18results about How to "Retain performance" patented technology

The invention discloses a preparation method of a pre-fibration polymer solid electrolyte as well as a product and application of the pre-fibration polymer solid electrolyte. The method comprises the following steps: demulsifying and drying polytetrafluoroethylene (PTFE) emulsion to prepare pre-fiberized PTFE dry powder, then mixing sulfide lithium phosphorus sulfur chlorine (Li6PS5Cl) electrolyte and fiberized PTFE in a proper proportion, and directly preparing the sulfide solid electrolyte composite membrane through a room temperature rolling solvent-free dry process. The pre-fiberized solid sulfide electrolyte film avoids high-cost and high-energy-consumption operations such as ball milling, hot pressing and the like, and the preparation energy consumption and the equipment cost are reduced. The material can be effectively applied to the field of preparation of solid-state batteries, and the structure and performance of an electrolyte material are fully reserved. The method is a solid electrolyte film forming technology with an application prospect, and has an important application value for the development of a solid-state battery.

The invention discloses a deep network model quantization compression method, which relates to the technical field of model quantization and comprises the following steps of: preprocessing a deep network model, compressing the deep network model by combining model quantization and pruning, performing low-rank decomposition on the deep network model and verifying a compression quantization effect of the deep network model; according to the method, a series of preprocessing operations are carried out on the deep network model to be processed, the deep network model can be ensured to be smoothly carried out in a subsequent quantization process, the performance of the model is reserved as far as possible, and redundant parameters in the deep network model can be deleted by carrying out model pruning on the deep network model in combination with model quantization. According to the method, the complexity of the model can be reduced, the precision of the model can be kept, floating-point number parameters in the model are converted into low-precision fixed-point numbers, and the storage and calculation cost of the model can be reduced, so that the deep network model is effectively compressed and quantized, and the deep network model is lighter.

The invention relates to the field of data forgetting, in particular to a data forgetting method based on error maximization noise and neuron pruning, which comprises the following steps: generating an error maximization noise matrix; initializing the model parameters according to the connection sensitivity of each model parameter so as to induce the model to be forgotten; updating the target forgotten data set and the reserved data set; periodically detecting the performance change amplitude, and determining whether to perform adaptive adjustment for the data forgetting process or not according to the performance change amplitude; if self-adaptive adjustment is carried out on the data forgetting process parameters, iterative updating is periodically carried out on the noise intensity and the pruning scale according to the performance change amplitude; and when the forgetting effect is met or the preset number of iterations is reached, iteration updating for the forgetting process parameters is stopped, and the overall performance of the model is reserved to the greatest extent while the data forgetting effect is ensured.

The invention discloses a CNN image classification model similarity calculation method based on multi-dimensional feature fingerprints, and the method comprises the steps: extracting static parameter fingerprint features (weight matrixes and bias vectors) and dynamic output fingerprint features (convolution kernel output, convolution layer output and model output) of an original CNN image classification model and a pirated CNN image classification model; a unique CNN image classification model fingerprint is formed, and a CNN image classification model similarity judgment mechanism is provided according to the fingerprint and can be used for CNN image classification model traceability comparison analysis. According to the method, the fingerprints of the CNN image classification model can be extracted through a CNN image classification model fingerprint extraction technology, and whether the pirated CNN image classification model plagiarizes the original CNN image classification model or not is judged based on the extracted fingerprints, so that a reliable method is provided for intellectual property protection of the CNN image classification model.

The invention discloses a chitosan nanofiber-based hemostatic sponge and a preparation method thereof, and belongs to the field of dressing preparation. The chitosan nanofiber-based hemostatic sponge is mainly prepared from chitosan nanofibers, alginate, amorphous calcium salt and a cross-linking agent. The chitosan nanofiber is used as a main framework of the hemostatic sponge, and blood cells and platelets can be conveniently adsorbed due to a large number of amino groups and high porosity; due to the existence of alginate, blood clots can be conveniently formed after the hemostasis sponge absorbs blood; the existence of calcium promotes the initiation of a blood coagulation cascade reaction and synergistically accelerates the hemostasis effect. The three components are combined to complement each other, so that the hemostatic sponge has more excellent hemostatic performance, and an excellent blood coagulation and hemostasis effect is achieved after all aspects are integrated.

The invention relates to the field of gas detection, and discloses a high-sensitivity decomposition gas sensor for an air switch cabinet, which takes a Co3O4 / Cr2O3 / YSZ ternary composite material as a sensitive electrode, and is prepared by the following method: step (1): ball-milling yttrium oxide stabilized zirconia powder, tabletting and then firing a solid electrolyte; (2) coating two fine Pt welding wires on the surface of the YSZ solid electrolyte, and calcining the YSZ solid electrolyte to prepare a sensor substrate; and (3) weighing Co3O4, Cr2O3 and YSZ powder, adding absolute ethyl alcohol and zirconium oxide balls for wet ball milling, naturally drying after ball milling, mixing with an organic carrier, performing silk-screen printing, coating on a sensor substrate, and calcining to form a sensitive electrode layer. The sensor adopts the ternary composite material as the sensitive electrode, the performance of each component is retained, the detected NO is as low as 10 * 10 <-6 >, and the reproducibility is good.

The invention discloses a machine learning data forgetting method based on Hebb plasticity inspiration. The method comprises the following steps: dividing a given data set into a reservation set and a forgetting set according to a forgetting task; establishing a first recursive lateral connection and a second recursive lateral connection between layers of the deep network model, which are respectively used for extracting feature subspaces of a given data set and a forgetting set; based on the feature subspaces of the given data set and the forgotten set, obtaining a core gradient space of the reserved set; constructing an orthogonal projection matrix of the core gradient space; forward propagation is carried out on samples in the forgotten set, and an error signal is generated through a loss function; performing back propagation on the samples in the forgotten set, and correcting the weight gradient of the deep network model based on the error signal and the orthogonal projection matrix; and updating the weight of the deep network model by using the corrected weight gradient to complete the forgetting task. By means of the method, forgetting of specific data can be efficiently and accurately achieved under the condition that the model is not retrained.

This application relates to the field of solar cell technology and discloses a method for recycling the bottom cell in a perovskite tandem solar cell. The perovskite tandem solar cell includes a top cell, an inorganic functional layer, and the bottom cell stacked sequentially. The top cell includes a perovskite light-absorbing layer disposed on the side of the inorganic functional layer facing away from the bottom cell. The inorganic functional layer and the perovskite light-absorbing layer have different coefficients of thermal expansion. The recycling method includes the following steps: cooling the perovskite tandem solar cell; and heating the cooled perovskite tandem solar cell to peel the perovskite light-absorbing layer off from the inorganic functional layer. Using the recycling method provided in this application, the performance of the bottom cell can be effectively preserved, ensuring its stability during secondary use.

The invention discloses a hybrid expert model pruning method and device and a medium, and the method comprises the steps: collecting a calibration data set through a predetermined data collection module of a hybrid expert model, so as to obtain performance data, and the performance data comprises a gating value and an activation vector; calculating according to the performance data to obtain a significance score corresponding to the hybrid expert model, determining experts needing to be removed according to the significance score, and pruning the hybrid expert model according to the experts needing to be removed; determining a reserved expert according to the pruned hybrid expert model, reconstructing the reserved expert to determine a new expert list, determining a new router, and replacing the hybrid expert model according to the expert list and the router; and storing the replaced hybrid expert model as a new model file, and carrying out iterative pruning on the model file so as to carry out fine tuning on the hybrid expert model. The device is accurate in evaluation, good in performance maintenance, stable in structure and flexible and efficient in pruning mode.

The application discloses a phase change microcapsule composite heat storage material with photocatalytic characteristics and a preparation method thereof. The phase change microcapsule composite heat storage material has a whole core-shell structure, paraffin as a core, crystalline TiO2 with photocatalytic characteristics as a shell wall coated on the surface of the paraffin, and nano-titanium nitride and carbon nanotube composite blended photothermal conversion particles as core-shell modification materials uniformly attached to the surface of the TiO2. The phase change microcapsule composite heat storage material has a stable core-shell structure, full-spectrum solar photothermal conversion capacity, and effectively improves the coating rate, the thermal conductivity and the photocatalytic efficiency, retains the thermal properties of paraffin and inhibits leakage and corrosion.

The invention belongs to the technical field of superconducting, and relates to an insulation enhanced superconducting cable and a preparation method thereof. The method comprises the following steps: firstly, reinforcing a superconducting wire rod through a NiCoCr / Cu composite matrix, and carrying out multi-channel coating and on-line monitoring by adopting BN modified polyimide paint to form a uniform insulating layer; and then wrapping and hot pressing are carried out by using a B-stage epoxy-impregnated aromatic polyamide tape to construct an insulation enhanced structure. Stranded cable forming is carried out through tension control and a numerical control positioning die, and the composite cable is formed by combining the core wire and the multiple layers of insulation reinforced wires. Finally, through secondary wrapping and pretension braiding protection, the superconducting cable with the coverage area larger than or equal to 99.9% and the void ratio smaller than 1% is obtained, the withstand voltage between strands is larger than 5000V, the withstand voltage between cables is larger than 7000V, the mechanical stability and the thermal management performance are remarkably improved, and the superconducting cable is suitable for being applied to high-field magnets and fast pulses.

The application relates to the technical field of high polymer materials, and particularly discloses a bio-based polyether ether ketone as well as a preparation method and application thereof. The bio-based polyether ether ketone is prepared by the following steps: mixing a bio-based diphenol monomer, alkali, 4,4'-difluorobenzophenone, a water-carrying agent and a polar organic solvent, refluxing the water-carrying agent at 30-150 DEG C, evaporating the water-carrying agent at 40-150 DEG C, performing a final polymerization reaction at 150-260 DEG C, and then sequentially performing precipitation, washing and drying to obtain the bio-based polyether ether ketone. The bio-based polyether ether ketone provided by the application uses renewable biomass as a raw material to construct a core molecular structure, retains the core properties of traditional polyether ether ketone, such as high-temperature resistance and chemical corrosion resistance, and optimizes and improves the solubility and processing fluidity.

This application discloses a composite high-flowability, high-conductivity conductive masterbatch and its preparation method, relating to the technical field of conductive materials. The masterbatch uses a fiber-forming polymer as a matrix, compounded with conductive carbon black, graphene, silver nanosheets, carbon nanotubes, and modified MXene nanosheets, along with coupling agents, dispersants, antioxidants, and lubricants. The modified MXene nanosheets are modified with an amino-functionalized modifier and a polysaccharide mixture of sodium alginate and agar. Amino grafting improves interfacial compatibility, while polysaccharide coating inhibits filler agglomeration and oxidation, optimizing the dispersion of the conductive filler in the matrix. This masterbatch constructs a multi-level continuous conductive network, significantly improving melt flowability, conductive stability, and mechanical properties, effectively solving the problems of easy agglomeration, poor flowability, and insufficient spinnability of traditional conductive masterbatches.

The application discloses a method for prolonging the service life of a centrifugal air compressor impeller, which comprises the following steps: disassembling and cleaning a stainless steel base impeller, placing the stainless steel base impeller and copper alloy powder material in a steel plate mold, placing the steel plate mold in a high-temperature furnace and carrying out high-temperature heating and heat preservation, so that the copper alloy powder material is fully cladded on the surface of the stainless steel base impeller and the cladding strength is ensured, taking out the steel plate mold from the high-temperature furnace and carrying out water cooling, taking out the stainless steel base impeller cladded with the copper alloy powder material from the steel plate mold, and carrying out finish machining according to an impeller drawing, so as to obtain the stainless steel base impeller with the copper alloy coating on the blade tooth top surface. The method clads the copper alloy on the blade tooth top surface of the steel structure base of the impeller, retains the performance of the stainless steel base impeller, increases the wear resistance and corrosion resistance of the surface copper alloy, significantly prolongs the service life of the impeller, and reduces the cost of spare parts and maintenance.

The invention belongs to the technical field of isotope labeling, provides a 13C labeled carbon nanotube and a preparation method thereof, and solves the problem of difficult tracing of carbon nanotubes. The method comprises the following steps: mixing 13C powder, graphite powder and a catalyst under protective gas, and discharging to obtain carbon ash; and sequentially refluxing, dissolving and filtering the carbon ash to obtain the 13C-labeled carbon nanotube. According to the invention, the carbon nanotube is prepared from a solid material, the labeled carbon nanotube of stable isotope 13C substitution skeleton atoms is obtained through an arc discharge method, the method is comprehensive and efficient, and the original morphological structure and performance of the carbon nanotube can be retained. 13C is randomly distributed on the skeleton of the carbon nano tube, so that the reaction change of the carbon nano tube in an organism is effectively traced. The preparation method provided by the invention is simple and low in cost, and brings a new direction for in-vivo tracing of the carbon nanotube.

The invention relates to a method for decolorizing waste cotton-polyester blended fabric through low-temperature plasma synergistic chemical oxidation. Through the synergistic effect of low-temperature plasma and an acidic chemical oxidant (composed of inorganic acid, hydrogen peroxide, a penetrant and a hydrogen peroxide stabilizer), color development and auxochrome groups of dye on the waste polyester-cotton blended fabric are effectively destroyed, and efficient decolorization of the waste polyester-cotton blended fabric is achieved. According to the technical scheme, the decolorizing efficiency is improved, the original performance of polyester and cotton fibers can be reserved, the secondary utilization performance of the recycled waste polyester-cotton mixed textile fabric is guaranteed, and an effective means is provided for resource utilization of the waste polyester-cotton mixed textile fabric. Compared with a traditional chemical oxidation technology, the method has the advantages that the required acid amount and treatment time are remarkably reduced, the temperature required by the reaction is greatly reduced, the environmental protection performance is remarkably improved, the energy consumption is reduced, and an economic and green new way is provided for recycling the waste polyester-cotton mixed textiles.

The invention discloses a flame-retardant heat-conducting organic silicon pouring sealant and a preparation method thereof, and relates to the technical field of pouring sealants, the flame-retardant heat-conducting organic silicon pouring sealant comprises a resin main body and a hardening composition which are matched with each other, the resin main body at least comprises a base adhesive, a flame retardant and a heat-conducting filler, and the hardening composition comprises a marginal hardening agent and an internal curing agent. The acid ester content of the marginal hardening agent is larger than that of the internal curing agent, and a first mixture formed by the marginal hardening agent and the resin body is used for making contact with the potting side boundary and surrounding a second mixture formed by the internal curing agent and the resin body. According to the invention, the potting effect of the organic silicon potting adhesive can be improved.

The invention relates to the technical field of high polymer materials, and particularly discloses bio-based polyether-ether-ketone as well as a preparation method and application thereof. The preparation method of the bio-based polyether-ether-ketone provided by the invention comprises the following steps: mixing a bio-based diphenol monomer, alkali, 4, 4 '-difluorobenzophenone, a water-carrying agent and a polar organic solvent, refluxing at 30-150 DEG C to carry out water carrying, steaming out the water-carrying agent at 40-150 DEG C, carrying out final polymerization reaction at 150-260 DEG C, and sequentially precipitating, washing and drying to obtain the bio-based polyether-ether-ketone. The bio-based polyether-ether-ketone is obtained. According to the bio-based polyether-ether-ketone provided by the invention, a core molecular structure is constructed by taking renewable biomass as a raw material, the core performance of high temperature resistance and chemical corrosion resistance of traditional polyether-ether-ketone is reserved, and the dissolving property and the processing fluidity are optimized and improved.