CN110765723A - Routing modeling optimization method and device based on BP neural network - Google Patents
Routing modeling optimization method and device based on BP neural network Download PDFInfo
- Publication number
- CN110765723A CN110765723A CN201911119685.0A CN201911119685A CN110765723A CN 110765723 A CN110765723 A CN 110765723A CN 201911119685 A CN201911119685 A CN 201911119685A CN 110765723 A CN110765723 A CN 110765723A
- Authority
- CN
- China
- Prior art keywords
- routing
- modeling
- neural network
- optimization
- thickness
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Production Of Multi-Layered Print Wiring Board (AREA)
Abstract
The invention discloses a routing modeling optimization method and device based on a BP neural network, which comprises the following steps: acquiring the laminated information of the board cards, and performing routing modeling by using the parameter information of the board cards; obtaining an optimization parameter of the routing to be modeled, which meets impedance and loss indexes, by using a BP neural network; and calling the optimized intermediate file to establish a plurality of routing optimization models. The invention adds the BP neural network to realize the adaptivity of the routing impedance and the loss optimization, thereby improving the precision of the routing model and the efficiency of modeling optimization. The intermediate file generated by the routing modeling is utilized, and the modeling software is combined to realize one-time establishment of a plurality of routing models and realize batch processing of routing optimization, so that the efficiency is improved.
Description
Technical Field
The invention relates to the technical field of board card routing modeling, in particular to a routing modeling optimization method and device based on a BP neural network.
Background
With the continuous improvement of clock frequency, the method puts forward higher and higher requirements on the signal integrity of high-speed signals, evaluates link risks in the early stage of a project, and simulates a risk link, thereby having important significance for avoiding risks and shortening the project period. For the early stage of high-speed link simulation, models of device, connectors, cables, vias and wires in a link need to be collected. Where for device, connectors and cables, typically supplier supplied, the vias and traces need to be self-modeled. The routing is used as an important component in a high-speed link, and the precision of the model plays an important role in a simulation result.
At present, the wiring model in the industry is established by using ADS, IMLC and other software, the lamination information and the physical size and attribute of the wiring are input in the software so as to establish the wiring model, and the impedance and insertion loss of the wiring are calculated through simulation. In order to ensure the accuracy of the simulation result, the established routing model is required to meet the requirements of impedance and insertion loss, and therefore, routing needs to be optimized. The factors affecting the impedance and loss are many, for example, the line width, line spacing and dielectric constant of the dielectric material of the trace mainly affect the impedance, and the roughness of the copper foil, the dielectric loss of the dielectric material and the etching degree of the trace affect the loss of the trace. When the routing is optimized, in the prior art, each parameter needs to be manually adjusted according to experience so that the routing meets the requirements of impedance and loss. The manual adjustment mode is adopted for wiring optimization, so that the experience requirement of a designer is high, the optimization time is long, and when a large number of wirings are required to be established for link simulation, a large amount of time is required for early-stage modeling.
Disclosure of Invention
The invention aims to provide a routing modeling optimization method and device based on a BP neural network, which can improve the accuracy of a routing model and the efficiency of modeling optimization.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention provides a routing modeling optimization method based on a BP neural network, which comprises the following steps:
acquiring the laminated information of the board cards, and performing routing modeling by using the parameter information of the board cards;
obtaining an optimization parameter of the routing to be modeled, which meets impedance and loss indexes, by using a BP neural network;
and calling the optimized intermediate file to establish a plurality of routing optimization models.
With reference to the first aspect, in a first possible implementation manner of the first aspect, the acquiring the board stack information specifically includes:
and acquiring the laminated information of the board card, including the thickness of the signal layer, the thickness of the reference layer, the weight of copper, the type of copper foil, the conductivity, the roughness, the thickness of green oil, the thickness of the insulating substrate, the thickness of the core plate, the dielectric constant, the dielectric loss, the etching thickness of the routing, the impedance and the insertion loss of different signals of different routing layers and corresponding line width and line distance.
With reference to the first aspect, in a second possible implementation manner of the first aspect, the performing routing modeling by using board parameter information specifically includes:
inputting the wiring logarithm, the attack line, the victim line and the line width and line distance into a wiring modeling unit;
and inputting the thickness of the inner-layer wiring core board and the insulating substrate, the thickness of the wiring layer, the dielectric constant, the dielectric loss value, the etching thickness, the conductivity and the roughness of copper into the wiring modeling unit.
With reference to the first aspect, in a third possible implementation manner of the first aspect, the obtaining, by using a BP neural network, an optimized parameter that a to-be-modeled trace meets impedance and loss indexes specifically includes:
establishing an objective function formula of impedance Z and insertion loss IL:
IL＝f(ω1DK+ω2TW+ω3TD)
Z＝f(ω1*Df+ω2*rough+ω3*etch)
in the formula, DK is dielectric constant, TW is line width, TD is line distance, Df is dielectric loss, rough is roughness of copper, etc. is etching factor, omega1、ω2、ω3Is the weight;
determining the number of neurons of each layer of the neural network, the transfer function of each layer of neurons and the name of a function for training according to the target function formula;
initializing a weight value and a threshold value by utilizing a newff () function before the initialization step of the neural network;
the neural network transmits the error signal back according to the forward propagation path, and adjusts the connection weight system of each neuron of each hidden layer so as to lead the expected error signal to be minimum.
With reference to the first aspect, in a fourth possible implementation manner of the first aspect, the invoking the optimized intermediate file to establish a plurality of routing optimization models specifically includes:
acquiring an optimized intermediate file in an xml format generated in the routing modeling process;
and calling the optimized intermediate file and the wiring modeling unit to process the laminated information of the board cards in batch, and establishing a plurality of wiring models.
The invention provides a routing modeling optimization device based on a BP neural network, which comprises:
the board card parameter information acquisition module acquires board card lamination information and utilizes the board card parameter information to perform routing modeling;
the routing parameter optimization module is used for obtaining the optimization parameters of the routing to be modeled, which meet impedance and loss indexes, by utilizing a BP neural network;
and the batch routing modeling module calls the optimization intermediate file to establish a plurality of routing optimization models.
The routing modeling optimization device based on the BP neural network in the second aspect of the invention can realize the method in the first aspect and achieve the same effect.
The effect provided in the summary of the invention is only the effect of the embodiment, not all the effects of the invention, and one of the above technical solutions has the following advantages or beneficial effects:
the invention adds the BP neural network to realize the adaptivity of the routing impedance and the loss optimization, thereby improving the precision of the routing model and the efficiency of modeling optimization. The intermediate file generated by the routing modeling is utilized, and the modeling software is combined to realize one-time establishment of a plurality of routing models and realize batch processing of routing optimization, so that the efficiency is improved.
Drawings
FIG. 1 is a flow chart of one step of an embodiment of the method of the present invention;
FIG. 2 is a flow chart of two steps of an embodiment of the method of the present invention;
FIG. 3 is a flow chart of three steps of an embodiment of the method of the present invention;
FIG. 4 is a schematic diagram of an embodiment of the apparatus of the present invention.
Detailed Description
In order to clearly explain the technical features of the present invention, the following detailed description of the present invention is provided with reference to the accompanying drawings. The following disclosure provides many different embodiments, or examples, for implementing different features of the invention. To simplify the disclosure of the present invention, the components and arrangements of specific examples are described below. Furthermore, the present invention may repeat reference numerals and/or letters in the various examples. This repetition is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed. It should be noted that the components illustrated in the figures are not necessarily drawn to scale. Descriptions of well-known components and processing techniques and procedures are omitted so as to not unnecessarily limit the invention.
As shown in fig. 1, a routing modeling optimization method based on a BP neural network includes the following steps:
s1, acquiring board card lamination information, and performing routing modeling by using board card parameter information;
s2, obtaining the optimization parameters of the wiring to be modeled, which meet impedance and loss indexes, by using a BP neural network;
and S3, calling the optimized intermediate file to establish a plurality of routing optimization models.
As an embodiment of the present invention, in step S1, the acquiring the board stack information specifically includes:
and acquiring the laminated information of the board card, including the thickness of the signal layer, the thickness of the reference layer, the weight of copper, the type of copper foil, the conductivity, the roughness, the thickness of green oil, the thickness of the insulating substrate, the thickness of the core plate, the dielectric constant, the dielectric loss, the etching thickness of the routing, the impedance and the insertion loss of different signals of different routing layers and corresponding line width and line distance.
As an embodiment of the present invention, in step S1, performing routing modeling by using board parameter information specifically includes:
inputting the wiring logarithm, the attack line, the victim line and the line width and line distance into a wiring modeling unit;
and inputting the thickness of the inner-layer wiring core board and the insulating substrate, the thickness of the wiring layer, the dielectric constant, the dielectric loss value, the etching thickness, the conductivity and the roughness of copper into the wiring modeling unit.
In this embodiment, IMLC software is used for modeling the routing.
As shown in fig. 2, as an embodiment of the present invention, in step S2, obtaining an optimized parameter that a trace to be modeled satisfies impedance and loss indexes by using a BP neural network specifically includes:
s21, establishing an objective function formula of impedance Z and insertion loss IL:
IL＝f(ω1DK+ω2TW+ω3TD)
Z＝f(ω1*Df+ω2*rough+ω3*etch)
in the formula, TW is line width, TD is line distance, rough is roughness of copper, etc. is etching factor, omega1、ω2、ω3Is a weight value.
For the output variable insert loss, the relevant input variables are dielectric constant, linewidth TW, and linedistance TD. For the output variable impedance Z, the relevant input variables are dielectric loss, roughness of copper, etch factor etch.
Weight omega1、ω2、ω3This is achieved by training the network, and in general, MATLAB is used without setting itself, and net is newff (minmax (p)), [12, 4]After { 'tansig', 'logsig' }, 'train lm') it is automatically assigned, ranging from 0, 1]In the meantime. The weight is variable, and the line width TW, the line distance TD, Df, the roughness of copper roughnesss and the etching factor etch are variable.
Taking the above stack as an example, take a 50ohm single ended line as an example. DK1: 3.65; DK2: 3.71; TW [3.01, 4.01], TD 6 mil; df [0.08,0.1 ]; roughnesss 5.5; etch [0.3,0.5], target value for IL is 0.65Db, and target value for Z is 50 ohm.
S22, determining the number of neurons of each layer of the neural network, the transfer function of each layer of neurons and the name of a training function according to the target function formula;
s23, initializing a weight and a threshold by utilizing a newff () function before the neural network initialization step;
and S24, the neural network reversely transmits the error signal back according to the forward propagation path, and adjusts the connection weight system of each neuron of each hidden layer so as to enable the expected error signal to tend to be minimum.
The specific algorithm for optimizing the routing parameters by using the BP neural network is as follows:
% constructing input vector p in training sample
p＝[3.653.713.016；]
3.653.713.026；
3.653.713.036；
……
3.653.714.016]
% creating a BP network, the hidden layer has 3 neurons, and the transfer function is tansig
% of the intermediate layer has 3 neurons, the transfer function is logsig and the training function is trainlm
net＝newff(minmax(p),[3,3],{'tansig','logsig'},'trainlm')；
% training number default to 100
net.trainParam.epochs＝500；
Target for% training defaults to 0
net.trainParam.goal＝0.01；
% neural network training
net＝train(net,p)；
y＝sim(net,p)。
As shown in fig. 3, as an embodiment of the present application, in step S3, invoking an optimized intermediate file to establish a multiple trace optimization model specifically includes:
s31, acquiring an optimized intermediate file in an xml format generated in the routing modeling process;
and S32, calling the optimized intermediate file and the wiring modeling unit to process the laminated information of the board cards in batch, and establishing a plurality of wiring models.
And (3) calling IMLC software and an intermediate file in an xml format by using a batch processing script in a Bat format, carrying out batch processing, and establishing a plurality of wires at one time. The intermediate file generated by the routing modeling is utilized, and the modeling software is combined to realize one-time establishment of a plurality of routing models and realize batch processing of routing optimization, so that the efficiency is improved.
As shown in fig. 4, a routing modeling optimization apparatus based on a BP neural network includes:
the board parameter information acquisition module 11 is used for acquiring board lamination information and carrying out routing modeling by utilizing the board parameter information;
the routing parameter optimization module 12 obtains the optimization parameters of the to-be-modeled routing, which meet impedance and loss indexes, by using a BP neural network;
and the batch routing modeling module 13 calls the optimization intermediate file to establish a plurality of routing optimization models.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (6)
1. A routing modeling optimization method based on a BP neural network is characterized by comprising the following steps:
acquiring the laminated information of the board cards, and performing routing modeling by using the parameter information of the board cards;
obtaining an optimization parameter of the routing to be modeled, which meets impedance and loss indexes, by using a BP neural network;
and calling the optimized intermediate file to establish a plurality of routing optimization models.
2. The routing modeling optimization method based on the BP neural network according to claim 1, wherein the acquiring of the board card stack information specifically comprises:
and acquiring the laminated information of the board card, including the thickness of the signal layer, the thickness of the reference layer, the weight of copper, the type of copper foil, the conductivity, the roughness, the thickness of green oil, the thickness of the insulating substrate, the thickness of the core plate, the dielectric constant, the dielectric loss, the etching thickness of the routing, the impedance and the insertion loss of different signals of different routing layers and corresponding line width and line distance.
3. The routing modeling optimization method based on the BP neural network according to claim 2, wherein the modeling routing by using board parameter information specifically includes:
inputting the wiring logarithm, the attack line, the victim line and the line width and line distance into a wiring modeling unit;
and inputting the thickness of the inner-layer wiring core board and the insulating substrate, the thickness of the wiring layer, the dielectric constant, the dielectric loss value, the etching thickness, the conductivity and the roughness of copper into the wiring modeling unit.
4. A routing modeling optimization method based on a BP neural network according to claim 3, wherein the obtaining of the optimization parameter that the to-be-modeled routing satisfies the impedance and loss indexes by using the BP neural network specifically comprises:
establishing an objective function formula of impedance Z and insertion loss IL:
IL＝f(ω1DK+ω2TW+ω3TD)
Z＝f(ω1*Df+ω2*rough+ω3*etch)
in the formula, TW is line width, TD is line distance, rough is roughness of copper, etc. is etching factor, omega1、ω2、ω3Is the weight;
determining the number of neurons of each layer of the neural network, the transfer function of each layer of neurons and the name of a function for training according to the target function formula;
initializing a weight value and a threshold value by utilizing a newff () function before the initialization step of the neural network;
the neural network transmits the error signal back according to the forward propagation path, and adjusts the connection weight system of each neuron of each hidden layer so as to lead the expected error signal to be minimum.
5. The routing modeling optimization method based on the BP neural network according to claim 4, wherein the invoking an optimization intermediate file to establish a plurality of routing optimization models specifically comprises:
acquiring an optimized intermediate file in an xml format generated in the routing modeling process;
and calling the optimized intermediate file and the wiring modeling unit to process the laminated information of the board cards in batch, and establishing a plurality of wiring models.
6. A routing modeling optimization device based on a BP neural network is characterized by comprising:
the board card parameter information acquisition module acquires board card lamination information and utilizes the board card parameter information to perform routing modeling;
the routing parameter optimization module is used for obtaining the optimization parameters of the routing to be modeled, which meet impedance and loss indexes, by utilizing a BP neural network;
and the batch routing modeling module calls the optimization intermediate file to establish a plurality of routing optimization models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911119685.0A CN110765723B (en) | 2019-11-15 | 2019-11-15 | Routing modeling optimization method and device based on BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911119685.0A CN110765723B (en) | 2019-11-15 | 2019-11-15 | Routing modeling optimization method and device based on BP neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110765723A true CN110765723A (en) | 2020-02-07 |
| CN110765723B CN110765723B (en) | 2022-06-10 |
Family
ID=69337960
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911119685.0A Active CN110765723B (en) | 2019-11-15 | 2019-11-15 | Routing modeling optimization method and device based on BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110765723B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310400A (en) * | 2020-02-16 | 2020-06-19 | 苏州浪潮智能科技有限公司 | BP neural network-based capacitance anti-pad optimization method and system |
| CN111832247A (en) * | 2020-06-24 | 2020-10-27 | 苏州浪潮智能科技有限公司 | Method and device for determining size of via hole anti-pad based on BP neural network |
| CN112100817A (en) * | 2020-08-20 | 2020-12-18 | 上海机电工程研究所 | Intelligent heterogeneous IO data conversion method and system based on distributed simulation system |
| CN112395807A (en) * | 2020-11-12 | 2021-02-23 | 苏州浪潮智能科技有限公司 | Method and system for optimizing coupling of via hole and in-out wire after capacitance |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107358021A (en) * | 2017-06-01 | 2017-11-17 | 华南理工大学 | DO prediction model establishment method based on BP neural network optimization |
| CN108920841A (en) * | 2018-07-05 | 2018-11-30 | 中南大学 | New antenna design method |
| CN109543939A (en) * | 2018-10-11 | 2019-03-29 | 北京信息科技大学 | A kind of method of green building productions certification risk evaluation model building |
-
2019
- 2019-11-15 CN CN201911119685.0A patent/CN110765723B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107358021A (en) * | 2017-06-01 | 2017-11-17 | 华南理工大学 | DO prediction model establishment method based on BP neural network optimization |
| CN108920841A (en) * | 2018-07-05 | 2018-11-30 | 中南大学 | New antenna design method |
| CN109543939A (en) * | 2018-10-11 | 2019-03-29 | 北京信息科技大学 | A kind of method of green building productions certification risk evaluation model building |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310400A (en) * | 2020-02-16 | 2020-06-19 | 苏州浪潮智能科技有限公司 | BP neural network-based capacitance anti-pad optimization method and system |
| CN111310400B (en) * | 2020-02-16 | 2022-06-07 | 苏州浪潮智能科技有限公司 | BP neural network-based capacitance anti-pad optimization method and system |
| CN111832247A (en) * | 2020-06-24 | 2020-10-27 | 苏州浪潮智能科技有限公司 | Method and device for determining size of via hole anti-pad based on BP neural network |
| CN111832247B (en) * | 2020-06-24 | 2022-06-03 | 苏州浪潮智能科技有限公司 | Method and device for determining size of via hole anti-pad based on BP neural network |
| CN112100817A (en) * | 2020-08-20 | 2020-12-18 | 上海机电工程研究所 | Intelligent heterogeneous IO data conversion method and system based on distributed simulation system |
| CN112395807A (en) * | 2020-11-12 | 2021-02-23 | 苏州浪潮智能科技有限公司 | Method and system for optimizing coupling of via hole and in-out wire after capacitance |
| CN112395807B (en) * | 2020-11-12 | 2022-07-12 | 苏州浪潮智能科技有限公司 | Method and system for optimizing coupling of via hole and in-out wire after capacitance |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110765723B (en) | 2022-06-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110765723B (en) | Routing modeling optimization method and device based on BP neural network | |
| CN114417781B (en) | PCB wiring crosstalk evaluation method, system, device, equipment and storage medium | |
| CN110717310B (en) | Method and device for adjusting circuit layout | |
| CN112784519A (en) | PCB routing parameter setting device and method and storage medium | |
| CN106604550B (en) | A kind of line impedance adjusting method and system | |
| CN109299534A (en) | A kind of modeling method and device of printed circuit board | |
| Lho et al. | Eye-width and eye-height estimation method based on artificial neural network (ANN) for USB 3.0 | |
| CN104915496A (en) | Method for wiring differential signal wires and device | |
| CN107066764A (en) | A kind of copper foil roughness design method suitable for high-speed line model extraction | |
| Cheng et al. | Enhanced microstrip guard trace for ringing noise suppression using a dielectric superstrate | |
| CN111934637A (en) | Jellyfish-shaped low-loss negative group delay circuit and implementation method | |
| US20230168723A1 (en) | Thermal relief implementing method, electronic device and storage medium | |
| Roy et al. | Macromodeling of multilayered power distribution networks based on multiconductor transmission line approach | |
| Ravelo et al. | Fast estimation of RL‐loaded microelectronic interconnections delay for the signal integrity prediction | |
| CN115587057A (en) | Isometric design method and system for high-speed signals in server system | |
| CN114638077A (en) | Electromagnetic transient simulation method and device for power distribution network | |
| CN113033140A (en) | Simulation method for accurately obtaining difference of dielectric constants of upper layer and lower layer of PCB (printed Circuit Board) wiring | |
| Wu et al. | High Speed Muti-board Signal Integrity Simulation and Implementation | |
| CN112131825A (en) | Expansion and shrinkage prediction method, device, equipment and storage medium for copper plugging process FPC | |
| CN109117530B (en) | Method, device, equipment and storage medium for calculating loss of copper foil of transmission line | |
| Bandler | The journey to automated design optimization and a vision for the future | |
| CN111159970B (en) | Multi-physical-field analysis method and device for flexible interconnection reliability | |
| Narayana et al. | ANN models for coplanar strip line analysis and synthesis | |
| Win et al. | A frequency-domain high-speed bus signal integrity compliance model: Design methodology and implementation | |
| CN116738905A (en) | Method, device, equipment and medium for establishing mathematical model of high-speed differential line |
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 |