11results about How to "Improve your own performance" patented technology

The invention provides a code large model self-evolution method based on Monte Carlo tree search, which comprises the following steps: S1, inputting a code generation task into a pre-training reasoning model, and generating a reasoning step according to a reasoning task; s2, constructing an inference tree, and obtaining an optimal inference path and a reward score according to an improved Monte Carlo tree search algorithm; s3, according to the optimal reasoning path, based on the automatically generated reasoning and reward data, performing pre-training model self-evolution to improve the reasoning model; and S4, enabling the pre-training model to generate a code of the reasoning scheme, and training the pre-training model to obtain a code large model. According to the method, under the condition that a large amount of fine adjustment data is not needed, the capacity and accuracy of large model code generation can be effectively improved, and the adaptability of the model in a complex reasoning task is further enhanced.

The invention belongs to the technical field of ceramic preparation, and particularly relates to a method for preparing a high-performance ceramic rod by isostatic pressing, which comprises the following steps of: performing isostatic pressing treatment on polyvinyl alcohol serving as an adhesive, aluminum isopropoxide serving as a filling agent and raw materials to form a primary green body; and performing surface spraying on the basis of permeation, hydrolysis and sintering of the primary green body to obtain a to-be-sintered green body, and performing sintering and surface polishing to obtain the ceramic rod. The defect that a green body is damaged due to existing isostatic pressing treatment is overcome, the hard damage of an isostatic pressing mold is effectively reduced by combining the adhesiveness of polyvinyl alcohol and a flexible silica gel material, meanwhile, the performance of a secondary isostatic pressing green body is improved by means of sintering curing pretreatment, and the service life of the secondary isostatic pressing mold is prolonged. Therefore, the overall performance and the yield of the ceramic rod are ensured.

The present application relates to the field of computer vision, and specifically discloses a DHA algal oil soft capsule filling amount insufficient detection method based on a deep learning neural network, comprising: accurately separating each capsule instance from a production line transmission image through a lightweight instance segmentation model; synchronously analyzing the global contour and local density features of a single capsule region using a multi-scale feature extraction network, which innovatively introduces an auxiliary density estimation module to enhance the perception of the content filling state; and fusing the above features to complete the accurate classification of the filling state. The present application effectively solves the practical problems of feature confusion, insensitivity to minor changes, poor dense detection effect, and the difficulty in balancing model efficiency and accuracy, and realizes high-precision and high-efficiency automatic detection of DHA algal oil soft capsule filling amount insufficient.

The invention discloses an energy storage load distribution method, system and device and a computer readable storage medium, and relates to the technical field of energy storage systems. Dividing the energy storage unit into energy storage areas based on the power grid data set; for each energy storage region, based on the power grid data set, constructing a local load distribution game model taking the energy storage unit as a game main body; based on the power grid data set, constructing a global load distribution game model taking the energy storage area as a game main body; calculating the local load distribution game model and the global load distribution game model through a reinforcement learning algorithm to generate a target configuration strategy; and performing load configuration on the energy storage unit according to the target configuration strategy. The dynamic real-time load distribution between the energy storage units and the energy storage areas is realized, the optimization cooperation inside the energy storage units and the power balance between the energy storage areas are also realized, finally, the accurate global load balance and the optimal resource distribution are realized, and the energy efficiency of the energy storage system is improved.

The invention provides a display panel and a display device.The display panel is provided with a first area and a second area arranged around the first area, the display panel comprises an array substrate, a light-emitting functional layer, a thin film packaging layer, a first conductor layer and a light shielding layer, the light-emitting functional layer is arranged on one side of the array substrate, and the thin film packaging layer is arranged on the other side of the array substrate; the light-emitting function layer comprises a light-emitting structure located in the first region, the thin film packaging layer is located on the side, away from the array substrate, of the light-emitting function layer, and the thin film packaging layer comprises an organic packaging layer and a first inorganic packaging layer located on the side, away from the light-emitting function layer, of the organic packaging layer; the first inorganic packaging layer covers at least part of the structure of the array substrate in the second region. The first conductor layer is arranged on the side, away from the light-emitting functional layer, of the first inorganic packaging layer and is attached to the first inorganic packaging layer. The shading layer is arranged on the side, away from the array substrate, of the first conductor layer, and the projection of the first conductor layer in the thickness direction of the display panel is located in the projection of the shading layer in the thickness direction.

The invention provides a molded case circuit breaker and a molded case circuit breaker electrical life prediction method, and the method comprises the steps: sampling the instantaneous voltage and instantaneous current of the input end and the output end of each phase when the circuit breaker acts for a single time, calculating the arc power according to the instantaneous voltage and the instantaneous current, and predicting the electrical life of the molded case circuit breaker. Calculating a difference value between the input end arc power and the output end arc power which are in the same phase, defining the difference value as instantaneous arc power, and defining the instantaneous arc power as electric wear loss; carrying out accumulation calculation on the instantaneous arc power after each action of the circuit breaker to obtain accumulated arc energy, and calculating the residual electric life of each phase of contact according to the accumulated arc energy; and comparing the residual electric service life of each phase of contact, and selecting the residual electric service life with the minimum value as the residual electric service life of the molded case circuit breaker. The accumulated arc energy is calculated by utilizing the arc power generated when the contact system of each phase of the circuit breaker is disconnected, and the service life of the circuit breaker is calculated according to the electric abrasion loss, so that guidance is provided for overhaul and maintenance time of the contact of the circuit breaker.

The invention discloses a reference electrode battery and a preparation method thereof, the method comprises the following steps: S1, a double-sided lithium pole piece and a single-sided graphite electrode lamination form a soft package battery, the double-sided lithium pole piece is located in the middle, the two sides of the double-sided lithium pole piece are separated by a diaphragm, a copper / nickel pole piece and a graphite pole piece are respectively placed, and a unit battery can be formed; s2, carrying out mixing treatment on the unit battery and an electrolyte to obtain a mixed battery; s3, carrying out electroplating treatment and charging treatment on the mixed battery to obtain a charged battery; s4, disassembling the charged battery to obtain a lithium-plated copper sheet / lithium-plated nickel sheet and a lithium-embedded graphite sheet; and S5, implanting the lithium-plated copper sheet / lithium-plated nickel sheet and / or lithium-embedded graphite sheet into a battery cell. The method is easy to operate, and the problems that the reference electrode is low in manufacturing success rate, difficult in standardization process, limited in use scene and poor in effect are solved.

The utility model provides a D-column structure and a non-bearing type automobile, the D-column structure comprises a D-column inner plate assembly and a D-column outer plate assembly, and further comprises a tail door hinge mounting plate assembly and a floor rear cross beam assembly; a tail door hinge upper supporting plate, a tail door hinge lower supporting plate and a D column reinforcing plate are sequentially arranged on the inner side of the tail door hinge mounting plate assembly from top to bottom, and the lower portion of the D column reinforcing plate extends to the floor rear cross beam assembly. The force transmission path of the D-column structure is improved by arranging the D-column reinforcing plate, matched lap joint between the D-column reinforcing plate and the D-column inner plate is improved by arranging the hinge upper supporting plate and the hinge lower supporting plate, the sinking rigidity of the tail door is improved, the sinking phenomenon of the tail door is relieved, and therefore the problem that the tail door cannot be normally opened due to sinking of the tail door is effectively solved. And even the problems of failure and water leakage of a tail door sealing system, loose tail door state, poor automobile appearance and the like are indirectly caused.

The invention belongs to the field of aluminum processing and manufacturing, and relates to a 7020 aluminum alloy profile and an extrusion process thereof. The aluminum alloy profile has the following component contents in percentage: Si≤0.15, Mg1.15-1.25, Fe≤0.25, Cu≤0.15, Mn≤0.25, Ti≤0.05, Zn: 4.80-5.00, Zr: 0.10-0.16, Pb: —, Cr: —, a single impurity is ≤0.05, a total of ≤0.15, and the rest is Al. The extrusion process comprises material inspection, heating treatment, extrusion profile, stretching and straightening, finished product sawing and artificial aging, wherein the artificial aging method is to first heat the sawed 7020 aluminum alloy profile to 105±3°C in an aging furnace, and keep the temperature for 7-8h, and then heat it to 135±3°C, and keep the temperature for 6-8h. The 7020 aluminum alloy profile prepared by the 7020 aluminum alloy profile extrusion process provided by the present invention has an elongation of 70-80 MPa higher than the standard value, and a tensile strength of 40-50 MPa higher than the standard value, has excellent product performance, and a yield rate of more than 75%.

The invention discloses a thermal power plant equipment operation optimization technology based on intelligent diagnosis, relates to the technical field of thermal power plant equipment operation, maintenance and optimization technologies, and aims to solve the problems that when a current intelligent operation and maintenance system applied to thermal power plant equipment processes complex data, an effective feature extraction step is often lacked, and the operation efficiency is high. The method comprises the following steps: S1, a real-time monitoring and data acquisition process: deploying a high-precision sensor on key equipment of the thermal power plant, acquiring equipment operation data in real time, and transmitting the data to a data processing center in real time in combination with a data acquisition system supporting remote communication; according to the method, an advanced feature extraction technology is introduced, and an autonomous learning mechanism is combined, so that the system can continuously learn and optimize own performance, and powerful support is provided for intelligent operation, maintenance and optimization of thermal power plant equipment.

The cylindrical battery structure comprises a shell, a BMS (Battery Management System) protection module, a roll core and a battery cathode, a mounting cavity is formed in the shell, a battery anode is arranged at one end of the shell, an opening is formed in the other end of the shell, a contact is arranged at the bottom of the BMS protection module, an anode bulge is arranged at the top of the BMS protection module, and a cathode side edge is arranged on the outer ring of the BMS protection module; one end of the roll core is provided with a positive electrode end, the other end of the roll core is provided with a negative electrode end, the roll core is arranged in the mounting cavity of the shell, the positive electrode end is in contact with the contact for welding, and the negative electrode of the battery is arranged at the opening of the shell; the utility model has the advantage of effectively solving the problems of poor positioning precision, limited sealing structure, easy electrolyte overflow to damage electric equipment, difficult accurate control of electrode connection, easy poor contact, complex positive electrode connection, reduced production efficiency and increased short circuit risk due to the adoption of a positive electrode part assembly mode in the traditional process.