163results about How to "Increased propulsion speed" patented technology

The invention relates to a super large mining height hydraulic support which comprises an opened rigid integrated base, wherein, two double-telescopic hydraulic upright posts are arranged on the base, and the upper ends of the upright posts are connected with a top beam; the rear of the top beam is connected with a screening beam; an edge protection device is connected with the telescopic beams which are in inner extension connection with the top beam; a balance jack is arranged between the top beam and the screening beam; the screening beam and the base are respectively connected with the front and rear connecting rods; a push jack is arranged on the base and is connected with a push rod; the front part of the top beam is provided with the telescopic beams, the middle part of the top beam is provided with a falling prevention device for the support, and bottom lifting, skid prevention and support regulation devices are arranged on the base; and a levelling device is arranged on a cross binding head on the front end of the push rod; the support adopts 2-meter width between centers, the maximum height ranges from 6.3 meters to 7.5 meters, and the front end of the telescopic beam is provided with a third-stage edge protection device which is not lower than half of the mining height. The support has obvious effects of controlling the wall caving of the coal wall of a super thick coalseam, the roof fall of the ceiling, and the stability of the hydraulic support.

The invention relates to the coal mine safety technology and discloses a gob natural fire control method which has good effect of fire preventing and extinguishing. The gob natural fire control method comprises the following steps: pouring nitrogen, wherein a nitrogen pouring point is 30 to 60 meters from a working face, and the amount of the liquid nitrogen is 2000m3 / h; reducing the air leaking volume of a gob, and reducing the pressure difference of the working face, and controlling the pressure difference of the working face within 120Pa; blocking the leakage within an area in which the tight caving cannot be achieved due to the cantilever beams formed at an air inlet side and an air return side of the gob through a top plates, and timely closing a transverse rail at the gob; accelerating the advance velocity of the working face, wherein the working face is at the speed of 4m / d; increasing the recovery ratio; and controlling the thickness of the residual coal to be below 1.3 meters. According to the gob natural fire control method, a comprehensive working face gob comprehensive fire preventing and extinguishing measure based on 'nitrogen pouring first, leakage blocking, wind resisting, accelerating of advance velocity and recovery rate increasing second' is proposed, and measure implementing parameters are provided; with the adoption of the method, the effect of preventing and extinguishing the fire of the gob can be improved; and the mining on the working face can be ensured to be smoothly carried out.

The invention relates to a large-diameter earth pressure balance shield machine. The machine comprises a cutterhead, a cutterhead drive part, a propulsion oil cylinder, an erector, spiral conveyors and a frame; the cutterhead is a double-cutterhead which comprises a peripheral large cutterhead and an independent central cutterhead; the cutterhead drive part comprises two driving mechanisms which drive the peripheral large cutterhead and the independent central cutterhead respectively; and the peripheral large cutterhead and the independent central cutterhead rotate independently and bidirectionally. Compared with the prior art, the shield machine adopts the double-cutterhead design, greatly optimizes a cutting speed of the cutterhead on a soil body, further improves a propulsion speed of the shield, optimizes the disturbance effect of the cutterhead on the soil body, and prevents the rolling of a shield machine body; different opening rates are designed on the cutterhead and are suitable for different situations of a central part and a peripheral part when a tunnel is advanced; and three spiral conveyors discharge soil, effectively control the posture of the shield, and prevent the deviation of a tunnel axis.

The invention provides an improved genetic algorithm-based method for distributing and optimizing thermal and electrical load of a steam extraction and heating unit, and belongs to the technical field of energy saving and monitoring of a power station. The invention aims to distribute and optimize the thermal and electrical load of a plurality of heating units of the power station, optimally distribute the electrical load and the thermal load of the power station when the electrical load and the thermal load meet requirements of users, and reduce total energy consumption and save energy. An actual heat consumption and consumption difference curve of the unit is set, and a design heat consumption curve of the unit is obtained; the electrical load and steam extraction amount of each unit are acquired; and the electrical load and the steam extraction amount value of each unit are obtained based on the improved genetic algorithm when the total thermal consumption values of the units are minimum, and the electrical load, the steam extraction amount value optimal solution and the corresponding minimum total thermal consumption of each unit are output when the total thermal consumption values of the units are minimum by improving genetic coding and fitness function, and selecting, crossing and mutating the genetic algorithm under the condition that the optimization process meets constraint conditions. By the method, the speed of optimization process and accuracy of an optimization result are improved.

The embodiments of the invention relate to a synchronous location and mapping method. The method comprises the following steps: based on magnitude of interframe motion, selecting a key frame sequence from image frames, based on an interframe matching result of the key frame sequence, introducing the image frames into an attitude graph, and based on the attitude graph, constructing one or more effective detection loops for performing bundle adjustment and correcting the image frames based on the attitude graph after the bundle adjustment. The method can effectively reduce the scale of optimization calculation.

A self-propelled device for locomotion through a lumen, comprising a set of serially arranged inflatable chambers, adjacent chambers being fluidly connected, and a fluid source attached to one end of the set, such that the chambers inflate sequentially. The chambers are constructed of an elastic material and have a wall thickness and dimensions such that they have a characteristic with a non-monotonous relationship between the inflation pressure within the chamber and the chamber's inflated size. The characteristic is such that after an initial inflation pressure peak, the non-monotonous relationship adopts a negative slope, such that the volume of the chamber increases more rapidly than the volume of fluid flowing into it, and the inflation pressure of the chamber falls. This effect causes the chamber to inflate and anchor rapidly, while essentially slowing down the inflation of the succeeding chamber until inflation of the first is complete.

Provided is an image quick defogging optimized method based on a black channel. Edges of an image and non-edge areas of the image are processed through different templates to obtain a transmission picture, and a precise atmosphere light value is obtained by the fact that a sky area or an area with thickest fog is divided. The method greatly reduces consumption, improves the image recovery speed, and improves image recovery effects.

The invention discloses an optimization design method of a low-noise amplifier (LNA) based on a genetic algorithm, and solves the problem of multi-objective optimization in LNA design. Circuit parameters in the LNA such as transistor sizes and passive device values are used as variable quantities, impedance matching of the LNA and a current equation of transistors are used as constraint conditions, a circuit performance evaluation technology based on the equation is used as a circuit performance evaluation method, a parallel genetic algorithm with elitist strategy is used as a global search algorithm, and the gain, noise coefficient and power consumption of the LNA are optimized simultaneously. Through the optimization method, the optimization result of the circuit can be quickly obtained, and the method is extremely suitable for circuit design with particular restriction, performance and function. The method can be used for deep sub-micron radio frequency CMOS (Complementary Metal Oxide Semiconductor) integrated circuit, and is widely applied to electronic systems in the aviation and aerospace fields.

A double-pump hydraulic control loop of a key drilling action of a cutting drill rig comprises a double linkage load-sensitive variable pump set, a double oil intake linkage load-sensitive multi-way valve, a rotary motor, a propulsion motor, a drilling logic control block, a rotation guide handle, a propulsion guide handle, and a hydraulic oil tank. An oil inlet of the double linkage load-sensitive variable pump set is connected with the hydraulic oil tank. An oil outlet and a load-sensitive opening are connected with an oil inlet and a load-sensitive opening of the double oil intake linkage load-sensitive multi-way valve respectively. Each valve pressure difference compensator of the load-sensitive multi-way valve is provided with an external control opening. The external control openings are connected with the drilling logic control block. The rotation guide handle and the propulsion guide handle are connected with the drilling logic control block or a guide control opening of the load-sensitive multi-way valve. The double-pump hydraulic control loop has the advantages that power head rotation-proportion-controlled propulsion is achieved, independent load-sensitive control of rotation and propulsion is achieved, propulsion speed is increased, and the like.

The invention discloses an urban traffic signal control system based on traffic flow prediction, which comprises a data center, a prediction module and an optimization module, wherein the data centerwhich is used for storing the neural network training of a data center and a prediction module and the prediction of traffic data at a next time interval; the data center adds the original real-time traffic data into the historical traffic data before the real-time traffic data is transmitted each time, and then updates the real-time traffic data; the prediction module is used for carrying out short-term traffic flow prediction, for a multi-intersection road section, the to-be-predicted road section is i, assuming that the current moment is in a time interval k, the to-be-predicted amount is the traffic flow of the next time interval, and the prediction basis is the known road section i and the past traffic flow of all the upstream or downstream road sections adjacent to the known road section i; and the optimization module is used for evaluating the congestion degree of the city according to the signal lamp timing scheme in the road and the prediction data output by the prediction module. According to the method, the traffic jam condition of the next time period is obtained in advance, and the signal lamp timing scheme with the minimum jam at the next moment is obtained.

The invention discloses a high-dispersive-distribution nano-TiB2 particle reinforced aluminum matrix composite material and a preparation method thereof. Firstly, a uniform-structure Al-TiB2 intermediate alloy is prepared through an ultrasound assisted mixed salt reaction (K2TiF6 / KBF4-Al), wherein the average particle diameter of in-situ endogenous TiB2 particles is smaller than 100nm; the Al-TiB2 intermediate alloy and Al are used as raw materials, or Al-TiB2 intermediate alloy, Al and alloy elements are used as raw materials, the nano-TiB2 particles are introduced into an aluminum (alloy) matrix through an intermediate alloy dilution method, ultrasound agitation treatment is used as assistance, then pouring into a mold is performed, and ultrasound is applied during solidification (introduction through a bottom introduction method); and thus the high-dispersive-distribution nano-TiB2 particle reinforced aluminum matrix composite material is obtained.

The invention provides an underwater bionic vehicle based on mixed propulsion of a pectoral fin and a propeller, and belongs to the field of bionic underwater vehicles. The underwater bionic vehicle comprises a body part, a pectoral fin propulsion module, a propeller vector propulsion module and a skin shell; a control part of the vehicle is mounted in the body part; the pectoral fin propulsion module is provided with two symmetrical components mounted on two sides of the body part correspondingly; the propeller vector propulsion module is mounted at the tail end of the body part; and the skinshell wraps the outer side. According to the underwater bionic vehicle, the propulsion efficiency and the propulsion speed are obviously improved, and the movement is more flexible; an phase angle and a frequency of each motor are adjusted so that the vehicle can quickly and accurately complete actions such as steering, pitching, swimming back, zero-radius turning and the like; and secondly, thevehicle can complete the pitching and rotating actions under the condition of no motion speed through the cooperative motion of the propeller, and the maximum propulsion speed of the propeller can reach 15 knots to 20 knots, and has a significant improvement compared with the maximum propulsion speed of 2 knots to 3 knots of the vehicle propelled only propelled by the pectoral fin.

The invention relates to an ultrasonic flaw-detecting machine for steel pipes. The ultrasonic flaw-detecting machine for the steel pipes is characterized in that: a flaw detection host machine comprises a box body, and two adjustable three-wheel centering devices are arranged at an inlet and an outlet on two sides of the box body respectively; a hollow spindle is arranged in the box body in a rotatable way, and a measured steel pipe passes through the hollow spindle, the inlet and the outlet; and the middle part of the hollow spindle is provided with at least three probe adjustment devices, an ultrasonic probe is arranged on each probe adjustment device, a power generation device and a signal pre-processing device are arranged outside the hollow spindle, and the signal pre-processing device is connected with the ultrasonic probes, and transmits detection data to an ultrasonic flaw detector of a control system by using a wireless signal. A structural form that the steel pipes are linearly pushed and the ultrasonic probes rotate is adopted, so that the pushing speed of the steel pipes is improved, and the flaw detection efficiency is improved; and the centrifugal force produced by the high-speed rotation of the hollow spindle extrudes the air in a detection cavity, and water is supplied by using an interlayer between an inner sleeve and the hollow spindle, so that the water supply quantity is large, bubbles are reduced, and the detection accuracy is improved.

The present invention relates to the railway traffic and tunnel, in particular to a synchronous construction method used for a shield tunnel. The method has the steps as follows: firstly, square prefabricated components are prefabricated; secondly, the square prefabricated components are arranged orderly at the position with certain distance behind a shield tail in a lining ring along with the pushing of a tunnel shield. The synchronous construction method has the advantages that under the precondition that the tunnel shield pushing speed is not influenced, the construction of the tunnel internal structure is processed at the same time; the square prefabricated components are used for realizing the synchronous construction in the tunnel, thereby not only reducing the cast-in-situ reinforced concrete block in the tunnel, but also shortening the construction time of the tunnel and realizing the high speed of pushing the tunnel shield.

Provided is a cutting blade of a mower. The cutting blade comprises a cutting blade body in the shape of a long strip. Two cutting edges, in proximity to the end portion of the cutting blade body, are arranged in a central symmetry manner. The end portions, opposite to end corners of side edges of the cutting edges, of the cutting blade body bend towards the direction away from the cutting blade body in order to from bending parts. Each bending part has an arc-shaped bending surface and is tangent to the cutting blade body in a transitional mode in order to form a tangent line. The connecting line for the center of the cutting blade body and pointed ends of the cutting edges runs parallel to the tangent line. When the cutting blade body of the mower rotatably works, the bending part drives the air to generate additional turbulent flow moving from the position below the cutting blade body to the upward part such that upper ends of grass mowed rapidly enter a cutting region of the mower under the suction effect of turbulent flow. As a result, the cutting blade of the mower performs cutting operation. Accordingly, grass needed to be cut can be rapidly sucked into the cutting region of the mower in order to finish pruning operation. The advance speed of the mower is improved. The cutting blade of the mower has advantages such as large turbulent flow for suction, smoother cutting and fine quality of broken grass.

The invention discloses a method for accelerating analysis and optimization of a physical layout of an integrated circuit, relating to the technical field of integrated circuits. The method comprises the following steps of: executing region division on the physical layout based on calculation nodes according to input data of the physical layout of the integrated circuit, executing isomorphic analysis on the regions in different calculation nodes, determining the analysis sequences of same regions in each calculation node and among the calculation nodes, analyzing and optimizing the regions of the physical layout according to the sequences, and multiplexing the analyzed and optimized results of the isomorphic regions among the calculation nodes during the analysis and optimization. According to the method for accelerating the analysis and optimization of the physical layout of the integrated circuit disclosed by the invention, the physical layout is divided by utilizing a multi-stage region segmentation algorithm, and the divided regions are analyzed by utilizing an isomorphic analysis method to find out the same regions in the divided regions, so that the posterior regions to be analyzed and optimized in the time sequence are not calculated any more, while the calculation results of the prior regions which are totally same as the posterior regions in the time sequence are multiplexed directly, thus the time is saved and the speed is increased.

The invention discloses a parameter optimization method for a control system based on orthogonal optimization and a particle swarm optimization method. The parameter optimization method for the control system comprises the steps that an orthogonal optimization method and the particle swarm optimization method are combined, the optimizing range of each controller parameter, the sensitivity relation between the controller parameters and the weight relation of performance indexes are determined through an orthogonal test, optimization is carried out by applying the particle swarm optimization method on this basis, loop iteration is carried out, and an optimization result is obtained. According to the method, a particle swarm is guided by the result of orthogonal optimization for optimization, optimizing speed is high, the optimization direction is clear, and the optimization effect is remarkable.

The invention discloses a deep learning hyper-parameter tuning improvement method based on Bayesian optimization. The method comprises the steps that a fast optimal Latin square experimental design algorithm is used to generate an initial point set, and the generation rate of effective evaluation points is improved; parallel computing is used to acquire the response of the initial point set in a computer, and the response of multiple points on an objective function is calculated at the same time to construct a data set; and a Bayesian optimization iterative process is initiated, and a parallelcomputing method is used in the computing process to accelerate the whole optimization process. According to the invention, the defects of long time consumption and large performance fluctuation of atraditional Bayesian optimization algorithm can be effectively solved; the optimization speed is accelerated through parallel computing; and the optimization efficiency is significantly improved.

The invention discloses a planar array near-field multipoint focusing system and method based on time reversal, and relates to the field of electromagnetic wave regulation and control. The planar array near-field multipoint focusing system comprises a signal generator for generating a channel detection signal, a transmitting module which is used for receiving the detection signal and transmittinga TRM return signal, a control module which is used for regulating and controlling the return signal, and a focusing module which transmits a detection signal. According to the planar array near-fieldmultipoint focusing system, SVD decomposition is utilized to represent a focusing field receiving signal as linear superposition of a plurality of orthogonal sub-signals so as to avoid redundancy caused by superposition of non-orthogonal signals; through threshold discrimination, a part of sub-channels are discarded, and the number of optimization variables is reduced, and the optimization speedis increased; a power fluctuation threshold is set to improve the transmission efficiency; and the energy ratio of each focus point is regulated and controlled by changing the power distribution ratio. Multi-point focusing can be generated in a near field with extremely uneven electromagnetic field intensity distribution according to a preset energy ratio, and the problems of microwave thermal therapy and multi-target selective wireless energy transmission can be effectively solved.

The invention relates to a polymer flooding production optimization method and system based on a Monte Carlo algorithm. The method comprises the steps that an optimum control function for oil reservoir optimization is generated by a simulator according to an oil reservoir value; a covariance matrix is generated by a solving model, an initial disturbance vector is generated nearby an optimum control variable by utilizing the Monte Carlo approach algorithm, the initial disturbance vector is processed by adopting the covariance matrix to generate a target disturbance vector; the Monte Carlo algorithm including the target disturbance vector is adopted to solve the optimum control function, and an optimum injection-production parameter corresponding to optimum exploitation benefit is generated. The calculation process of the method is simple and easy to achieve, the optimization efficiency is high, the convergence rate is high, the system is easily combined with any oil reservoir value simulator to calculate polymer flooding production, injection-production well production parameters and injection polymer concentration can be simultaneously optimized, an oil reservoir development effect can be remarkably improved, and a basis can be provided for scientific, reasonable and efficient oil field development. The method can be widely applied to the field of oil-gas field development.

The invention discloses an outdoor weather image classification method based on a transfer learning technology. The method comprises the following steps: firstly, establishing a deep network model, and then converting the trained deep network model in a checkpoint format into a deep network model in a pb format; and finally, judging the weather type of the image quickly and automatically only by inputting the image to be judged into the deep network model in the pb format. Classification accuracy is high, meanwhile, according to the outdoor weather image classification method, weather types can be classified; ready-to-classify fog, rain, sun and snow, meanwhile, subdivision can be carried out according to snowy weather; that is to say, subdivided into snows, medium snow and heavy snow, meanwhile, based on the transfer learning technology, the method has the advantages of being small in training data demand, short in model training time and high in optimization speed, the excellent deepnetwork ResNet is shown on an ImageNet data set in a transfer mode, and the high classification accuracy is obtained. The method is suitable for popularization and application in the technical fieldof image processing.

The invention relates to a ship propeller, which comprises a pushing mechanism and a driving mechanism. The ship propeller is characterized in that the pushing mechanism comprises a cylinder sleeve, a piston, a piston connecting rod and a pressure-limiting automatic valve opening device, wherein the piston is sleeved in the cylinder sleeve, one end of the piston connecting rod is hinged with the piston, the other end of the piston connecting rod is connected with the driving mechanism, and the piston is provided with a valve rod hole and a ventilation hole; and the pressure-limiting automatic valve opening device comprises a valve, a valve rod, a spring and a spring bottom plate, wherein the valve is of a disk shape, the valve rod is inserted into the valve rod hole, one end of the valve rod is connected with the valve, the other end of the valve rod is connected with the spring bottom plate, and the spring is arranged between the piston and the spring bottom plate and is sleeved on the valve rod. The ship propeller has the beneficial effects that the ship propeller is simple in structure, low in manufacturing cost, reliable in work, high in propelling speed and low in noise, and is energy-saving and environment-friendly; and due to the simple structure, the loss in energy and power is reduced.

The invention provides a fully mechanized coal mining face excessive fall fault novel method, which comprises the following steps of digging an exploratory roadway vertical to the fault trend in a haulage gate along a coal seam floor in advance; building a fault trend calculation model according to fault disclose information in the exploratory roadway; designing the blasting roadway trend along the fault trend; designing the blasting roadway cross section dimension and arrangement position according to the operation track of a coal cutting machine and a support frame; forming pre-fracturing deep holes at intervals in the blasting roadway; forming a groove pulling shallow hole in the middle of each pre-fracturing deep hole; installing a lug type chamber in a key position of the blasting roadway; forming fan-shaped drill holes in the chamber. During the blasting construction, the fan-shaped drill holes in the lug type chamber firstly detonate; then, the pre-fracturing deep holes in the blasting roadway detonate; the groove pulling shallow holes detonate at last. Before the blasting construction, a single body hydraulic support post is arranged in the blasting roadway according to requirements. The deep hole pre-fracturing blasting technology is used for pre-processing the fault zone rock; a coal mining machine can be effectively helped to pass through the fault; high yield and high efficiency of the coal mining face can be ensured.

The invention discloses a parallel mechanism multi-objective optimization design method considering parameter nondeterminacy. The parallel mechanism multi-objective optimization design method comprises the following steps that 1, an analysis mapping model among optimization objectives, constraint conditions and design parameters of a parallel mechanism is constructed; 2, the design parameters obtained after disturbance are regarded as random variables, a probability factor is calculated with the help of Monte Carlo simulation, an analysis mapping model of the probability factor is constructed,probability factor analysis mapping models of all the constraint conditions are fit through a least square method, precision evaluation is carried out on the probability factor analysis mapping models of all the constraint conditions, and the model with highest precision is adopted as the probability factor analysis mapping model of all the constraint conditions, and on this basis, the probability factors larger than or equal to 1 of probability factors are set as the constraint conditions of all parameters; 3, solving is carried out through an objective optimum matching method, and an optimum design parameter is obtained. According to the method, the mechanism optimizing speed is increased.

The invention discloses a photovoltaic cell parameter identification method based on an elite mass differential evolution algorithm. The implementation process framework comprises establishing a photovoltaic cell model and confirming the parameters required to be identified; collecting the data and setting a target function; carrying out population initialization, calculating a fitness value, and dividing the population into an elitist level and a crowd level, and performing mutation operation and crossover operation on the elitist level and the mass level, judging whether the elitist level and the mass level exceed a boundary, calculating the fitness value, performing selection operation, updating a scaling factor F and a crossover probability factor CR, removing the duplicates, and finally outputting an optimal parameter value needing to be identified in a photovoltaic cell. The method is used for identifying the parameters of the photovoltaic cell with the m * n type photovoltaic array in a photovoltaic grid-connected power generation system, the undetermined parameters of the photovoltaic cell I-V equation can be identified, the output power of the photovoltaic cell can be solved, and the identified parameters are used for analyzing the fault of the photovoltaic cell.

The present invention provides a method for describing the heterogeneity of a carbon dioxide drive reservoir layer. The method comprises: Step 1, calculating the permeability difference of blocks, anddetermining a reasonable permeability differential limit according to the permeability difference of study areas; Step 2, processing a permeability interpretation result in a certain study area and calculating the permeability difference of a grid model; Step 3, using a numerical simulation means to study the influence of different methods on the carbon dioxide drive gas breakthrough time; Step 4, analyzing the influence of physical and chemical properties of crude oil in the carbon dioxide drive on the ratio of gas to oil; Step 5, using the numerical simulation means to analyze the flow difference of different permeability value models; and Step 6, using the ratio of gas to oil in the actual model to fit, compare and correct the average permeability value. By virtue of the construction method provided by the present invention, according to the parameters that can be provided in the numerical simulation, a method for discretizing the permeability curve is established, and a numericalsimulation basis is provided for the improvement of the numerical simulation accuracy of the carbon dioxide drive and the quantitative characterization of the displacement mechanism thereof.

The invention discloses a multi-view data missing complementing method for multi-manifold regularization non-negative matrix factorization, which comprises the following steps of: obtaining manifold and global clustering in which unmarked multi-view data tends to be consistent by utilizing a multi-manifold regularization non-negative matrix factorization algorithm through consistency assumption among multiple views; and constructing a multi-view collaborative discrimination model by adopting a view collaborative improved Gaussian mixture method. Pre-calibration of a cluster to which a sample belongs is realized by calculating the cluster relevancy level of the sample with missing data under a non-missing view angle; and establishing a missing data prediction model under a specific view angle by utilizing the consistency of multiple view angles in a low-dimensional space and adopting a multiple linear regression analysis method, thereby realizing accurate data completion under a multi-attribute missing condition. According to the method, large-scale labeled samples are not needed for training, the pre-defined category relation and related characteristics are avoided, and the understanding and discovering capacity of an existing multi-view mining technology for unlabeled multi-source data is improved.

The invention discloses a fibrobronchoscopy intubation aid decision making method based on deep learning. The method comprises the steps that an intubation video is collected and divided into image frames; a decision instruction set is formulated; operation instructions are taken as category labels of image decision to label the image frames frame by frame; original images are processed to make adata set; a training set and a verification set are extracted; feature extraction is performed on a training sample, and high-level feature maps are obtained step by step; results after the feature maps are conversed are sent into a Naive Bayes classifier and a softmax activation function for calculation; the training sample is input, a loss function loss is set, and separate training is performedon an operation instruction decision model and a Naive Bayes classifier model separately; the training step is repeated to cross-validate a network model. According to the method, an end-to-end method is adopted, and a decision result is directly obtained by inputting the images, so that the algorithm cost is greatly reduced, the decision speed is high, and the real-time performance is very high.

The invention specifically relates to a method for optimizing the design on the number of regulating stage nozzles of a steam turbine based on an improved genetic algorithm. The method disclosed by the invention aims at optimizing the number of the regulating stage nozzles of the steam turbine under multiple load points, so that a unit can stably run under different loads, and the nozzles can constitute more valve points as far as possible, thereby allowing the damper loss to be minimum and the intra-regulating-stage efficiency to be maximum. The method mainly comprises the following steps of: based on the improved genetic algorithm, determining the optimal nozzle number combination, the regulating-stage rear flow, the downstream pressure of each valve and the air flow force which can be achieved through outputting a load point in a situation that the deviation value Y between actual flow and theoretical flow is minimum: and through improved genetic codes and a fitness function, carrying out a genetic algorithm operation, and then outputting the optimal nozzle number combination achieved in the situation that the deviation value Y between the actual flow and the theoretical flow is minimum. The method disclosed by the invention can improve the speed of an optimization process and the accuracy of optimization results.

The invention provides a plan implementation method and apparatus based on predicted dose guidance and Gaussian process optimization, and relates to the field of radiotherapy technology. The method comprises: calculating a predicted dose of a case by using a dose prediction model; calculating a plan score of the predicted dose according to the scoring rule as an optimal plan score; determining a plurality of sets of plan parameters based on the a priori database according to organ anatomy information; calculating corresponding plan scores of the plurality of sets of plan parameters, and constituting a Gaussian data set; and based on the Gaussian data set, calculating new plan parameters by using the Gaussian process, calculating the corresponding plan scores, adding the calculated scores to the Gauss data set, iteratively executing the step, and finally calculating the intensity-modulated optimization results under the plan parameters corresponding to the highest score in the Gauss data set. By the use of the predictive model to predict the dose distribution of the case to optimize the guidance, the quality of the plan is guaranteed, and by the use of the Gaussian process to calculate the posterior distribution based on the a priori data, the number of trial and error is reduced, thereby speeding up the optimization.