70results about How to "Balanced utilization" patented technology

The invention discloses a broadband wide frequency spectrum analysis method and system based on deep learning. The method comprises the steps of spectrum signal acquisition, spectrum preprocessing, abnormal interference detection, spectrum trend prediction and spectrum scheduling management. The invention relates to the technical field of wide-frequency spectrum analysis, in particular to a broadband wide-frequency spectrum analysis method and system based on deep learning, which adopts a three-step method of abnormal interference detection, spectrum trend prediction and spectrum scheduling management, and adopts a layer-by-layer progressive deep learning and reinforcement learning technology. The real-time detection of spectrum abnormity, the accurate prediction of future trend and the optimal allocation of dynamic resources can be realized; carrying out abnormal interference detection by adopting a joint improvement multi-classification support vector machine method combined with feature enhancement; performing spectrum trend prediction by adopting a time sequence integrated network combined with one-dimensional convolution feature extraction; and performing spectrum scheduling management by adopting a reinforcement learning method enhanced by a near-end strategy optimization algorithm.

The invention provides an adaptive retrieval enhanced large language model construction and question answering method and system, a storage medium and a program product, and the construction method comprises the steps: collecting question answering data including questions, standard answers and reference contexts; answering the question by using a large language model to obtain an initial answer; judging whether the initial answer is consistent with a standard answer or not, and constructing training data according to a judgment result; and based on the training data, performing fine tuning training on the large language model to obtain the adaptive retrieval enhanced large language model. The question-answering method comprises the steps of inputting questions; self-adaptively retrieving the enhanced large language model to answer questions, and judging whether extra knowledge needs to be acquired or not; and outputting an answer as a final answer according to a judgment result. According to the method, the balance between resource utilization and answer quality can be realized, and the problems of high resource consumption, low question and answer efficiency and poor answer accuracy are solved.

The invention belongs to the technical field of power distribution networks, and particularly discloses a multi-time scale voltage regulation method and system based on deep reinforcement learning. The invention provides a voltage regulation method for an active power distribution network with a time sequence characteristic and a hybrid equipment combined control capability, which comprises the following steps of: firstly, setting an intelligent agent capable of outputting discrete actions and continuous actions; and hour-level and minute-level dual-time-scale scheduling of an on-load voltage regulator, a capacitor bank and a renewable energy inverter is realized. In addition, the invention further designs a redundant multi-agent cooperative control system, and action output of all agents is coordinated through a multiplication mechanism of transverse interaction. In addition, a priority experience playback technology, a Thompson sampling technology and a security module mechanism are also integrated, and the learning efficiency and the control effect of the intelligent agent are improved.

The invention discloses a host power consumption and temperature balance control method and system, a medium and a program product, and relates to the field of electrical digital data processing.The method comprises the steps that a plurality of heat dissipation air channels are constructed according to air flow data; collecting temperature data in each heat dissipation air duct, and respectively calculating the temperature response rate and the temperature attenuation rate of each heat dissipation air duct; dividing the calculation tasks into a first type of calculation tasks and a second type of calculation tasks based on the temperature response rate; establishing migration rules of the first type of calculation tasks and the second type of calculation tasks according to the temperature attenuation rate; when it is detected that the temperature in any heat dissipation air duct is larger than a temperature upper limit threshold value, migrating a first type of calculation tasks being executed in the heat dissipation air ducts to target heat dissipation air ducts with the temperature lower than a preset temperature threshold value; and migrating the second type of calculation tasks being executed in the target heat dissipation air channel to the heat dissipation air channel. According to the invention, the accuracy of heat transfer and accumulation in the equipment is improved, and the accuracy of temperature control management is further improved.

The invention relates to a multi-task allocation method and device for unmanned transportation equipment, equipment and a medium, and the method comprises the steps: constructing a multi-task combination model, and setting related constraint conditions; obtaining an initial population through chaotic mapping; performing rapid non-dominated sorting on the initial population, and based on a dynamic congestion degree sorting strategy, selecting and determining a plurality of parent populations through a champion; for each parent population, generating a new individual through adaptive crossover operation, and then performing mutation operation through Gaussian mutation to obtain a filial population; combining the parent population and the child population, performing rapid non-dominated sorting operation, and screening excellent individuals through a probability distribution function to obtain a new parent population; performing iterative computation based on the new parent population, and gradually optimizing a result to obtain an optimal multi-task allocation scheme; and performing multi-task allocation on the unmanned transportation equipment through the optimal multi-task allocation scheme. The method can improve the transportation efficiency, and balance the task scheduling complexity and the utilization rate of unmanned transportation equipment.

The invention discloses an Internet of Things HTTP fuzzing method based on a large language model, and relates to the technical field of Internet of Things HTTP fuzzing. The method comprises the following steps: data acquisition: adopting a data acquisition strategy to obtain data for the HTTP fuzzy test of the Internet of Things; data preprocessing: preprocessing the collected data, and providing high-quality input data for subsequent fuzzy testing; hTTP structure sensing guided by LLM: using a prompt adjustment method to label fuzzy variables in request data by using the LLM, understanding a protocol data structure, and guiding a variation process; lLM-guided seed template expansion: expanding a seed bank by enriching data packets and expanding field values; lLM-guided seed variation and seed scheduling: optimizing test results by intelligently selecting and managing input seeds. According to the method, the efficiency and accuracy of fuzzy testing can be improved, and the vulnerability detection capability is enhanced.

The present application discloses a graph database load balancing method and device based on request type, which relates to the field of graph database query technology, and solves the problem in the prior art that several complex queries with long time consumption and heavy load may all fall on the same query service, causing the query service to be under great pressure and slow response, because only focusing on the resource utilization of the server and ignoring the characteristics of the query request itself. The method includes: constructing a mapping table, and according to the complexity of the query request, independently polling and routing the requests for complex queries according to the type, and independently polling and routing the requests for non-complex queries. Since the complex queries are routed to different query services, the effective dispersion of complex query requests of the same type is achieved, and the pressure on each query service to process data is balanced and small, thereby effectively improving the overall query efficiency and reducing the query time.

The invention relates to the technical field of concurrent processing, in particular to a concurrent processing method and system for requests with sequence requirements, equipment and a medium. The method comprises the following steps: receiving a plurality of request tasks with a sequence requirement, grouping the tasks according to a preset condition to form a plurality of request task groups, then adding all the request task groups into a to-be-processed pool, creating a unique group identifier, recording final modification time, forming a group list, and sending the request task groups to the to-be-processed pool. Whether idle slots exist in the concurrent slots or not is continuously checked, when the idle slots are found, the target request task group with the earliest modification time is selected from the group list, the first request task is taken out to serve as the target request task, meanwhile, the target request tasks distributed to the different concurrent slots are processed, and efficient concurrency is achieved; when a certain target request task is executed, releasing the concurrence slot occupied by the target request task; the contradiction between sequence requirements and concurrent processing is solved, balanced utilization of system resources is realized, and the processing efficiency is improved.

The invention provides an adaptive knowledge contrast learning recommendation method based on a dynamic attention network. The method comprises the following steps: step 1, acquiring a user-project interaction graph and a knowledge graph; wherein the user-project interaction graph is used for describing historical interaction records of a user on projects, and the knowledge graph describes a relationship between the projects in a triple form; step 2, constructing an AKCL-DA network framework, and training the AKCL-DA network framework by using the user-project interaction graph and the knowledge graph to obtain a recommendation model; 3, inputting the target user-project interaction diagram into the recommendation model, and generating a recommendation result of the project of the target user; the item recommendation result is an item set with a high matching score output by the recommendation model. According to the method, the AKCL-DA network framework is provided to obtain the recommendation model, and the problem that sparse information and personal interests are not matched with general knowledge in the knowledge graph can be solved.

The utility model relates to the technical field of vehicle motors, in particular to an oil-cooled driving motor cooling system, which comprises a stator core and a stator winding. The stator iron core is provided with an oil inlet section located in the middle, two parallel oil section sections located at the two end sides of the oil inlet section respectively, two end oil collecting sections located on the sides, away from the oil inlet section, of the two parallel oil section sections respectively, and two end oil spraying sections located on the sides, away from the parallel oil section sections, of the two end oil collecting sections respectively. The oil inlet section is provided with an oil inlet channel extending in the circumferential direction, the parallel oil section is provided with a plurality of parallel oil channels at intervals in the circumferential direction, the parallel oil channels are communicated with the oil inlet channel, the end oil collecting section is provided with a plurality of oil collecting channels at intervals in the circumferential direction, and one oil collecting channel corresponds to and is communicated with the parallel oil channels. A plurality of oil outlet channels are formed in the end oil injection section in the circumferential direction at intervals, and one oil collecting channel corresponds to and communicates with at least one oil outlet channel. The structure is compact, cooling uniformity is good, and cooling oil utilization rate is high.

The invention provides a modeling method for a flow model of an industrial dehydrogenation double-ring radial reactor, which comprises the following steps of: deducing the flow model of the industrial dehydrogenation reactor according to a modified momentum equation, an Ergun equation and a continuity equation, and simulating fluid mechanics behaviors of reactant flow in a main runner and a catalyst bed layer; a modeling method for solving the mathematical model of the double-ring radial reactor by adopting a numerical algorithm; a numerical algorithm is adopted to solve a mathematical model of the double-ring radial reactor, accurate mathematical description of fluid mechanics behaviors in the industrial dehydrogenation reactor is achieved, and quantitative reference is provided for optimization selection of the structural size of the industrial double-ring radial reactor on the basis. A newly-built industrial dehydrogenation reactor based on the method can realize more balanced utilization of the catalyst at different parts, so that the treatment capacity of the dehydrogenation reactor and the overall utilization efficiency of the catalyst are improved.

According to the precise pig feed feeding method and device, data of a plurality of feeders are obtained and transmitted to a host for processing in real time, it can be ensured that a feeding plan is generated based on latest accurate data, personal errors are avoided, precise feeding is achieved, the feeders conduct feeding according to the plan, the feeding amount is precisely controlled in combination with liquid level detection and a touch device, and the feeding efficiency is improved. Meanwhile, according to actual demands of pigs, personalized feeding plans are made, appropriate nutrition intake is ensured, excessive or insufficient feeding is avoided, the feed utilization rate is optimized, waste is reduced, breeding benefits are improved, real-time data collection and transmission and automatic generation and issuing of the feeding plans are achieved, management is simplified, manual intervention is reduced, and the feeding efficiency is improved. The efficiency and the quality are improved, the nutrition intake is balanced, the immunity of the pigs is improved, the growth speed and the meat quality are improved by optimizing the feeding plan, and remarkable economic benefits are brought to the breeding industry.

The invention discloses a multi-cluster container scheduling method based on a mirror image layer in an edge computing environment, which comprises the following steps of: establishing a system model of a multi-cluster edge computing environment, and comprehensively describing communication modes inside and outside clusters, resource constraints of nodes, a dependency relationship of a container mirror image layer and the like; analyzing task scheduling requirements and optimization requirements of mirror image layer downloading, establishing a corresponding mathematical model by combining node computing resource limiting conditions, task scheduling requirements and mirror image downloading time delay which need to be met in an edge computing environment, and proposing a method for optimizing the task scheduling requirements and the mirror image downloading time delay by minimizing container starting delay. A scheduling optimization problem aiming at ensuring resource load balancing between clusters is solved; using a heuristic method to quickly screen out candidate edge nodes, and using linear programming to optimize task allocation; calculating the load balance degree between clusters in the edge computing environment, and adjusting the task allocation condition according to the resource load degree of the edge cluster; and for the mirror image layer downloaded by the edge node, the downloading sequence is optimized by using dynamic planning.

The utility model relates to the field of irrigation devices, and particularly provides an agricultural water-saving irrigation device. The device comprises a supporting assembly, a water storage assembly and an adjusting assembly, the supporting assembly comprises a telescopic column, the bottom end of the telescopic column is fixedly connected with a base, the bottom end of a connecting frame is fixedly connected to the top end of a connecting base, two sets of connecting pipes are arranged at the inner end of the connecting frame, and nozzle connecting valves are arranged on the front sides of the connecting pipes. Through cooperation of the supporting assembly and the adjusting assembly, the problems that an existing spray irrigation type irrigation device can provide water irrigation for agricultural crops within the range, the height of the crops can not be changed along with growth of the crops, if the spray irrigation height is not adjusted, irrigation is not uniform enough, and the irrigation efficiency is high are solved. The technical problems that crops are gathered in soil with much water, healthy growth of the crops is affected, water sources are wasted are solved, the overall structure is simple, use is convenient and fast, sprinkling irrigation uniformity is improved, and waste of the water sources is reduced.

The invention discloses a short multimedia message sending method, a short multimedia message sending system, a medium and a product, and relates to the field of information transmission. Selecting short and multimedia message sending channels of which the real-time load rates are lower than a preset load rate threshold value from the plurality of short and multimedia message sending channels as candidate channels; determining a short multimedia message feature data set according to the short multimedia message sending request; determining channel feature data according to the candidate channels; inputting the short multimedia message feature data set and the channel feature data into a preset sending failure probability prediction model to obtain a sending failure probability; when the sending failure probability exceeds a preset probability threshold value, sending the short multimedia message to the receiver through the standby channel; and when the sending failure probability is lower than a preset probability threshold, sending the short multimedia message to the receiver through the candidate channel. By implementing the method, the success rate of sending the short multimedia message can be improved.

The invention discloses a multi-agent implicit collaborative decision-making method based on dark tactical guidance, which is used for solving the problems of unreliable communication and even complete lack of collaboration in a multi-agent system. A'dark tactical 'concept is introduced to guide behaviors of the intelligent agent. The method comprises three core components: a tactical distribution network, an intelligent agent strategy network and a group of dark tactics, and the three components are synchronously optimized in a training stage; during execution, each agent selects a dark tactics which is most adaptive to the current overall and individual states through the tactics distribution network according to information of other agents which are observed by the agent and can be acquired by the agent; the selected dark tactics are then entered into the policy network along with individual observations of the agents to generate action decisions. According to the method, cooperation and competition among the intelligent agents can be considered, the benefit of the overall strategy of the intelligent agent cluster is maximized, and the cooperation efficiency of the multiple intelligent agents in an uncertain environment is improved.

The invention discloses a resource constraint project scheduling method, system, equipment and medium, and belongs to the technical field of workshop intelligent scheduling, the method comprises the following steps: based on a generated process sorting code, by creating a decoding sequence set and a waiting process group, updating the decoding sequence set, determining the decoding sequence set corresponding to the updated process sorting code, and updating the waiting process group; determining the machining starting time, the machining ending time and the total time for completing all the working procedures of each working procedure; and taking the minimum maximum completion time of the worker as an optimization target, obtaining an individual with the highest fitness value through a genetic simulated annealing algorithm, and obtaining the maximum completion time of the worker according to the machining starting time, the machining ending time and the total time for completing all the processes of each process corresponding to the individual with the highest fitness value. And determining the total processing duration coefficient and the latest processing duration coefficient of each worker, and performing shift arrangement on each worker. The method can improve the utilization rate of human resource scheduling.

The invention discloses a method and a system for optimizing an evacuation path of people in a building based on shortest global evacuation time, and relates to the field of optimization of the evacuation path of people in the building. The problems that in the prior art, a path optimization method cannot truly and accurately reflect the influence of crowd crowd on the evacuation process, and the actual condition of crowd evacuation in the crowd state cannot be reasonably considered in the path optimization process are solved. The method comprises the following steps: generating a corresponding evacuation road network model according to an architectural plan of a to-be-analyzed case; determining an evacuation personnel distribution scheme; determining a shortest-path evacuation path scheme, calculating an evacuation simulation result of the scheme, and obtaining global evacuation time and the number of the finally evacuated group; determining an optimizable evacuation group set according to a bidirectional transfer strategy; setting an initial optimization scheme; and iteratively optimizing the evacuation paths of the groups in the first optimization adjustment area and the second optimization adjustment area. And outputting a final optimized evacuation path scheme and a corresponding evacuation process simulation calculation result.

The present invention discloses a method and system for optimizing material allocation strategy, which includes the following steps: collecting and preprocessing data related to material allocation; predicting material demand; formulating inventory optimization strategy; based on the above content, generating an optimal material allocation plan based on the critical path priority scheduling algorithm. The present invention can achieve more accurate demand forecasting, more efficient inventory management and more optimized allocation decision-making.

The utility model provides a modularized fluorine ion exchange column, which adopts a column body structure formed by connecting a plurality of sections of flanges to realize height adjustability and adapt to different fluorine concentration and flow requirements. Spiral flow guide ribs are arranged in the column to guide wastewater to spirally flow, fluid distribution is optimized through centrifugal force, and the fringe effect is reduced; the assembling precision between the columns is ensured through a concave-convex positioning mechanism, and the corrosion resistance is enhanced by matching with a polytetrafluoroethylene lining. The system integrates a water distribution plate and funnel combined structure, realizes flow equalization through groove water distribution holes, eliminates a bias flow phenomenon, and improves the resin utilization rate. And the saturation of the resin can be observed at any time through fluorine ion color-changing exchange resin and an observation window. The resin bracket is of a pull ring type quick release structure, and quick maintenance is achieved through a tap bolt and a sealing ring annular groove. According to the device, through modular combination, flow field optimization and visual monitoring, the fluorine ion deep purification efficiency and the operation convenience are remarkably improved.

The invention discloses a charging pile control method and device and a charging pile, and belongs to the field of new energy charging piles, and the charging pile control method and device comprises the steps that charging demand information of multiple to-be-charged vehicles connected with multiple charging guns is acquired, and the charging demand information at least comprises the battery charge state, the battery capacity and the charging duration preset by a user; according to the charging demand information, the charging priority and the demand power of each vehicle are calculated; generating a power distribution strategy according to the charging priority and the required power, and dynamically distributing a plurality of power modules in the charging pile to a plurality of charging guns; controlling the on-off state of the power distribution matrix according to a power distribution strategy so as to establish electric connection between the power supply module and the charging gun; in the charging process, the actual charging power and the battery state of each vehicle are monitored in real time, the on-off state of the power distribution matrix is dynamically adjusted according to the monitoring result, the power modules are redistributed, and the charging efficiency is maximized on the premise that the requirement for charging multiple vehicles at the same time is met.

This invention discloses a data access control method and related apparatus in a multi-agent collaborative environment, relating to the field of intelligent agents. Responding to a data query request initiated by a first intelligent agent, the method obtains the identity attributes and current task context of the first intelligent agent to determine the visible data range from a hierarchical data repository. Within the visible data range, it retrieves a target data unit matching the data query request and obtains the degree of difference between the first intelligent agent's permission level and the security level label of the target data unit. Based on this difference, it performs differentiated semantic processing on the target data unit and returns the processed target data unit to the first intelligent agent. This invention overcomes the rigidity of static isolation through dynamic permission adjustment and semantic desensitization. By performing field-level or digest-level desensitization based on the difference between permissions and security levels, it enables low-permission nodes to obtain the information required for the task, avoids the waste of non-sensitive knowledge in high-security data, and balances security and knowledge utilization.

This invention discloses a visualized intelligent scheduling system for SMT production lines, relating to the field of precision electronics. It solves the problem of poor scheduling performance in existing intelligent scheduling systems for production lines. The system includes a data acquisition module: analyzing production demand for target orders to be executed, obtaining the quantity of products in the target order, performing historical production quantitative analysis on the production line executing the order based on the quantity of products in the target order, and acquiring production line scheduling data based on the analysis results; a production line matching module: matching production lines for target orders to be executed, obtaining target-suitable production lines; a production analysis module: monitoring the production of sample production units included in the target-suitable production line, obtaining the overall efficiency of the unit equipment based on the monitoring results, and obtaining production scheduling analysis data; and a production scheduling module: performing real-time production scheduling for the product production units on the target-suitable production line. This invention can improve the scheduling efficiency of intelligent scheduling systems for production lines.

The present invention relates to a central tube, which comprises an inner tube (1) and an outer tube (2) which are coaxially sleeved and connected at the bottom, a gap is provided between the inner tube (1) and the outer tube (2), a plurality of filter windows (3) are provided on the side wall of the outer tube (2), the side wall of the inner tube (1) is a continuous structure, and the height of the inner tube (1) is lower than the height of the outer tube (2); the present invention also relates to a metal-free environmentally friendly filter element and a spin-on filter having the central tube. Compared with the prior art, the present invention has the advantages of simple structure, high filtering efficiency, wide application range, safety and reliability, and long service life.

The invention provides a wireless sensor network node configuration optimization method based on a VGPSO algorithm, and relates to the field of wireless sensor networks, and the method comprises the steps: defining the wireless sensor network node configuration optimization as a problem M; the particle swarm optimization is improved through a random Halton sequence, a multi-factor cosine dynamic parameter adjustment strategy, a multi-population multi-strategy evolution mode and Voronoi diagram guidance; and solving the problem M through an improved particle swarm algorithm, and outputting the global optimal position of each sensor node in the wireless sensor network. And a multi-factor cosine dynamic parameter adjustment strategy is introduced to balance global and local optimization searching capabilities and enhance the diversity of the particle population. A Voronoi polygon is used to guide an evolution direction, and a coverage hole direction is searched by calculating a Voronoi centroid so as to optimize coverage vulnerabilities in a population evolution process.

The invention discloses a method for preparing polyethylene by using hexene as a comonomer through a slurry method and high-molecular-weight polyethylene. Monomer ethylene and comonomer 1-hexene are used as raw materials; uniformly mixing a catalyst and supercritical ethane fluid to prepare a catalyst solvent; raw materials, a catalyst solvent, hydrogen and supercritical ethane fluid are continuously added into a first stirring reaction kettle for a prepolymerization reaction, slurry discharged from the first stirring reaction kettle enters a second stirring reaction kettle for a polymerization reaction, and after polymerization is completed, the slurry is subjected to post-treatment to obtain a finished product. According to the method for preparing the polyethylene by the slurry method by taking the hexene as the comonomer, the melt index, the density and the molecular weight of the product can be improved by the supercritical ethane, and the impact resistance of the material is enhanced.

The invention discloses a stereo image super-resolution method for progressive feature extraction based on frequency separation, and particularly relates to the technical field of image super-resolution. The method provided by the invention comprises the following steps: preprocessing a low-resolution image; the input image is sent to a dual-frequency feature extraction module, and the structural and texture dependency relationship of a single-view image is constructed through frequency perception attention; integrating related information of different levels through a progressive three-dimensional feature fusion module, and further guiding the fusion of low-frequency and high-frequency information between the left and right views; and finally, features of different levels of the left view and the right view are cascaded and then sent to a reconstruction part, a final super-resolution image is obtained, and high-frequency and low-frequency information is balanced by using frequency domain loss. According to the method, the quality of the three-dimensional super-resolution image is effectively improved on the basis of ensuring the global consistency.

The invention discloses a system and a method for regenerating carbon black by pyrolysis of waste rubber based on a molten salt liquid phase condition. The system comprises a waste rubber pyrolysis unit, a pyrolysis oil condensation unit and a hydrogen-rich combustible gas collection unit, the waste rubber pyrolysis unit comprises a pyrolyzer, a solar heat collector and a carbon storage tank; the pyrolytic oil condensation unit comprises an air cooler, an oil storage tank and a water cooler; the hydrogen-rich combustible gas collecting unit comprises a compressor and a gas storage tank which are connected in sequence; solar energy is absorbed and utilized for heating, the solar energy conversion efficiency of the adopted flat plate type solar heat collector is high, and the use of fuel oil and natural gas in a traditional heating mode is reduced through the heating mode. By using the mode, the energy conversion efficiency is greatly improved, and the energy consumption and the use of traditional energy are reduced.

The present invention discloses a safe self-driving method for an intelligent guided vehicle and a U-shaped automated container terminal, including laying out a number of magnetic nails in the operation site of the intelligent guided vehicle; writing an absolute position code corresponding to the laying position of each magnetic nail; installing an antenna on the intelligent guided vehicle; demarcating a synchronization area; configuring all intelligent guided vehicles entering the synchronization area to simultaneously start the antenna to charge the passing magnetic nails and read data to obtain the absolute position of the intelligent guided vehicle; planning the driving route of the intelligent guided vehicle according to the absolute position and target position of the intelligent guided vehicle; configuring the intelligent guided vehicle to automatically drive according to the driving route. The present invention configures adjacent intelligent guided vehicles to start measuring positions at the same time, thereby eliminating the problem that the strong electric field generated by the antenna of one intelligent guided vehicle charging the magnetic nails causes the antenna of another intelligent guided vehicle to fail to read the magnetic nail data and cannot be positioned, thereby ensuring the reliability and accuracy of the positioning of the intelligent guided vehicle.

The invention relates to the technical field of industrial electromechanical manufacturing, in particular to a cutting and welding machine tool for industrial electromechanical intelligent machining, which comprises a workbench, a base, a grinding device, a fixing mechanism and a linkage mechanism, the grinding device comprises a grinding mechanism and a first lifting mechanism, the first lifting mechanism comprises a U-shaped rack and a lifting driver, the grinding mechanism is arranged in the U-shaped rack, and the lifting driver is arranged at the bottom of the U-shaped rack; the fixing mechanism comprises a moving structure and a mounting plate, and the mounting plate is arranged at the upper end of the moving structure; the linkage mechanism comprises a second lifting mechanism and two linkage structures, the second lifting mechanism is arranged on the base, the two linkage structures are arranged on the two sides of the U-shaped rack correspondingly, and the linkage structures are connected with the U-shaped rack and the moving structure; the grinding mechanism and the two workpieces are linked through the linkage structure, the jumping phenomenon generated when the grinding disc makes contact with the workpieces is avoided, and therefore the machining stability is enhanced while the grinding face of the grinding disc is evenly utilized.