS transistor, and the current direction is from the base to the drain of the transistor.
When ions attack the corresponding transistor circuit, the circuit shown in fig. 2 and 3 can be used to simulate the equivalent row injection current source, so as to evaluate the influence of particles on the device under a certain distance.
The application of this model is illustrated below in the injection simulation of a single node of a specific application example 6T SRAM standard cell:
as shown in fig. 4, VDD is a power supply, GND is ground, transistors M1, M2, M3, M4, M5 and M6 form a standard 6T SRAM cell according to the connection mode shown in the figure, wherein M1, M2, M5 and M6 are NMOS transistors, M3 and M4 are PMOS transistors, W is a control input signal, B and BN are write signals, cord and cord_are internal level holding nodes, the standard SRAM cell is available in the relevant literature, current I (r, T) is a current source according to an embodiment of the present invention, and r represents a distance from an ion injection point to an M1 collection point.
The specific application process for the SRAM cell is as follows:
(1) Completing an SRAM cell according to the circuit structure design shown in FIG. 4;
(2) Based on the result of the previous machine learning, constructing a current source based on I (r, t) in a simulation platform of the circuit, and selecting different injection models for NMOS and PMOS according to transistors under particle attack; without loss of generality, the transistors for selecting particle attack are M1, M1 NMOS transistors, the connection relation of injection current source models of which is shown in figure 3, and the current sources are connected across the D end and the B end of the M1 transistor;
(3) Setting related parameters of I (r, t) according to the technological parameters and the particle types, and setting an injection distance r;
(4) Performing circuit simulation, and observing the damage degree of a current source of I (r, t) to the logic state of the circuit at the injection distance; if the logic state of the SRAM cell is flipped (single event flip effect) it is representative that the particle can affect the corresponding device within the implantation distance range.
Those of ordinary skill in the art will recognize that the embodiments described herein are for the purpose of aiding the reader in understanding the principles of the present invention and should be understood that the scope of the invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific modifications and combinations from the teachings of the present disclosure without departing from the spirit thereof, and such modifications and combinations remain within the scope of the present disclosure.
Claims (3)
1. A method for establishing a current source injection model based on machine learning comprises the following specific steps:
s1, obtaining a data set required by machine learning through 3D TCAD modeling simulation; the data set is I (r, t) data obtained by 3D TCAD modeling simulation by taking the injection distance as an input condition, wherein r is the injection distance, and t is a time variable;
s2, giving out a main body form of an injection current injection model based on device physical and mathematical transformation;
the main body form of the injection current injection model is expressed as follows:
wherein Q is L =10 (LET), which is the linear transmission energy, is the input parameter, depending on the irradiation environment,D n,p is the diffusivity of the carrier, D n Represents the diffusivity of electrons, D p Representing the diffusivity of holes, wherein T is a fixed amount of time, r is the injection distance, namely the distance between an ion point and a collection point, I (r) is the injection current with the injection distance r, and f (r) is a functional form which needs to be obtained by machine learning;
s3, selecting time information t to obtain a training set required by machine learning;
s4, selecting an f (r) machine learning model;
s5, training and optimizing based on the training set generated in the step S3 to obtain parameters corresponding to the f (r) model;
s6, recovering a specific expression of I (r, t) based on the f (r) parameter obtained in the step S5, wherein I (r, t) is the established current source injection model.
2. The method for building a current source injection model based on machine learning according to claim 1, wherein the training set in step S3 is I (r, t) data with a fixed time variable as a specified time variable and corresponding input condition variables.
3. The method for building a machine learning based current source injection model according to claim 2, wherein the restored expression of I (r, t) in step S6 is:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110850762.0A CN113569527B (en) | 2021-07-27 | 2021-07-27 | Current source injection model building method based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110850762.0A CN113569527B (en) | 2021-07-27 | 2021-07-27 | Current source injection model building method based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113569527A CN113569527A (en) | 2021-10-29 |
| CN113569527B true CN113569527B (en) | 2023-09-29 |
Family
ID=78167905
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110850762.0A Active CN113569527B (en) | 2021-07-27 | 2021-07-27 | Current source injection model building method based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113569527B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2845180A1 (en) * | 2002-09-27 | 2004-04-02 | St Microelectronics Sa | CMOS logic cell characterizing method for accelerating simulation of temporal dependence of delays effect in design of CMOS logic cells |
| CN102982216A (en) * | 2012-12-18 | 2013-03-20 | 电子科技大学 | Method for establishing current source model on the basis of implantation distance |
| CN108508351A (en) * | 2018-03-30 | 2018-09-07 | 西北核技术研究所 | A kind of single-particle direct fault location emulation mode based on double-two fingers number current source |
| CN111064182A (en) * | 2019-11-25 | 2020-04-24 | 国网浙江省电力有限公司湖州供电公司 | Short-circuit current calculation method based on power grid characteristics |
| CN111079366A (en) * | 2019-12-19 | 2020-04-28 | 电子科技大学 | Charge sharing-oriented current source model establishing method |
| CN111159841A (en) * | 2019-11-25 | 2020-05-15 | 国网浙江省电力有限公司湖州供电公司 | Power distribution network short-circuit current calculation method based on data fusion |
-
2021
- 2021-07-27 CN CN202110850762.0A patent/CN113569527B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2845180A1 (en) * | 2002-09-27 | 2004-04-02 | St Microelectronics Sa | CMOS logic cell characterizing method for accelerating simulation of temporal dependence of delays effect in design of CMOS logic cells |
| CN102982216A (en) * | 2012-12-18 | 2013-03-20 | 电子科技大学 | Method for establishing current source model on the basis of implantation distance |
| CN108508351A (en) * | 2018-03-30 | 2018-09-07 | 西北核技术研究所 | A kind of single-particle direct fault location emulation mode based on double-two fingers number current source |
| CN111064182A (en) * | 2019-11-25 | 2020-04-24 | 国网浙江省电力有限公司湖州供电公司 | Short-circuit current calculation method based on power grid characteristics |
| CN111159841A (en) * | 2019-11-25 | 2020-05-15 | 国网浙江省电力有限公司湖州供电公司 | Power distribution network short-circuit current calculation method based on data fusion |
| CN111079366A (en) * | 2019-12-19 | 2020-04-28 | 电子科技大学 | Charge sharing-oriented current source model establishing method |
Non-Patent Citations (6)
| Title |
|---|
| DTC无速度传感器运行方法综述;林琪;赵亮;李智强;;科技信息(第21期) * |
| 刘畅 ; 张庆范 ; 郑伟杰 ; .基于反向传播算法神经网络的谐波源模型分析.现代电子技术.2008,(第06期), * |
| 基于DSP的开关磁阻电机磁链特性检测与神经网络建模;薛梅;夏长亮;王慧敏;谢细明;;电工技术学报(第02期) * |
| 基于反向传播算法神经网络的谐波源模型分析;刘畅;张庆范;郑伟杰;;现代电子技术(第06期) * |
| 林琪 ; 赵亮 ; 李智强 ; .DTC无速度传感器运行方法综述.科技信息.2010,(第21期), * |
| 薛梅 ; 夏长亮 ; 王慧敏 ; 谢细明 ; .基于DSP的开关磁阻电机磁链特性检测与神经网络建模.电工技术学报.2011,(第02期), * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113569527A (en) | 2021-10-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Singhee et al. | Statistical blockade: a novel method for very fast Monte Carlo simulation of rare circuit events, and its application | |
| Lin et al. | Analysis and design of nanoscale CMOS storage elements for single-event hardening with multiple-node upset | |
| Zhang et al. | Analytical modeling of SRAM dynamic stability | |
| CN103577643A (en) | SRAM type FPGA single event upset effect simulation method | |
| CN109918723B (en) | Single-particle fault injection method based on particle incidence randomness | |
| Woodruff et al. | Three-dimensional numerical simulation of single event upset of an SRAM cell | |
| CN103198198B (en) | Single-event transient injection simulation method for high-frequency circuits | |
| Dong et al. | SRAM dynamic stability: Theory, variability and analysis | |
| CN108363893A (en) | A kind of single-particle pulse current source modeling method of complex condition | |
| Marukame et al. | Error tolerance analysis of deep learning hardware using a restricted boltzmann machine toward low-power memory implementation | |
| Rezaei et al. | Fundamental thermal limits on data retention in low-voltage CMOS latches and SRAM | |
| Raji et al. | Soft error rate estimation of combinational circuits based on vulnerability analysis | |
| CN113591320B (en) | Coupling simulation method for hot carrier effect and total dose effect | |
| CN113569527B (en) | Current source injection model building method based on machine learning | |
| CN106649920A (en) | IBIS-based integrated circuit total dose effect modeling method | |
| CN111079366B (en) | Charge sharing-oriented current source model establishing method | |
| Qi et al. | The device compact model based on multi-gradient neural network and its application on mos 2 field effect transistors | |
| CN102982216B (en) | Method for establishing current source model on the basis of implantation distance | |
| Royer et al. | Evolution of radiation-induced soft errors in FinFET SRAMs under process variations beyond 22nm | |
| Poudel et al. | Design and evaluation of a PVT variation-resistant TRNG circuit | |
| CN112836386B (en) | Novel total dose simulation method and single particle coupling simulation method and device thereof | |
| Cummings et al. | Comparison of discretization methods for device simulation | |
| Kauppila et al. | Impact of process variations on upset reversal in a 65 nm flip-flop | |
| Marques et al. | Soft Error Reliability of SRAM cells during the three operation states | |
| Camargo et al. | Use of SSTA tools for evaluating BTI impact on combinational circuits |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |