60results about How to "Fix performance degradation" patented technology

The invention discloses channel calibration method and device, wherein the method comprises the following steps of: dispatching the time slot of a preset frame; estimating an upper radio frequency channel and a lower radio frequency channel of equipment in the time slot, obtaining the gain and signal delay parameters of a sending part and a receiving part of the upper radio frequency channel and the lower radio frequency channel of the equipment, and computing the calibration compensation data of the equipment; and calibrating by applying the calibration compensation data in a weight value mode. The invention solves the problem of TDD-BF performance deterioration caused by the asymmetry or nonreciprocity of an upstream channel and a downstream channel of a TDD-BF, improves the calibration reliability and ensures the high-quality data transmission of beam shaping.

The invention discloses a rotating machinery fault diagnosis method based on a one-dimensional depth residual convolution neural network. Firstly, the network learns more deep and abstract fault characteristics of training samples through a stacked one-dimensional residual module. Then, the Adam optimization algorithm is used to optimize all the super-parameters to complete the deep-level featureextraction and fault classification, and the rotating machinery fault diagnosis system model based on one-dimensional depth residual convolution neural network is obtained. Finally, the test samples are inputted into the trained fault diagnosis model, and the deep-seated features are extracted automatically to diagnose the health status of the rotating machinery.

The invention provides a test device and detection method for detecting comprehensive performance degradation of a rolling linear guideway pair, and belongs to the technical field of precise numerical control manufacturing. A three-direction force sensor is utilized in the design of the sliding block and guideway loading operation test device; the test device adopts a cylinder to carry out loading on the vertical direction and side direction of the rolling linear guideway pair; and the test device is driven to move through a numerical control machine tool spindle. A sliding block is connected with the sensor and the spindle through specially-designed connecting devices respectively; the numerical control system is utilized to regularly drive the sliding block to reciprocate linearly; and theoretically, the connection mode can adopt the rigid connection mode, and however, considering a workbench of the numerical control system has a certain parallel misalignment and angle error, steel springs having appropriate elastic coefficients are adopted as buffer elements, which are arranged in the axial direction and side direction of the sliding block respectively. The sliding block can be loaded through descending or side bias amount of the spindle to ensure that loading of the spindle to the sliding block is more stable and has no destructive effect.

The invention discloses a radar radiation source recognizing method based on a VAE-ResNet network. The problems that network training cannot be completed under few signals, the radar radiation sourcefeature extraction is not sufficient and the recognition accuracy is low in the prior art are mainly solved. By means of the scheme, seven different types of radar radiation source signals are generated; the radar signals are output in the form of a sequence, then time-frequency converting is conducted on the signals, and the signals are output in the form of a time-frequency graph; the data output in the two forms is concentratedly marked with the signal types, and a training set and a testing set are prepared; the VAE-ResNet network is established; the VAE-ResNet network is trained through the training set; the testing set is sent into the trained network, and the network output is of a radar radiation source type. The radar radiation source recognition feature extracting completeness degree and recognition precision are improved, and the method can be used for radar signal recognition under a complicated electromagnetic environment.

The present invention discloses a newborn pain expression recognition method and system based on a deep 3D residual network. The method comprises: establishing a newborn expression video library containing pain expression category tags, and dividing samples in the newborn expression video library into a training set and a verification set; constructing a deep 3D residual network for newborn pain expression recognition, pre-training the network by using a publicized large-scale video database with category tags to obtain initial weight parameter values, and then performing fine-tune on the network by using the samples in the training set and in the verification set in the newborn expression video library to obtain a trained network model; and inputting a to-be-tested newborn expression video segment into the trained network model for expression classification recognition, and obtaining a pain expression recognition result. According to the technical scheme of the present invention, a deep 3D residual network is used to extract temporal and spatial dynamic features capable of reflecting time information from the video, and the change of the facial expression can be better characterized, so that the accuracy of the classification recognition can be improved.

The embodiment of the invention discloses a low-voltage differential signaling (LVDS) transmitter, which relates to the technical field of data transmission and solves the problem that the performance of the LVDS transmitter in the prior art is degraded. The transmitter comprises a working power supply, a main transmitter part, a first selective input module and a second selective input module, the main transmitter part comprises a first PMOS (P-channel Metal Oxide Semiconductor) transistor, a second PMOS transistor, a first NMOS (N-channel Metal Oxide Semiconductor) transistor and a second NMOS transistor, and the sources of the first PMOS transistor and the second PMOS transistor are connected with the working power supply; both the first selective input module and the second selective input module are loaded with a first signal, a second signal, first bias voltage and second bias voltage, the first bias voltage or the second bias voltage is selectively inputted into the grids of the first PMOS transistor and the second PMOS transistor according to the first signal and the second signal, the voltage loaded into the grid of the first PMOS transistor is different from the voltage loaded into the grid of the second PMOS transistor in real time, and the first signal and the second signal are a pair of different-mode signals in order to carry out the function of the LVDS transmitter.

The invention discloses non-reactive casting powder for high-aluminum steel. The non-reactive casting powder is composed of, by weight percent, 26%-40% of CaO, 18%-28% of Al2O3, 14%-28% of BaO, 2%-10%of Li2O, 6%-14% of F-, not larger than 2% of MgO, 2%-8% of C and the balance inevitable impurities, and it is kept that (Na2O+K2O+SiO2) in the impurities is smaller than or equal to 2%. The casting powder basically includes no components of SiO2, Na2O, Fe2O3, B2O3 and the like capable of reacting with active elements of [Al], [Ti] and the like in molten steel; after use, the components and performance of the casting powder are still stable; and a series of problems of casting powder performance deterioration, slag entrapment, longitudinal division and the like brought by traditional high-aluminum steel casting powder steel slag reacting are fundamentally solved. According to the casting powder, lubrication and heat transfer of a cast blank can be effectively and coordinately controlled, the high capacity of absorbing Al2O3 inclusions is achieved, and it can be guaranteed that the continuous casting process of the high-aluminum steel is smooth.

The invention relates to the technical field of silicon light, and provides an optical transceiver assembly based on silicon light. The optical transceiver assembly comprises a silicon optical chip element and an electric chip element, and specifically, the silicon optical chip element comprises a silicon optical chip step, a silicon optical modulator, a first detector and a laser driving unit; wherein the silicon optical modulator, the first detector and the laser driving unit are manufactured on the silicon optical chip in a semiconductor processing mode; the electric chip element comprisesa microprocessor, a modulator driving unit, a trans-impedance amplifier and a clock data recovery unit; wherein the microprocessor, the modulator driving unit, the transimpedance amplifier and the clock data recovery unit are integrated on a single electric chip and are arranged on a PCB (Printed Circuit Board). The silicon optical chip modulator and the detector achieve modulation and detection of optical signals respectively, the electric chip integrates functions of a plurality of electronic elements such as TIA, LA, CDR and MZM DRIVER, and the size and the number of packaging parts are simplified.

The invention discloses an accident-tolerant nuclear fuel cladding double-layer coating structure and a preparation method thereof. The double-layer structure composite coating structure design of a diffusion barrier layer bottom layer with a high temperature oxidation resistant top layer can have comprehensive characteristics of both high temperature oxidation corrosion resistance and strong interface bonding; the components of the diffusion barrier layer material comprise Ta, Ti, W, V and Cr; the mass percentages of each element of the diffusion barrier layer are that Ta+Ti+W is greater thanor equal to 70% and less than or equal to 100%, and V+Cr is less than or equal to 30.0%, the material of the anti-oxidation layer is one of FeCrAl, Ti2AlC, TiN, ZrC or SiC, the thickness of the diffusion barrier layer accounts for 5%-40% of the entire coating thickness, the thickness of the high temperature oxidation resistant layer accounts for 60%-95% of the entire coating thickness, the totalthickness of the double-layer coating structure is 5[mu]m-500[mu]m, and the total thickness of the coating structure does not exceed 25% of the thickness of a zirconium alloy substrate.

The invention discloses a flexible display panel and a display device. The flexible display panel comprises a bending region and a metal layer which is correspondingly distributed in the bending region; and the metal layer is provided with a plurality of metal wires, and at least one side edge of each metal wire is provided with a protrusion or a groove. According to the flexible display panel andthe display device, the groove or the protrusion is formed in at least one side of the metal wires, and when the metal wires are bent, the stress concentration phenomenon of the metal wires in the bending region is effectively improved, so that the problems of performance degradation, failure and the like caused by the phenomena such as peeling and the like in bending the metal wires can be solved effectively.

The invention relates to an improved Adaboost-SVM model generation method suitable for fault diagnosis of a wind power converter, and the method comprises the steps: giving the weight of each trainingsample, and updating the weight of each sample according to the classification error rate of the previous round. In the algorithm training process, the weight of the mistakenly classified samples isincreased, and the weight of the correctly classified samples is reduced. Through multiple iterations, a plurality of weighted weak classifiers are combined to finally obtain a strong classifier. Through the design of the improved Adaboost-SVM model, the problem of performance degradation caused by different samples can be solved to a certain extent, the generalization capability of the classifieris improved, and the method has a very good effect in application to fault diagnosis of the converter.

The present invention discloses a method for confirming the actual transmitting power of the transmitter, and the method comprises the following steps: a.calculating the saturated transmitting power Pcsi and other maximum transmitting power; b. confirming the least value in other maximum transmitting power and the saturated transmitting power Pcsi, and setting the least value as the actual transmitting power Ptol of the transmitter; c. confirming the final actual transmitting power Pk of each subchannel of the transmitter according to the confirmed actual transmitting power Ptol. The scheme of the invention facilitates that the transmitter does power distribution according to the actual need of each subchannel. The scheme of the invention settles the problem that the saturated transmitting power exists when the CSIT of the transmitter generates error, and the problem that the integral performance of the system deteriorates causing by the continuous increasing of the transmitting power when the target SNR which is corresponding to the set SNR or the supported frequency spectrum efficiency is larger than the saturated SNR that is caused by the CSIT error and is corresponding to the saturated transmitting power.

The invention discloses an FPGA (a field-programmable gate array) operational circuit based on generalized correlation coefficients. The architecture of the FPGA operational circuit comprises an array multiplication accumulator, a shifting register, a multiplying unit, an assembly line, a square root extractor, a divider and a register. When environmental noise contains an impulse noise component, a matched filter basically loses efficacy while the SR (the Spearman rank correlation coefficients) and the KT (the kendall rank correlation coefficient) contained in the generalized correlation coefficients still can show excellent robustness when interfered by impulse noise; by utilizing the rich logical resources and the high-speed parallel processing capability of the FPGA, the operational circuit achieves real-time and high-speed correlation processing of signals based on the SR, the KT and the PPMCC (the product moment correlation coefficient) and can be applied to radar, a sonar sound ranging system and a GNSS (a global navigation satellite system) for signal capturing, tracking and other signal processing technologies.

The invention relates to a ball mill load parameter soft measuring method and belongs to the technical field of soft measurement. The invention provides a ball mill load parameter soft measuring method based on multi-state transfer learning, comprising: recording a known label of a ball mill, acquiring auxiliary variables under the known and unknown labels, subjecting the acquired auxiliary variables to FFT, combining and modeling a data input Xs and a real-time data input Xt to obtain X, unitizing the label Ys in the source domain to obtain Ysa, subjecting the X and the Ysa to dimension reduction by a joint distribution adaptive method to obtain a new feature representation Z, dividing Z into a source domain feature Zs and a target domain feature Zt, normalizing the label Ys in source domain data to obtain Ysy, training a multitask-LS SVM model with Zs and Ysy, and predicting the Zt using the model. The ball mill load parameter soft measuring method is applied to the field of soft measurement.

The invention provides a nonvolatile memory device and a manufacturing method thereof. The manufacturing method of the nonvolatile memory device comprises the step of: providing a substrate comprising an array region; forming at least two polycrystalline silicon gate stack structures on the array region of the substrate; forming an insulating layer between the at least two polycrystalline silicon gate stack structures, wherein a groove is formed in the insulating layer between the at least two polycrystalline silicon gate stack structures; forming a first metal layer on the side walls and the bottom part of the groove; filling a protective layer in the groove to cover the first metal layer; and forming a second metal layer on the top part surfaces of the at least two polycrystalline silicon gate stack structures and the protective layer, so that the protective layer is surrounded by the first metal layer and the second metal layer. The metal layer provided by the invention can effectively solve the problems of performance degradation and electrical property failure of the nonvolatile memory device.

The invention relates to the field of communication transmission, in particular to a method and a device for alignment of return-to-zero modulating pulse and transmission data. The method comprises the following steps: modulating and outputting a return-to-zero modulating optical signal carrying transmission data; carrying out photoelectric conversion to the return-to-zero modulating optical signal, and then, filtering the obtained return-to-zero modulating electric signal to obtain a frequency component; and adjusting the return-to-zero modulating optical signal in a mode of time delay according to the size of the frequency component. The device comprises a return-to-zero modulating unit for modulating and outputting the return-to-zero modulating optical signal carrying the transmission data, a detecting unit for carrying out the photoelectric conversion to the return-to-zero modulating optical signal and filtering the obtained return-to-zero modulating electric signal to obtain the frequency component, and a control unit for adjusting the return-to-zero modulating optical signal in the mode of time delay according to the size of the frequency component. The invention solves the problem that the prior art can not align the RZ modulating pulse and data to cause the system performance to be poor, and improves the performance of transmission systems.

The invention discloses a voltage controlled oscillation device and method for driving voltage controlled oscillation device, and the device and the method only use simple structures to controle power voltage supplied to the voltage controlled oscillation device accroding to the state of the frequence signal outputted from the voltage controlled oscillation device, thereby the frequency signal is in the nomal range and the change of surrounding environment can be compensated, such as technology bias, voltage, temperature (technology, voltage, temperature change: PVT), etc.

The invention discloses a gas path structure of a solid oxide fuel cell and an adjustment method thereof. The solid oxide fuel cell is formed by multiple cells which are combined in series and parallel and acts as the battery stack of a generating unit. Each cell is provided with a gas duct for transporting gas required for reaction. The gas path structure comprises a cavity, a porous baffle, a gas intake manifold and an gas outlet manifold. The cavity surrounds the battery stack in a sealing way and the cavity has an insulating effect. The gas intake manifold is connected on one side of the cavity. The gas outlet manifold is connected on the other side of the cavity. The porous baffle is arranged in the cavity and close to the gas intake manifold. The porous baffle simply and ingeniouslychanges the pressure field in the cavity of the battery stack, thereby changing the distribution of the gas flow, achieving the goal that the gas flow uniformly flows through the gas duct of each celland providing sufficient fuel gas and air for each cell.

The invention discloses an array antenna adopting an artificial magnetic conductor. According to one embodiment of the invention, the array antenna comprises a dielectric substrate provided with a first layer, a second layer, and a third layer, which are in an overlapped arrangement; a conductor area, which comprises a grounding surface and a feeder, which are disposed on the first layer of the dielectric substrate; a patch antenna, which is disposed on the third layer of the dielectric substrate, and is provided with a plurality of unit antennas having parts connected with the feeder by passing through conductor through holes; and the artificial magnetic conductor, which is disposed on the second layer of the dielectric substrate, and is used to block interferences between the patch chipand the conductor area, and is provided with a plurality of grid structures having shapes arranged at intervals in a staggered manner.The artificial magnetic conductor is configured such that the sizeof a pattern shape adjacent to the conductor via hole among the plurality of pattern shapes is smaller than the size of the remaining pattern shapes.

The embodiment of the invention discloses a memory array operation method and device. The method may comprise the steps of receiving an operation instruction issued by a target business object for a cache of a memory array, wherein service objects supported by the memory array are divided into at least one performance group, and traffic allowed to be operated is pre-calculated for each performance group; selecting a target performance group to which the target business object belongs from the at least one performance group and judging whether the traffic allowed to be operated of the target performance group has residual traffic or not; if the traffic allowed to be operated of the target performance group has the residual traffic, responding to the operation instruction; and if the traffic allowed to be operated of the target performance group has no residual traffic, refusing the operation instruction. According to the embodiment of the method and the device, the performance deterioration problem of integrated memory can be solved.

The invention provides a self-adaptive control method for characteristic aging of a device. The method comprises that a representative value of the present characteristic aging state of the device of an integrated circuit is obtained in real time; a calibration control signal value is determined according to the representative value of the characteristic aging state of the device; and the calibration control signal value is used to control a calibrated power voltage output by a voltage regulator, and the calibrated power voltage serves as a present working voltage of the integrated circuit. The problem that aging of an MOSFET device causes performance degeneration of the circuit is solved effectively, and the service life of an integrated circuit chip is prolonged.

The invention provides a self-adaptive control method for characteristic aging of a device. The method comprises that a representative value of the present characteristic aging state of the device of an integrated circuit is obtained in real time; a calibration control signal value is determined according to the representative value of the characteristic aging state of the device; and the calibration control signal value is used to control a calibrated power voltage output by a voltage regulator, and the calibrated power voltage serves as a present working voltage of the integrated circuit. The problem that aging of an MOSFET device causes performance degeneration of the circuit is solved effectively, and the service life of an integrated circuit chip is prolonged.

The invention discloses a method for pre-processing a substrate in a carbon-based semiconductor device manufacturing technology. The method comprises the steps that the washed substrate is placed in a drying oven, and drying processing is carried out on the substrate at 80 DEG C-200 DEG C for 5-60 minutes; BCB solutions are prepared, and are diluted through organic solutions, and the diluted BCB solutions coat the surface of the substrate in a spinning mode through a spin coater; under the protection of N2 or inert gas, the substrate with the surface coated with the BCB solutions in the spinning mode is heated to be 200 DEG C-400 DEG C, a cross-linking reaction occurs to benzocyclobutene monomers to generate a benzocyclobutene polymer, and therefore a BCB organic film layer is formed on the surface of the substrate. The method solves the problem that due to the fact that the surface of an original substrate is scattering in polarity, rough and fluctuant and impurity absorption occurs, the performance depredation of a grapheme device occurs, the BCB organic film passivates and decorates the surface of the original substrate, and meanwhile carrier mobility degeneration of grapheme cannot be caused.

The invention provides an optical path device of an optical fiber gyroscope and an optical fiber coiling method, the optical path device comprises an optical fiber ring, a waveguide, a coupler, a detector, a light source and a middle hollow annular fiber coiling structure, the optical fiber ring and the annular fiber coiling structure are arranged in parallel; the waveguide, the coupler, the detector and the light source are all arranged at the middle hollow position of the annular fiber coiling structure; and tail fibers among the optical fiber ring, the waveguide, the coupler, the detector and the light source are coiled on the annular fiber coiling structure. Therefore, through the optical path device of the fiber-optic gyroscope and the device tail fiber single-layer symmetric coiling method for improving the optical path symmetry of the fiber-optic gyroscope through optical fiber coiling, the problem of gyroscope performance deterioration caused by improper treatment of the device tail fiber during assembly of the fiber-optic gyroscope is solved.

The invention discloses a low-temperature heating refrigerant flow rate matching control method and an air conditioner. The low-temperature heating refrigerant flow rate matching control method comprises the following steps that S1, the air conditioner enters a heating mode, whether a condition of entering defrosting control is achieved or not is judged, and if the condition is achieved, S2 is executed; S2, the air conditioner performs defrosting operation, when defrosting is finished, an electronic expansion valve recovers an initial opening degree, defrosting timing is started, and after thedefrosting time is achieved with defrosting timing, S3 is executed; S3, valve regulation timing is started; S4, after a regulation time of the expansion valve is achieved with regulation timing, theelectronic expansion valve performs one-step valve closing, and whether the opening degree of the electronic expansion valve is declined to a lower limit opening degree or not is judged, and if the opening degree of the electronic expansion valve is not declined to the lower limit opening degree, S5 is executed; S5, whether the condition of entering defrosting control is achieved or not is judged,and if the condition is achieved, S6 is executed; S6, the opening degree of the electronic expansion valve is kept invariable; and S7, when the condition of entering defrosting control is achieved ornot is judged, the step S2 is executed, or otherwise, the step 6 is executed. By adopting the low-temperature heating refrigerant flow rate matching control method provided by the invention, the refrigerant flow rate in the defrosting process and the heat exchange ability of an evaporator perform dynamic matching, the low-temperature heating ability is improved, the defrosting period is prolonged, the energy consumption is lowered, noise pollution is reduced, and the use comfort is improved.

The invention discloses a multi-source distillation-migration mechanical fault intelligent diagnosis method based on high-order moment matching, and the method comprises the steps: building a multi-source data set through the operation data collected from a plurality of mechanical devices, carrying out the preprocessing, and dividing the multi-source data set into a source domain data set, a target domain training data set, and a target domain test data set; constructing a multi-source distillation-transfer learning network model based on high-order moment matching, and performing high-order moment matching, maximum classifier difference and multi-source distillation training by using the source domain data set and the target domain training data set; and taking the target domain test data set as test input, and synthesizing outputs of the plurality of classifiers by using an adaptive weighting strategy to complete cross-domain fault diagnosis. According to the method, features of a source domain and a target domain are aligned at domain and category levels by utilizing multi-source data, the classification capability of the model on target samples is improved through multi-source distillation, and adaptive weighting is provided to integrate diagnosis results, so that the problem that the performance of a traditional method is reduced in cross-domain diagnosis is solved, and the performance of a deep model is greatly improved.

The invention relates to a power amplifier control method and device, a power amplification system and a storage medium. The method is applied to the power amplification system, and the power amplification system comprises a power amplifier. The method comprises the following steps: when a starting moment arrives, outputting a first driving voltage to the power amplifier, and enabling the power amplifier to enter a working state and work under the driving of the first driving voltage; and under the condition that the control end voltage of the power amplifier reaches the power amplifier working bias voltage, outputting a second driving voltage to the power amplifier, and enabling the power amplifier to work under the driving of the second driving voltage, wherein the first driving voltage is greater than the second driving voltage, and the power amplifier working bias voltage is determined according to the second driving voltage. Even if the power amplifier works at the same temperature, the driving voltage output to the power amplifier can change along with time, so that the starting speed of the power amplifier is increased, and the problem of performance deterioration of the power amplifier is solved.

The invention provides a burst error code detection method based on coding constraints, and belongs to the technical field of channel coding. The method comprises the following steps: before decoding, carrying out hard decision on decoding initial information; calculating a syndrome according to the hard decision result and the LDPC code check matrix, and generating an error code indication sequence according to the syndrome; constructing an initial search window for the generated error code indication sequence, counting an error rate estimation value in a sliding window by adopting a sliding search mode, comparing the error rate estimation value with a preset threshold value, and judging whether continuous burst error codes exist or not; if the continuous burst error code exists, determining the starting position and the ending position of the burst error code by adopting a bidirectional iteration window search method, calculating the length of the burst error code according to the obtained starting position and the ending position of the burst error code, and deleting the corresponding channel information; if the continuous burst error code does not exist, directly decoding and then outputting a result. By adopting the method and the device, the burst error code resistance of the LDPC code can be improved.

The invention discloses an axis alignment method for direct coupling of a polarization maintaining optical fiber ring terminal and an integrated optical chip. Direct coupling of the polarization maintaining optical fiber ring terminal and the integrated optical modulator chip is achieved through two coupling processes of end face observation of the polarization maintaining optical fiber ring terminal and clamping of a coupling clamp. By means of the technical scheme, the problem that the performance of the fiber-optic gyroscope is degraded due to the fact that additional optical nonreciprocalerrors are generated due to the melting point in the fiber-optic gyroscope is solved, the precision and speed of coupling angle error identification are improved, and accurate control over the coupling angle of the polarization axis is achieved.

The present invention discloses a method for confirming the actual transmitting power of the transmitter, and the method comprises the following steps: a calculating the saturated transmitting power Pcsi and other maximum transmitting power; b. confirming the least value in other maximum transmitting power and the saturated transmitting power Pcsi, and setting the least value as the actual transmitting power Ptol of the transmitter; c. confirming the final actual transmitting power Pk of each subchannel of the transmitter according to the confirmed actual transmitting power Ptol. The scheme of the invention facilitates that the transmitter does power distribution according to the actual need of each subchannel. The scheme of the invention settles the problem that the saturated transmitting power exists when the CSIT of the transmitter generates error, and the problem that the integral performance of the system deteriorates causing by the continuous increasing of the transmitting power when the target SNR which is corresponding to the set SNR or the supported frequency spectrum efficiency is larger than the saturated SNR that is caused by the CSIT error and is corresponding to the saturated transmitting power.