Product
Patsnap Eureka
For R&D, Patsnap Eureka makes reading and utilizing patents & technical documents easy.
Patsnap Eureka AIR
Designed for self-driven R&D workflows. Generate viable solutions, solve complex R&D challenges, empower your innovation with AI.
Patsnap Eureka Materials
Designed for material experts only. Revolutionize your material R&D, from search, analyze, to developing new materials.
Feature
TechResearch
Generate reliable direction feasibility study reports for your R&D in just a few steps.
TechSeek
Discover and master advanced knowledge NOW. Basics, ideas, possibilities, all at once.
TechMind
As an expert in R&D Theories, TechMind can generates customized viable solutions instantly.
TechRisk
Analyze your overall solution with one click, know your potential R&D risks in advance.
TechMonitor
Get weekly tech updates, stay abreast of the latest tech innovations and key insights.
Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
50results about How to "Achieve streamlining" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
Neural network relation classification method and an implementation system thereof, which fuse discrimination degree information
ActiveCN109344244ASolve the confusionWork lessSpecial data processing applicationsText database clustering/classificationFeature vectorRelation classification
The invention relates to a neural network relation classification method and an implementation system thereof, which fuse discrimination degree information, comprising: (1) data preprocessing; (2) training word vector; (3) extracting position vector; Combining the word vector of each word with the high-dimensional position feature vector, the joint feature is obtained. (4) computing the semantic representation of sentences; The context information and semantic information of entities are encoded by bi-directional LSTM. (5) calculating a discrimination vector; After the two entity vectors are subtracted from each other and cascaded with a Bi-LSTM unit encode that input; (6) inputting the outputs of (4) and (5) to the CNN, outputting the feature vector fused with the discrimination information, and inputting the feature vector to the classifier for classification; (5) adopting a Loss function to train the model. The invention does not need to manually extract any features, and the jointmodel does not need to preprocess the data by means of other natural language processing tools, the algorithm is simple and clear, and the effect is best at present.
Owner:SHANDONG UNIV
Emotion classification method based on a tensor fusion mode
ActiveCN109614487AWork lessAvoid interferenceSemantic analysisSpecial data processing applicationsFeature vectorAlgorithm
The invention relates to an emotion classification method based on a tensor fusion mode. The method comprises the steps of (1) data preprocessing; (2) training word vectors; (3) modeling specific tasks; the context information and the semantic information of each sentence are coded through bidirectional LSTM, and the feature vector representation of each sentence is obtained through the effect ofeach layer of network; and (4) tensor fusion: fusing the three feature vectors output by the model by adopting a tensor fusion mode, obtaining an optimal and comprehensive feature representation through information fusion, and then sending the feature vectors formed after fusion to a classifier for emotion classification. (5) training model. Any feature does not need to be extracted manually, themodel does not need to preprocess data by means of an additional natural language processing tool, meanwhile, emotion classification does not need to be conducted by recognizing current words in advance, the algorithm is simple and clear, and the effect is obvious.
Owner:SHANDONG UNIV
CPU power consumption test method and CPU power consumption test system
ActiveCN104502690AAchieve streamliningReduce adverse effectsElectric devicesElectric power measurementTest requirementsCPU power dissipation
The invention discloses a CPU power consumption test method and a CPU power consumption test system. According to the CPU power consumption test method and the CPU power consumption test system, service programs configured in an operating system booting configuration file are screened in advance according to test requirements, unnecessary service programs for a test process are eliminated, and an operating system is simplified; when a CPU (central processing unit) power consumption test for a target device is performed on the basis of the simplified operating system, environmental parameters of the target device are controlled according to preset corresponding target numerical values, CPU load is generated through simulation, operation pressure is applied to a CPU of the target device, output current of a battery of the target device is acquired on that basis, and CPU power consumption measurement of the target device is realized; therefore, the CPU load is generated through simplification of the operating system and simulation, negative influences of the system and application programs on test results are reduced, and accurate data support is provided for system power consumption optimization.
Owner:THUNDERSOFT
Permission dispatching method and device
ActiveCN104598778AMinimization and Optimal SchedulingReduce workloadDigital data authenticationUser needsWorkload
The invention provides a permission dispatching method and device, and relates to the technical field of permission control. The permission dispatching method and device is designed for solving the problems of a traditional static setting permission dispatching method that administrators have large workloads and the permission can not realize optimized and minimized dispatching. The method comprises the following steps: obtaining permission information which comprises permission use frequency; according to relationship between the permission information and a preset function, calculating a permission dispatching factor; and comparing the permission dispatching factor with a preset threshold value to obtain a comparison result, and dispatching the permission according to a preset dispatching instruction mapped by the comparison result. The permission dispatching method and device realizes the intellectualization of the permission dispatching, causes the distribution and the recall of the permission to be simpler, more conforms to user requirements and improves information security.
Owner:CHINA MOBILE GROUP JIANGSU
Market risk assessment method based on power transaction data
PendingCN111950918AMonitor market irregularitiesAchieve streamliningMarket predictionsBuying/selling/leasing transactionsFinancial transactionIndex system
The invention relates to a market risk assessment method based on power transaction data, and the method comprises the steps: 1), carrying out the dimension reduction of a power market risk monitoringfull-index library based on a Lasso regression model, and constructing a power market risk monitoring index system; 2) determining an index threshold value by adopting a mean number principle method,screening abnormal data, and carrying out preliminary identification on a risk object; 3) calculating the subjective weight and the objective weight of each index by using an improved CRITICG1 method, calculating the comprehensive weight of each index by using an improved game theory combination weighting method, constructing an aggregation model, and determining the risk level of the screened market subject by using a comprehensive evaluation method; and 4) performing risk assessment on the market risk of the market subject according to the risk level, and establishing different early warning mechanisms to warn the market risk. Compared with the prior art, the method has the advantages of effectively monitoring potential market violation behaviors of different market subjects, improvingthe precision and speed of power market risk assessment and the like.
Owner:江苏电力交易中心有限公司 +1
Method for simplifying five-loudspeaker system to four-loudspeaker system
The invention discloses a method for simplifying a five-loudspeaker system to a four-loudspeaker system. The method comprises the following steps of arranging a substitutive loudspeaker group including four loudspeakers, obtaining a signal distributed by synthesizing four loudspeakers into one virtual sound source, simulating the virtual sound source by using a loudspeaker, wherein the four loudspeakers and the virtual sound source are on the same spherical surface, and a sphere center position of the spherical surface is a receiving point; calculating initial distribution coefficients of the four loudspeakers and determining a final distribution coefficient; for a multi-sound channel system including four loudspeakers, applying the initial signal of single loudspeaker and the corresponding signal distributed by the loudspeakers in the step 1. The conventional method for synthesizing Ando four loudspeakers into one virtual sound source gives a series of loudspeaker distribution coefficients, but does not give the optimal loudspeaker distribution coefficient. The method disclosed by the invention can obtain the optical distribution coefficient for synthesizing four loudspeakers into one virtual sound source, and is beneficial to improving effects for synthesizing the virtual sound source.
Owner:WUHAN UNIV
Multifunctional installation and investigation vessel
ActiveCN103395479AAchieve streamliningImprove efficiencyWaterborne vesselsSpecial purpose vesselsWinchEnergy converter
The invention relates to a multifunctional installation and investigation vessel, which comprises a vessel body and is characterized in that an A-type lifting frame, an oceanographic engineering crane and a derrick are sequentially arranged in the X direction on a main deck of the vessel body; a crane winch is arranged below the main deck, and a steel wire rope passes through the main deck; the crane and the crane winch are respectively arranged in positions on two sides of the vessel body; multiple empty tanks are arranged between the inner bottom and the outer bottom of the vessel body and in positions on two sides of the vessel body, and multiple lower water ballast tanks are arranged on the inner sides of the two sides of the vessel body; opposite tanks on the two sides of the vessel body are communicated via communication passages; an energy converter tank is also arranged at the bottom of the vessel body. When the tanks are damaged, water can enter the empty tanks uniformly, so that the stability is improved; when water in the empty tanks on one side of the vessel body is excessive, water can flow from one side to the other side of the vessel body via the communication passages, so that the transverse inclination after damage of the vessel can be reduced and the damaged stability is improved.
Owner:SHANGHAI MERCHANT SHIP DESIGN & RES INST
Control tree generation method and terminal equipment
ActiveCN109656827AInitial control tree simplificationSmall scaleSoftware testing/debuggingTree generationTarget control
The invention discloses a control tree generation method, which comprises the following steps: generating an initial control tree of an application to be tested, nodes in the initial control tree being in one-to-one correspondence with controls in the application to be tested; Determining at least one target node according to an area where each node on the initial control tree is located; And based on the at least one target node, generating a target control tree so as to test the to-be-tested application. According to the embodiment of the invention, the control in the control tree can be efficiently searched, so that the efficiency of automatic testing of the application is improved.
Owner:BEIJING YUNCE INFORMATION TECH CO LTD
Mechanism for locating and guillotining terminal
InactiveCN101350490ASave man hoursAchieve streamliningContact member manufacturingPistonSalient point
The invention discloses a terminal locating and cutting mechanism, comprising a main body, wherein, a first opening is arranged near the top part, and a C-shaped first slide block is flexibly arranged in the first opening, a second opening is flexibly arranged near the bottom end of the main body, and a L-shaped second slide block is arranged in the second opening; a cutting tool fixed on the first slide block; a cutting table which is fixed on the second slide block by the bottom end provided with a plurality of bottom holes, and at least one locating salient point is arranged on the cutting table; a supporting block which is arranged under the second fixing block on the main body, and a plurality of top holes are arranged at the top end of the supporting block; a plurality of springs, one end of each spring is arranged in a bottom hole, and the other end is arranged in a top hole; a first elbow lever with one end arranged in the opening of the C-shaped first slide block; a second elbow lever with one end arranged on the bottom end of the L-shaped second slide block; a cylinder connected with the first elbow lever and the second elbow lever by the piston at one end away from the main body.
Owner:MITAC PRECISION TECH(KUNSHAN) CORP
Precise adjustment auxiliary device for track board of high speed railway and adjusting method thereof
InactiveCN102936868AAvoid groundingAchieve streamliningRailway track constructionEngineeringMechanical engineering
Owner:SHANGHAI BAOSTEEL IND TECHNOLOGICAL SERVICE
Interest list pushing system based on heterogeneous information network
ActiveCN112532755APush implementationEasy to filterDigital data information retrievalTransmissionTime informationComputer network
The invention relates to an interest list pushing system based on a heterogeneous information network, and the system comprises an obtaining module which is used for obtaining the login information and keyword information of a user; an establishing module which is used for establishing an information matrix I according to the login information of the user; a determining module which is used for determining an interest list of a current user according to the information matrix and the login information; a pushing module which is used for pushing the interest list to the second user; a central processing unit which is connected with the obtaining module, the establishing module, the determining module and the pushing module and used for adjusting real-time information in the interest list inthe determining module according to an information matrix in the establishing module and pushing the interest list of the first user to the second user, In addition, due to the fact that the pushinginformation received by the second user can obtain the list which the second user is interested in, screening and filtering of the information by the user are facilitated, and the information spreading efficiency is improved.
Owner:广州锐竞信息科技有限责任公司
Terrain section data thinning method, device and equipment
ActiveCN112562079AAchieve streamliningReduce manual workloadGeometric CADDetails involving 3D image dataTerrainAlgorithm
The embodiment of the invention discloses a terrain section data thinning method, device and equipment, and the method comprises the steps: obtaining terrain section data of a target section; recursively obtaining a section highest point and a section lowest point in each data point corresponding to the topographic section data according to a preset horizontal spacing threshold value; according tothe angular point characteristics of the target section, obtaining forced reservation data points, and segmenting terrain section data based on the forced reservation data points, the highest point of the section and the lowest point of the section to obtain segmented terrain data; and according to a preset rarefaction threshold value, adopting an improved Douglas pucker algorithm to rarefactionthe segmented terrain data to obtain rarefied terrain section data, wherein the threshold value index of the improved Douglas pucker algorithm is the ratio of the vertical distance between the data point and the connecting line of the end points on the two sides to the horizontal distance between the end points on the two sides. The method, device and equipment are suitable for thinning the multi-section line data of the engineering terrain section, especially the bridge terrain section.
Owner:CHINA RAILWAY SIYUAN SURVEY & DESIGN GRP
Signal line fixing structure of resolver
The invention provides a signal line fixing structure of a resolver. The signal line fixing structure is fixed on a shaft end of a stator of the resolver, and comprises the following components: a circuit board; a fixing part arranged on the circuit board and used for combining the circuit board on the end face of the shaft end of the stator; a terminal combination base arranged on the circuit board; multiple poroid connection parts respectively arranged on the circuit board in a penetration manner; and multiple bridge circuits, which are respectively located between one corresponding connection part and the terminal combination base, and used for bridging the corresponding connection part and the terminal combination base. The product disclosed by the invention can be suitable for the stator with less or more teeth, can greatly reduce the extending length of a lead, and can also avoid the situation where each bridge circuit is in short circuit with a stator coil or each wind-out end, thereby ensuring the reliability of the product; and the product disclosed by the invention can further integrate the electronic parts needed for the resolver on the circuit board, thereby realizing the effect of simplification of components.
Owner:HIWIN MIKROSYST
Method for hidden danger and fault feature extraction for train running system
ActiveCN103674586APlay an early warning functionAvoid the shortcomings of poor time domain resolution and insufficient sensitivity of non-stationary signal detectionRailway vehicle testingMultiple sensorFeature extraction algorithm
The invention relates to a method for hidden danger and fault feature extraction for a train running system. The method comprises the following steps that (1) real-time operation data of a train are collected by a sensor installed on the train and speed signals in the running process of the train are obtained through a train-mounted network; (2) multiple sensor network data fusion processing is conducted on the real-time operation data and the speed signals, so that source data of hidden danger and fault feature extraction are obtained; (3) a feature extraction algorithm is conducted on the source data, so that feature quantities of different components of the running system are obtained; (4) a feature quantity value which is representative and capable of reflecting the operation state of the running system is selected from the feature quantities obtained in the step (3), the redundancy among the feature quantities is removed, and the final hidden danger and fault feature quantity of the running system is obtained. Compared with the prior art, the data extracted through the method is high in accurate and the running safety can be effectively improved.
Owner:TONGJI UNIV
Similarity detection method and system based on power information system code file
ActiveCN110471835AAchieve streamliningImprove detection efficiencyCharacter and pattern recognitionSoftware testing/debuggingProgramming languageCall tree
The invention discloses a similarity detection method and system based on a power information system code file, and the method comprises the steps: obtaining a first code file and a second code file which need to judge the similarity, and respectively obtaining a first text and a second text after preprocessing; obtaining the text semantic word vector according to the TF-IDF value of the word, respectively searching function call tree structures of the first text and the second text from the function call entrances of the first text and the second text, and calculating to obtain a first text structure vector and a second text structure vector; calculating an intermediate semantic word vector through a text semantic word vector, taking a union set of the first text structure vector and thesecond text structure vector, and calculating a first intermediate structure vector and a second intermediate structure vector; and obtaining the similarity between the first text and the second text.Firstly, a preprocessing function is adopted to perform code simplification on a code file, the detection efficiency is improved, and the detection error rate is reduced.
Owner:NANJING NARI GROUP CORP +2
Method and system for classification and pre-processing of big data under Internet environment
InactiveCN106055557AAchieve streamliningWeb data indexingCharacter and pattern recognitionData setThe Internet
The invention relates to a method and system for classification and pre-processing of big data and especially relates to the method for the classification and the pre-processing of the big data under an Internet environment. The method and the system belong to the field of data exaction. The method provided by the invention comprises the steps that multiple types of network data in the Internet is used to compose a complete pre-processing basic dataset, and the data is simplified through operations such as dimension reduction; and then, the different types of data in the dataset is analyzed and pre-processed respectively, and a dataset used for classification is obtained, so that a data preparation is made for further classification.
Owner:INST OF SCI & TECHN INFORMATION OF CHINA
Sound field reconstruction method based on high-order microphone array
ActiveCN111263291AImprove reconstruction accuracyReduce measurement errorStereophonic systemsHigher order harmonicsFourier transform
The invention relates to a sound field reconstruction method based on a high-order microphone array. The method comprises the steps that firstly, a high-order microphone is used for collecting sound pressure of a local sound field, and a local sound field coefficient is calculated according to a discrete form of inverse Fourier transform; then regularization is carried out on a high-order harmoniccomponent in a low-order sound field according to a truncation order limit of the low-order sound field, and a regularized low-order sound field coefficient is obtained; a conversion matrix is calculated according to the orders of the high-order microphones and the topological structure of the array, and the local sound field coefficient obtained at each high-order microphone is converted into aglobal sound field coefficient of a to-be-reconstructed region; and finally, sound pressure distribution of the reconstructed sound field is acquired according to the harmonic expression of the soundfield.
Owner:西北工业大学太仓长三角研究院 +1
A GPU card cluster configuration control system and method
InactiveCN109739560AReduce development costsFlexible applicationResource allocationBootstrappingNetwork expansionNetwork interface
The invention discloses a GPU card cluster configuration control system. The system comprises at least one GPU node and a network expansion unit connected with a network communication end of the GPU node. The GPU node comprises a BMC control module, a management network interface module and at least one GPU card expansion module, wherein the management network interface module and the at least oneGPU card expansion module are respectively connected with the BMC control module, and the GPU card expansion module is connected with a GPU card. The invention further provides a GPU card cluster configuration control method. By integrating the management resources, the redundant resources in the GPU nodes are integrated, the system resources are reasonably utilized, and the development cost is saved.
Owner:ZHENGZHOU YUNHAI INFORMATION TECH CO LTD
Three-dimensional audio downsizing method and system
A three-dimensional audio downsizing method and system are disclosed. The method includes acquiring the spatial and positional information of L loudspeakers and the spatial and positional information of human head characteristics in an original three-dimensional multitrack audio system, converting the input time-domain signals of the L loudspeakers to obtain corresponding frequency domain signals, calculating the total sound pressure of the sound signals played by the L loudspeakers at left ear, the right ear and the head center, fining and removing one loudspeaker with minimal distortion of an original sound field from the current loudspeaker set to be downsized, continuously downsizing the current new loudspeaker set to be downsized until the optimal spatial and positional assignment of loudspeakers of an M channel system is obtained, and converting the frequency-domain signals on corresponding loudspeaker of the M channel system into time-domain signals through inverse Fourier transform. A 3D spatial sound field of a head area can be reconstructed with fewer loudspeakers, and the method and system are applicable to household environment.
Owner:WUHAN UNIV
Pipeline Interface Wave Ultrasonic Tracking System and Method
ActiveCN104237377BReduce data volumeLower requirementAnalysing solids using sonic/ultrasonic/infrasonic wavesPipeline systemsSonificationInterface waves
Owner:NANTONG UNION DIGITAL TECH DEV
RUP quantitative process cutting method aiming at software features
InactiveCN101620534AGuaranteed reliabilityImprove development efficiencySpecific program execution arrangementsSoftware development processSoftware engineering
The invention discloses an RUP quantitative process cutting method aiming at software features, comprising RUP flow expansion cutting and RUP expansion flow applicability cutting; the RUP flow expansion cutting aims at requirements of software features and carries out expansion to an RUP standard flow, so as to form special working flow for ensuring the requirements of a certain software feature; the RUP expansion flow applicability cutting leads the expanded special working flow to carry out process applicability cutting relevant to elements and element attribute, the corresponding cutting operations are ensured by the relevancy degree among all the elements, and the relevancy degree among all the elements is ensured by a fuzzy comprehensive evaluation method. The method solves the problems that the cost is overhigh and the development is overloading in the software developing process, ensures the reliability of the software developing process, improves the efficiency of the software development, the quality of software, and the applicability of final flow of software development.
Owner:SUN YAT SEN UNIV
RUP quantitative process cutting method aiming at software features
InactiveCN101620534BGuaranteed reliabilityImprove development efficiencySpecific program execution arrangementsSoftware development processSoftware engineering
The invention discloses an RUP quantitative process cutting method aiming at software features, comprising RUP flow expansion cutting and RUP expansion flow applicability cutting; the RUP flow expansion cutting aims at requirements of software features and carries out expansion to an RUP standard flow, so as to form special working flow for ensuring the requirements of a certain software feature;the RUP expansion flow applicability cutting leads the expanded special working flow to carry out process applicability cutting relevant to elements and element attribute, the corresponding cutting operations are ensured by the relevancy degree among all the elements, and the relevancy degree among all the elements is ensured by a fuzzy comprehensive evaluation method. The method solves the problems that the cost is overhigh and the development is overloading in the software developing process, ensures the reliability of the software developing process, improves the efficiency of the softwaredevelopment, the quality of software, and the applicability of final flow of software development.
Owner:SUN YAT SEN UNIV
Real-time tracking and identifying method and device for steel coil loosening
ActiveCN113269043AAutomate monitoringRealize intelligent decision-makingKernel methodsCharacter and pattern recognitionAutomatic routingStructural engineering
The invention discloses a real-time tracking and identifying method and device for steel coil loosening, and the method comprises the steps: adjusting a camera, and completing the setting of the initial position and the initial size of a hoisted steel coil; obtaining a video frame, and calculating the accurate position of the hoisted steel coil according to the initial position information; calculating the accurate size of the hoisted steel coil in the video frame according to the initial size information; intercepting a hoisting steel coil image in the video frame according to the accurate position and the accurate size of the hoisting steel coil, and identifying the hoisting steel coil; whether the coil is loosened or not is judged according to an identification result; judging whether the camera needs to be adjusted; acquiring a new video frame, tracking the hoisted steel coil, and determining the accurate position and the accurate size of the hoisted steel coil; and sending out a steel coil loosening alarm, and sending parking instruction information to a sharing area of the crown block hoisting system. According to the method, manual routing inspection and decision making are replaced, the steel coils hoisted by the crown block are tracked in real time, the steel coils with scattered outer sides and scattered strip steel heads are automatically recognized, and therefore automatic routing inspection and intelligent decision making of the steel coils in the storehouse in the hoisting process are achieved.
Owner:UNIV OF SCI & TECH LIAONING
Motor fixing support and air conditioner
The invention provides a motor fixing support and an air conditioner. The motor fixing support comprises a fixing sleeve and a connecting plate. A containing space in sleeve connection with at least part of a shell of a motor is formed in the fixing sleeve, and a mounting port is formed in the fixing sleeve and used for inserting the motor. The connecting plate is arranged on the fixing sleeve andconnected with an air duct. Firstly, the fixing sleeve is in sleeve connection with at least part of the shell of the motor, so that the motor is inserted into the containing space through the mounting port. Secondly, the connecting plate is connected with the air duct, so that the motor can be fixedly mounted. By the aid of the technical scheme, the fixing mode of pressing the motor by an uppercover and a lower cover is directly cancelled, materials are simplified, and assembly effects are improved. Besides, the motor is fixed onto the motor fixing support in advance, automatic assembly anddistribution can be realized by the aid of a tooling, and operation efficiency is improved.
Owner:GREE ELECTRIC APPLIANCES WUHAN +1
Food forming method provided with split-type forming mold
ActiveCN110897181ASimple structureAchieve streamliningFood coatingFood shapingMechanical engineeringFood processing
The invention belongs to the technical field of food processing, and particularly relates to a food forming method provided with a split type forming mold. The method comprises the following steps: flour wrappers are treated by a flour supply device, the flour wrappers fall on a forming mold located at a flour supply station, the forming mold moves to a stuffing injection station, a stuffing injection device injects stuffing into the flour wrappers of the forming mold, the forming mold continues to move to a forming station, a first forming plate on a forming device and a second forming plateon the forming mold are in press fit with each other, the flour wrappers and the stuffing are subjecting wrapping and forming, the forming mold moves to a push-out station, and a push-out device pushes out formed foods. According to the application, the forming action connection relationship of dumpling foods in the forming process is changed, and the forming mold is split into independent structures, so that the simplification of a complete machine equipment track is realized, a large amount of space can be saved, and meanwhile, the condition that the push-out device is easily interfered or damaged by forming plates on the forming mold in a push-out procedure is also avoided.
Owner:CHENGDU SOONTRUE MECHANICAL EQUIP CO LTD
Authority scheduling method and device
Owner:CHINA MOBILE GROUP JIANGSU
A Neural Network Relation Classification Method and Its Realization System Fused with Distinguishing Degree Information
ActiveCN109344244BWork lessAvoid interferenceSpecial data processing applicationsText database clustering/classificationRelation classificationFeature vector
The invention relates to a neural network relation classification method and an implementation system thereof, which fuse discrimination degree information, comprising: (1) data preprocessing; (2) training word vector; (3) extracting position vector; Combining the word vector of each word with the high-dimensional position feature vector, the joint feature is obtained. (4) computing the semantic representation of sentences; The context information and semantic information of entities are encoded by bi-directional LSTM. (5) calculating a discrimination vector; After the two entity vectors are subtracted from each other and cascaded with a Bi-LSTM unit encode that input; (6) inputting the outputs of (4) and (5) to the CNN, outputting the feature vector fused with the discrimination information, and inputting the feature vector to the classifier for classification; (5) adopting a Loss function to train the model. The invention does not need to manually extract any features, and the jointmodel does not need to preprocess the data by means of other natural language processing tools, the algorithm is simple and clear, and the effect is best at present.
Owner:SHANDONG UNIV
A method for early warning of hidden troubles in train traction system
InactiveCN104236933BObserve the degree of deteriorationReflect the development trend of hidden troublesRailway vehicle testingFeature vectorTraction system
The invention relates to a method for early warning of potential failures of a train traction system, comprising the following steps: 1) collecting real-time operating data of the train traction system through sensors, and performing multi-sensor fusion to obtain an operating data set; After preprocessing the data, extract its hidden danger feature value; 3) Combine the hidden danger feature value into a hidden danger feature vector to represent the running state of the traction system; 4) Process the hidden danger feature vector through the fuzzy algorithm to obtain the change trend of the hidden trouble; 5) Use charts or curves to output changing trends, and carry out early warning of hidden troubles. Compared with the prior art, the invention has simplified data and high accuracy, and can effectively improve the driving safety of the train.
Owner:TONGJI UNIV
N600 one-point grounding test device and application thereof
InactiveCN106054014AEasy to carry outEliminate potential safety hazardsElectrical testingCurrent meterElectrical resistance and conductance
The invention relates to an N600 one-point grounding test device and an application thereof. The test device comprises a voltage meter and a resistor unit which are mutually arranged in parallel, wherein the resistor unit comprises a current meter, a first resistor, a second resistor and a third resistor which are sequentially connected in series, the first resistor, the second resistor and the third resistor are respectively connected in parallel with a relay K1, a relay K2 and a relay K3, and two ends of the voltage meter are test ends. The test device is advantaged in that the test device is convenient to use on site, same type work is simplified, N600 one-point grounding test work of a transformer station is convenient to develop, work efficiency is improved, and hidden safety trouble of incorrect protection motion caused by N600 multi-point grounding of the transformer station in operation can be effectively eliminated.
Owner:国网山东省电力公司聊城供电公司 +1
Knowledge distillation optimization rnn short-term power outage prediction method, storage medium and equipment
ActiveCN110232203BSimplify the complexity of parameter adjustmentImprove operational efficiencyDesign optimisation/simulationNeural architecturesData predictionComputer science
This disclosure provides a knowledge distillation optimized RNN short-term power outage prediction method, storage medium and equipment to obtain features with a high degree of correlation with power outages as initial fault features, divide the fault data into a linear main part and a nonlinear main part, and use ARIMA The algorithm realizes the short-term data prediction of the linear main part, and realizes the prediction of the non-linear main part by RNN. Both of them are used as the input of softmax, and finally the regional short-term power outage prediction value is given, so as to realize the prediction accuracy under the premise of ensuring the prediction accuracy. The simplification and compression of the RNN model improves the running speed and parameter tuning performance of the model.
Owner:SHANDONG UNIV
Who we serve
- R&D Engineer
- R&D Manager
- IP Professional
Why Patsnap Eureka
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com