166results about How to "Meet the needs of engineering applications" patented technology

The invention provides a wind channel model tail support rod structure capable of actively damping vibration by utilizing a piezoelectric stack. The front end of the tail support rod is used for mounting a model internally provided with a strain meter balance, and the rear end of the tail support rod is fixed on a support seat by a locking nut; the tail support rod is divided into a front tail support rod section and a rear tail support rod section, which are connected by a hole shaft in a match mode; a pair of symmetrically connected flange plates in the connecting part are positioned at the front and rear tail support rod sections respectively; and four positioning slots, mutually forming an included angle of 90 degrees, are arranged on the two flange plates and are provided with the piezoelectric stack. Compared with the traditional passive vibration damping structure, the wind channel model tail support rod structure capable of actively damping vibration has the advantages of high frequency and fast responding as adopting the intelligent piezoelectric material as a vibration damping element, can effectively control the vibration, has no effect on the flow field of a test model as the whole vibration damping part is small in size, is simple in structure and design and suitable for the vibration damping of tail support rod structures in different sizes, and has better usability and maintainability.

The invention discloses a continuous rigid frame bridge prestress damage identification method based on deflection monitoring. The continuous rigid frame bridge prestress damage identification method comprises the following steps: I, establishing a bridge deflection monitoring system, that is, the bridge deflection monitoring system comprises n deflection monitoring devices and one data acquisition device, and the n deflection monitoring devices are respectively arranged on n deflection monitoring points; II, monitoring deflection of a bridge, namely, respectively monitoring deflection data of n deflection monitoring points on a main beam by using the bridge deflection monitoring system in real time, and synchronously transmitting the monitored deflection monitoring data to data processing equipment; III, identifying prestress damage, namely, establishing a finite element model of the bridge, processing the bridge deflection monitoring data, acquiring deflection data caused by prestress damage, establishing a damage identification rigidity matrix and identifying prestress damage. The continuous rigid frame bridge prestress damage identification method is simple in step, reasonable in design, convenient to realize, good in use effect, capable of easily and conveniently completing the prestress damage identification process of a continuous rigid frame bridge, and relatively high in identification result reliability.

The invention discloses a commutation failure judging and simulation method for a direct current transmission system based on an electromechanical transient state. The commutation failure judging method comprises the steps: establishing an alternative current-direct current electrical power system electromechanical transient simulation model of direct current transmission, calculating alternative current three-phase / single-phase short-circuit fault happening in a certain point of an alternating system, comparing and analyzing changes of bus three-phase voltage effective value on a failure instant high-voltage direct current inversion side to judge whether commutation occurs in the direct current system or not, and conducting commutation judgment to each point of the alternating current system, and calculating and confirming an alternating current fault area. The simulation method comprises the step of simulating direct current commutation failure by artificially modifying direct-current power according to the judging results of the direct current commutation failure. According to the commutation failure judging and simulation method for the direct current transmission system based on the electromechanical transient state, an electromechanical transient simulation program in which an electric power system takes the leading application part is adopted, the commutation failure judging and simulation method has strong adaptability and practical applicability and is simple and useful and conserved in the aspect of computational analysis of the interaction effect between the direct current and alternating current, meanwhile the simulation method is effective and feasible, and fulfills engineering application requirements effectively.

The invention relates to a multi-source solid waste-based grouting cementing material as well as a preparation method and application thereof. The grouting cementing material is prepared from the following raw materials in parts by weight: 60 to 80 parts of red mud, 60 to 80 parts of other solid wastes, 40 to 60 parts of an excitant and 10 to 20 parts of an additive, wherein the other solid wastescomprise the following raw materials in parts by weight: 15-25 parts of furnace slag, 15-25 parts of mineral slag, 10-20 parts of steel slag and 8-20 parts of desulfurized gypsum. The preparation method is simple; the prepared grouting cementing material is short in setting time, high in early strength, high in stone body mechanical strength, excellent in erosion resistance and suitable for coastal karst broken rock mass reinforcement and treatment engineering, engineering safety construction in coastal areas can be guaranteed, the solid waste stockpiling pressure is relieved, the problem ofenvironmental pollution caused by red mud is solved, high added value is achieved, and comprehensive utilization of the red mud is promoted.

The invention relates to a smart satellite earth observation time window generating method based space analytic geometry, comprising the following six steps of: (1) establishing a space analytic geometry relational model, and using the model for carrying out coordinate transformation on the given substellar point and satellite coordinate; (2) initialization: pretreating the substellar point and target data, and forming substellar point target pair sequence; (3) angle inversion: carrying out angle inversion on each substellar point target in the substellar point target pair sequence, and obtaining a roll sequence and a pitch sequence corresponding to the substellar point target pair sequence; (4) rotation matrix inverse transformation: converting the roll sequence and the pitch sequence which are obtained in the step (3) into dynamic angle under a dynamic rotation coordinate system by inverse transformation; (5) obtaining smart satellite observation angle curve for observing the target by calculation of interpolation; and (6) intercepting a time window. The smart satellite earth observation time window generating method has the advantages of simple and convenient operation, high accuracy, easy implementation and the like, and can be applied under the background of rotation with three degrees of freedom of a smart satellite.

The invention discloses a quick calculation method of super-low-altitude target and land-sea rough surface composite scattering. The method comprises the following steps: acquiring a central position of an irradiated area of a land-sea rough surface; acquiring back scattering RCS (Radar Cross-Section) value of a plurality of target mirror image strong scattering points by the high-frequency asymptotic method; calculating the scattering contribution of the target related to the land-sea rough surface mirror; arranging the observing position to the level under the target irradiated area; calculating the double-station scattering RCS value of a plurality of target double-station strong scattering points; calculating the scattering contribution of the land-sea rough surface related to the local target mirror image while calculating vertical incidence according to the kirchhoff approximation principle; combining and calculating the main composite scattering contribution of a super-low-altitude target and the land-sea rough surface to obtain the composite scattering RCS value. According to the method, the strong scattering points are introduced into the composite multi-scattering calculation of a near field target and the rough surface, and thereby, the calculation amount of the coupling scattering of the target and a land-sea surface is decreased; meanwhile, the application is not limited to a flat land-sea surface, and highly undulated land-sea rough surfaces also apply.

The invention relates to a through-hole connection structure and connection method of a titanium aluminum supercharger turbine impeller and a rotating shaft. The connection structure comprises the rotating shaft, the titanium aluminum turbine impeller and a shaft end locking nut. A through hole is formed in the center of the titanium aluminum turbine impeller. A thread is arranged at the end portion of the rotating shaft. Seen from the end of a turbine, the rotating direction of the thread is the same as the rotating direction of the turbine rotating shaft during working. The lower portion of the thread is the rotating shaft connection portion. The rotating shaft connection portion penetrates through the through hole of the turbine impeller. The rotating shaft connection portion and the turbine impeller through hole are in transition fit or interference fit. The shaft end locking nut is installed on the thread of the end portion of the rotating shaft. As for the connection problem of the titanium aluminum turbine impeller and the rotating shaft, the through-hole connection and threaded locking structure are adopted, the durable and reliable connection between the titanium aluminum turbine impeller and the 42CrMo rotating shaft can be achieved, the requirements for the connection strength of the supercharger turbine impeller and the rotating shaft are met, and meanwhile the connection structure has the advantages of being simple in structure, easy to machine and the like.

The invention discloses a calibration method of an optical inertial measurement unit with a redundant oblique shaft of a carrier rocket. The calibration method comprises the following steps: step one, leveling of a marble slab: placing a level gauge on the marble slab, and repeatedly adjusting the levelness of the marble slab until meeting the requirements; step two, electrifying and preheating of an inertial measurement unit: after installing the inertial measurement unit in a regular hexahedral tool, connecting a testing cable for preheating; step three, calibrating a scale factor and an installation error of a gyroscope through a velocity method; step four, calibrating zero bias of the gyroscope and a scale factor, an installation error and zero bias of an accelerometer through a position method; step five, data processing and analysis: solving the scale factors, the installation errors and the zero bias of the gyroscope and the accelerometer through methods of least square fit and the like.

The invention provides a camera high-precision calibration method based on an optimal polarization angle. The method aims at solving the problem that high-brightness area image texture details and color information in a calibration image are lost to influence camera calibration precision. According to reflection characteristics of a highlight area, a camera calibration model for two-dimensional pixel coordinates in checkerboard image corners is established by applying a principle that polarization imaging in physics can weaken highlight and Stokes variable linear feedback. The requirement on the camera calibration precision under the natural illumination condition is combined; and obtaining checkerboard calibration images under the optimal polarization angles of different poses in the space by utilizing the mathematical model, accurately positioning the images by utilizing a sub-pixel level detection algorithm and a Gaussian fitting method, and extracting a two-dimensional pixel calibration result of angular points in the checkerboard images. The camera calibration method based on the optimal polarization angle not only meets the high-precision requirement of the camera, but also eliminates a high-brightness area of an image part, can be applied to the fields of stereoscopic vision measurement systems, industrial manufacturing technologies and the like needing high-precision camera calibration under the condition of natural light, and has a good application prospect.

The present invention relates to a two-step integral inverse method for determining a geoidal surface based on band-limited aerial vector gravity. The method includes the following steps that: based on a generalized horizontal boundary value theory, band-limited aerial vector gravity at the altitude of a flight line is adopted to calculate band-limited aeronautical disturbing potential; and an inverse Poisson integral model is adopted to extend downwards the band-limited aeronautical disturbing potential to a sea surface, so that band-limited sea surface disturbing potential can be obtained, and the band-limited sea surface disturbing potential is converted into the geoidal surface through using the Bruns formula. The method of the invention is reasonable in design. With the method adopted, a new solution is provided for calculating the geoidal surface based on the band-limited aerial vector gravity, a calculation function for determining the geoidal surface by using the band-limited aerial vector gravity is realized, and the accuracy of the obtained geoidal surface can satisfy engineering application requirements.

The invention discloses a method for actively controlling a friction coefficient of a metal friction pair in aqueous solution of a surfactant. A metal side of a sliding friction pair is provided with a working electrode; an inert material with proper area is arranged nearby the friction pair to serve as an auxiliary electrode; the required friction coefficient is controlled actively and acquired by applying a corresponding control electric potential to the sliding friction pair; the sliding friction pair consists of a metal material and a ceramic or metal material; and a lubricant is the aqueous solution of the surfactant. The range and the size of a voltage applied in the method for actively controlling the friction coefficient of the metal friction pair in the aqueous solution of the surfactant can be determined according to a relationship between an electrode electric potential of the working electrode and a bath voltage. The method for actively controlling the friction coefficient of the metal friction pair in the aqueous solution of the surfactant is widely applicable to a friction-driven and friction-braked device and provides technical guarantee for the reliability and the controllability of the device.

The invention relates to a bi-arc self-locking interference threaded connection method and structure for a titanium aluminum turbine and a spindle. By designing the size, the interference magnitude and assembly technological parameters of a bi-arc self-locking interference threaded connection structure for a titanium aluminum supercharger turbine impeller and a 42 CrMo spindle, machining is performed on turbine impeller internal bi-arc threads and a matching end surface as well as spindle external bi-arc threads and a matching end surface. A turbine impeller heating measure is taken, assembly for the titanium aluminum turbine impeller and the 42 CrMo spindle is achieved through thread screwing, and machining is performed on other positions of the connected titanium aluminum spindle. By means of the connection method and structure, lasting reliable connection between the titanium aluminum turbine impeller and the 42 CrMo alloy spindle is achieved, connection strength requirements are met, the structure and the technology are simple, and the engineering application is facilitated.

The invention discloses a method and device for planning a distributed type power supply system. The planning method comprises the steps of establishing a resource layer model of the distributed type power supply system, wherein control variables of the resource layer model comprise a first vector for representing a system network topology and feeder linetype and a second vector for representing the installing positions of renewable energy generation units in the system and the configuration number, an object function of the resource layer model is min F=(DC+OC) / H, wherein DC represents hidden carbon emission, OC represents power generation carbon emission, and H represents a power consumption prediction value; establishing an operation layer model of the distributed type power supply system, control variables of the operation layer model comprise a third vector for representing renewable energy generation unit call power of the system during the running and calling electric energy from a superior power grid, wherein an objection function of the operation layer model is min f=OC / epsilon, wherein epsilon is a variation factor; adopting a predetermined algorithm to solve the resource layer model, obtaining optimal solutions of the first vector and the second vector, and adopting the optimal solutions to serve as a planning scheme for the distributed type power supply system. The method can effectively reduce carbon footprint of terminal power utilization.

The invention discloses a method for preparing an anisotropic nanocrystalline rare-earth permanent magnet block material, which includes: firstly, preparing an alloy ingot by the vacuum smelting process; secondly, melting the alloy ingot and injecting an external direct-current electric field to the melt to realize direct-current inoculation process; thirdly, preparing an anisotropic nanocrystalline rare-earth permanent magnet material by the rapid copper mould solidification process under action of an external direct-current magnetic field; and finally processing the material in heat properly according to specific conditions. The anisotropic nanocrystalline rare-earth permanent magnet block material prepared by the method has the advantages of evident magnetic anisotropy, high density and fine comprehensive permanent magnetic performance. The method is simple in procedure, convenient in operation and low in processing consumption.

The invention discloses a construction hoist lift car people counting method based on multi-information fusion. The method includes the steps that first, GRB color feature threshold values of a safety helmet are acquired; second, segmentation and binarization processing are performed on an image; third, background interference is removed; fourth, mathematical morphology processing is performed on the image; fifth, multiple geometric features in the image are extracted; sixth, the number of people is recognized in a multi-information fusion mode. The method effectively solves the problems caused when the number of people in a construction hoist is counted through an image processing method, is high in algorithm recognition accuracy and good in instantaneity and can meet the requirement of engineering application.

The invention provides a method and device for identifying script files, and relates to the technical field of network security, and the method comprises the steps: obtaining a plurality of sample script files marked with labels; converting each sample script file into a machine instruction sequence; extracting a feature word combination from the machine instruction sequence of the sample script file to obtain a first feature set of the sample script file; for the first feature set, according to a preset word frequency feature algorithm, respectively calculating word frequency features of eachfeature word combination in the first feature set to obtain feature vectors of the sample script file; training a script recognition model according to the feature vector of each sample script file and the label of each sample script file; and when the to-be-identified script file is obtained, identifying the to-be-identified script file by utilizing the script identification model, and determining whether the to-be-identified script file is a malicious script file or not. By adopting the method and the device, the accuracy of identifying the webshell file can be improved.

The invention discloses an on-line monitoring method for a short-circuit fault in a battery energy storage system caused by a low-temperature working condition, and the method comprises the steps: firstly building a lithium battery internal short-circuit equivalent model to simulate the transient response of voltage along with the change of a current load during the internal short circuit of a lithium battery; secondly, analyzing typical processes and interference factors of lithium battery internal short circuit faults under a low-temperature working condition based on a lithium battery chemical principle, calculating correlation coefficients among single batteries, and determining internal short circuit curve characteristic values caused by the low-temperature working condition; and finally, counting all internal short-circuit frequencies based on the battery correlation coefficient characteristic values, and performing safety state evaluation on the internal short circuit of the lithium battery. According to the invention, based on the internal short circuit waveform characteristics of the single cell, the precision is improved by using the moving window filter and the active noise reduction method, and the online monitoring of the low-temperature internal short circuit of the lithium battery is effectively realized.

The invention provides a deep convolutional network for human face identification in a complex scene and a learning method. The network structurally consists of four full convolutional modules and three full connection modules; and parameters of the deep convolutional network are optimized by utilizing a stochastic gradient descent algorithm. The designed deep convolutional network for human face identification is expected to have stable identification performance under multiple view angles for complex illumination in a real application scene; and engineering application demands are met based on relatively high calculation degree under the support of a high-performance GPU.

The invention relates to the technical field of space-air-ground-sea integrated communication, in particular to a low-orbit satellite low-signal-to-noise-ratio high-dynamic burst signal carrier synchronization method. According to the method, burst signal Doppler frequency offset, Doppler frequency offset change rate and carrier phase parameter estimation and correction are realized step by step by adopting a forward structure. The parameter estimation precision of Doppler frequency offset and frequency offset change rate is improved by combining FFT with a quadratic inner difference algorithm; and the phase parameter estimation value of the outburst signal is obtained by using a linear interpolation mode, the estimation precision can obtain higher carrier synchronization under the conditions of high dynamic and low signal-to-noise ratio of a low-orbit satellite, the operation complexity is moderate, and the requirement of engineering application can be met.

The invention discloses a method for identifying and optimizing a fragile line of an electric-gas coupling system. The method includes inputting original data, establishing an electric-gas coupling system optimization model, and processing to obtain initial operation state of coupling system; turning off the output power of each gas source of the natural gas network, calculating the gas source vulnerability, and carrying out weight processing to obtain a line coupling vulnerability factor; extracting a topological structure of the electric network and processing to obtain a line topological vulnerability factor; respectively disconnecting each power grid line to obtain a power flow distribution entropy so as to obtain a line operation vulnerability factor; comprehensively obtaining the comprehensive vulnerability of the lines, sorting according to the descending order, and the lines larger than a set threshold value are the fragile lines; and after the fragile line is identified, the transmission capacity of the fragile lines is improved through the thickened wire. According to invention, fragile line in electricity-gas coupling system can be rapidly identified, monitoring and protection of fragile line are enhanced, and probability of large-area power failure of power system can be effectively prevented.

The invention relates to a protective device of a building, in particular to a protective device for reducing the external vibration effects. An energy-dissipating and shock-absorbing supporting device comprises a quadrilateral framework and a damper arranged in the framework, the damper is connected with the framework through a supporting member, the framework comprises an upper beam, a lower beam, a left supporting leg and a right supporting leg, and the framework is of a trapezoidal structure. The energy-dissipating and shock-absorbing supporting device has convenient installation and dismounting, strong energy-dissipating capacity and wide shockproof and quake-proof application fields.

The invention discloses a vulnerability evaluation and improvement method for an electricity-gas comprehensive energy system. The method comprises steps of establishing a network topology structure ofthe electricity-gas comprehensive energy system, according to the given initial fault proportion, simulating the initial fault of the system by utilizing a Monte Carlo algorithm to obtain an initialrunning state of the system, establishing a fault propagation model of the electricity-gas comprehensive energy system to obtain a fault propagation seepage equation, continuously iterating the seepage equation according to the fault propagation termination condition until the fault propagation termination condition is met to obtain a final running state of the system, and establishing a vulnerability parameter of the electricity-gas comprehensive energy system, calculating to obtain a vulnerability result of the electricity-gas comprehensive energy system, and improving the vulnerability result. According to the method, the vulnerability result of the electricity-gas comprehensive energy system can be quickly evaluated, the weak link of the electricity-gas comprehensive energy system is searched, and the propagation of the fault in the system is inhibited by taking measures, so that the vulnerability of the electricity-gas comprehensive energy system is effectively improved, and the probability of large-area collapse of the system is reduced.

The invention discloses a task-distributing and path-planning method applied to multilayer SOM heterogeneous welding robots and belongs to the field of control of robots. The implementation method comprises the following steps: building a first layer of SOM neural network model, training the first layer of SOM neural network model and completing coarse distribution of a heterogeneous welding robotsystem, outputting task points to be welded of various types of heterogeneous welding robots, and improving robustness and universality of the heterogeneous welding robot system; planning paths of asecond layer of parallel SOM network heterogeneous welding robots, enabling multiple heterogeneous welding robots to reasonably complete tasks of the system in a self-organizing manner, simultaneouslyconsidering welding sequence constraint and welding robot interference constraint, achieving optimal path cost and outputting task-distributing and path-planning results; and enabling various heterogeneous welding robots to perform welding tasks according to the output task-distributing and path-planning results, improving the efficiency of the heterogeneous welding robot system and meeting the requirements of engineering application.

The invention provides an online monitoring system and method of an exhaust steam enthalpy value of a low pressure cylinder of a steam turbine. The system is provided with sensors at the positions corresponding to a steam pipeline of a condenser, an axis heater, and four groups of low pressure heaters where steam is exhausted by the low pressure cylinder. As monitoring for flow of circulating water and temperature of inlet water and outlet water is added, the heat of the exhaust steam of the low pressure cylinder can be calculated indirectly through a cold end system; when a heat balance equation is built for a condenser of a steam turbine thermodynamic system to calculate the exhaust steam heat of the low pressure cylinder, the heat which is taken away by the circulating water system andcan be measured directly is introduced to the balance equation, building of the heat balance equation is greatly simplified and accumulated error caused by a traditional calculation method is avoided.The monitoring system only monitors parameters of a low pressure heater system and parameters of a cold end system, calculation parameters are reduced to the maximum extent, and the calculation precision of the exhaust steam enthalpy value of the low pressure cylinder can be improved efficiently.

The invention discloses a distributed real-time control building intercom system, which comprises a main network and a sub-network. Each intelligent node of a controller area network (CAN) bus in the system is configured with the sub-network inside a building; the sub-network adopts a recommended standard 485 (RS-485) bus for networking; the intelligent node of the CAN bus is used for data exchange between the sub-network inside the building and the main network of the CAN bus outside the building and for switching between audio and video channels simultaneously; the intelligent node of the CAN bus adopts a microprocessor, namely a digital signal processor (DSP), into which an enhanced controller area network (ECAN) module is embedded; the ECAN module receives and sends information based on a mailbox; and each message can be configured for sending or receiving the mailbox. Based on a grading distributed control principle, the distributed real-time control building intercom system combines the high speed of the DSP with the high reliability, real time and maintainability of CAN bus communication effectively, so that the system is reliable in operation and good in anti-interference performance and can meet requirements of engineering application completely.

The invention relates to the field of all-vanadium redox flow batteries, and specifically discloses an all-vanadium redox flow battery SOC detection method based on thermochemistry measurement. For the requirements on the real-time detection of SOC (State-of-charge) state of positive and negative electrodes in the using process of an all-vanadium redox flow battery, the specific heat capacity anddensity of positive and negative electrode electrolytes, the inner resistances of positive and negative electrodes and the enthalpy changes of positive and negative electrode reactions at different temperatures are measured in an offline manner, energy conservation equations are established for the positive and negative electrodes inside an electric pile by using the first law of thermodynamics, the inlet and outlet electrolyte temperature changes of the positive and negative electrodes of the electric pile are associated with the heat changes of reactions, the inlet and outlet electrolyte temperature changes of the positive and negative electrodes of the electric pile, reaction current and electrolyte flow rate are measured in real time in an online manner, the heat changes of the positive and negative electrode reactions inside a sampling time period are calculated in real time, and the substance concentration change rule of adjacent sampling points is calculated and obtained in combination with the enthalpy change values of the positive and negative electrode reactions at specific reactions obtained in the offline manner, so that real-time detection for SOC is realized.

The invention provides a method and system for calculating the virtual flow of an underwater acquisition system of an offshore gas field group. The method includes the steps that operating data of at least two gas wells in the gas field group are acquired; the operating data are filtered, and the coefficient of dispersion of the operating data is acquired; a flow control equation is established according to a multi-phase flow physical transportation rule, the flow serves as an unknown value, and the flow control equation is converted into a flow calculation model; when the coefficient of dispersion of the operating data is higher than a preset threshold value, flow data of all the gas wells in the operating data are calculated through the flow calculation model; and when the difference between the total flow value of the flow data of all the gas wells and a measured flow value is smaller than a preset threshold value, yield data of the to-be-measured gas wells are acquired by calculation through the flow calculation model.

The invention discloses a preparation method for a CrN coating on a surface of a nuclear zirconium alloy and a product. The sputtering power, the matrix temperature and the bias voltage are adjusted by the multi-target magnetron sputtering technology; and the CrN coating is prepared on the surface of the Zr alloy matrix. According to the invention, the energy of deposited particles is increased byadjusting the bias voltage of the matrix, so the bombardment effect of the deposited particles on the surface of the Zr alloy matrix is enhanced, the bonding at the interface between the CrN coatingand the Zr alloy matrix is promoted, and then the bonding force of the CrN coating and the Zr alloy matrix is increased; the CrN coating with uniform compactness, excellent bonding performance and high thickness and deposited on the surface of the Zr alloy matrix is obtained; the CrN coating is of important scientific significance and engineering application value to a protective coating preparedon the surface of the zirconium alloy and meeting the engineering application thickness (above 10 microns).

The invention discloses an adjustable micro distributing pulverized coal distributor. The adjustable micro distributing pulverized coal distributor comprises a shell and a distribution grid arranged in the shell. An inlet of the shell is connected with an inlet pipe and an upstream pipeline, an outlet of the shell communicates with outlet branch pipes and is connected with a downstream pipeline, and the part, except the inlet pipe and the outlet branch pipes, of the shell is a closed cavity. A trapezoid distribution grid is adopted as the distribution grid, the top of the trapezoid distribution grid is an arc, and the shell cavity is divided into a plurality of grids through the trapezoid distribution grid. A first flow guiding plate and a second flow guiding plate are arranged in the grids. Openings in the two sides of the grids communicate with the two outlet branch pipes. Openings in the tops of the grids communicate with the two outlet branch pipes. A first adjusting baffle is arranged above outlets in the tops of the grids. The outlet branch pipes are provided with second adjusting baffles, and through the adjustment of the two baffles, the outlets can achieve the relatively uniform distribution effect no matter whether an inlet flow field is uniform or not.