475results about How to "Reduce instability" patented technology

The invention provides a servo hydraulic system optimization control method and system for a press machine test platform, and relates to the technical field of press machines, and the method comprises the steps: building a state space expression and building a mathematical model of a servo hydraulic system based on dynamic characteristics by deducing mathematical expressions of all components of the system; building a predictive control algorithm based on the mathematical model, solving a quadratic programming problem by adopting a rolling optimization strategy through building a state space predictive model and determining a multi-objective optimization function to obtain an optimal control sequence, and taking a first element of the optimal control sequence as control input at the current moment; feedforward compensation and pressure feedback compensation are introduced, the feedforward compensation constructs a compensation model based on an inverse system of a press transfer function and pre-corrects a control signal, and the pressure feedback compensation designs a digital filter based on a pressure error, fuses a feedforward compensation signal with a pressure feedback compensation signal and superposes the fused signal with the control signal; and obtaining a final control signal.

The invention relates to a real-time river water level monitoring and early warning method and system based on deep learning. The method comprises the following steps: acquiring a target watershed remote sensing image and real-time hydrological data of multiple monitoring points; performing water flow direction topology analysis on the remote sensing image of the target watershed, and constructing a water flow topology network; performing basin structure dynamic change feature analysis on the target basin remote sensing image to generate basin structure dynamic change data; carrying out topology situation evolution fitting on the water flow topology network based on the basin structure dynamic change data to obtain a dynamic topology evolution network; performing multi-scale time sequence trend evolution analysis on the multi-monitoring-point real-time hydrological data to generate water flow time sequence evolution characteristics; and evolution characteristic space position mapping is performed on the dynamic topology evolution network based on water flow time sequence evolution characteristics, and a three-dimensional hydrological evolution model is constructed. According to the invention, timely and accurate river water level monitoring and early warning are realized.

The invention discloses a light spot real-time monitoring and feedback focusing system and method for laser processing. The system comprises a laser processing module, an emission light source module, a power monitoring module, a light spot monitoring module and an automatic control module. The emitting light source module is used for emitting laser beams; the power monitoring module is used for monitoring the laser power emitted by the emitting light source module; the light spot monitoring module is used for monitoring light spots of diffuse reflection light after the laser is emitted to the surface of the workpiece; the automatic control module sets reference signals of the following parameters: processing laser power and shape, position and size of a processing light spot; and according to the reference signal and the corresponding feedback signal, the laser emission parameter and the laser incident workpiece surface parameter are automatically adjusted, so that the error between the reference signal and the corresponding feedback signal is within a set precision range. According to the method, the light spot form and the light intensity in the machining process are dynamically monitored, and the light spot form and the energy in the laser machining process are adjusted in real time through feedback.

The invention discloses a charging pile dynamic power scheduling method, and relates to the technical field of charging pile power scheduling, and the method specifically comprises the following steps: obtaining node state information of each remaining normal node in real time, and generating a remaining bearing capability index and a load balancing index of each normal node; and constructing a power distribution evaluation model for the generated residual bearing capacity index and the load balancing index of each normal node, generating a power distribution evaluation coefficient of each normal node, analyzing the power distribution evaluation coefficient after generation, and evaluating the power bearing capacity of each normal node. Dividing each normal node into a high-bearing node, a middle-bearing node and a low-bearing node according to an evaluation result; and respectively adopting a corresponding power scheduling strategy according to a division result of each normal node. According to the invention, the problem that the power resource cannot be quickly and efficiently allocated after the fault node exits is solved, and efficient and stable dynamic balanced power scheduling and system operation are realized.

The invention discloses a wind power prediction method based on improved LSTM and FA-KELM, and the method comprises the steps: obtaining the wind power data of a wind power plant, carrying out the normalization of the wind power data, and carrying out the decomposition of the wind power data through employing CEEMDAN; calculating the fuzzy entropy of the obtained subsequences, performing secondary decomposition on the subsequence with the maximum fuzzy entropy by using a PSO-VMD algorithm, and dividing other sequences into a high-entropy sequence and a low-entropy sequence according to the fuzzy entropy; the method comprises the following steps: improving an LSTM network, and replacing an input gate, a forgetting gate and an output gate in the LSTM network with a gate control structure with a cross coupling structure to obtain ConvsLSTM; mIC is adopted to screen out meteorological factors with large correlation with wind power in the meteorological factors of the wind power plant; and in combination with the obtained meteorological factor data, predicting the obtained high-entropy sequence by using ConvsLSTM, and predicting the low-entropy sequence by using FA-KELM. And superposing the predicted values of all the components, and carrying out reverse normalization to obtain a final ultra-short-term wind power prediction result. According to the method, the appropriate prediction model can be selected for the power subsequences with different complexities, and the overall wind power prediction precision is improved.

The invention discloses a sea ice target detection method based on multi-dimensional feature extraction, and the method specifically comprises the following steps: S1, data preprocessing: carrying out the denoising of an original sea ice image through a Gaussian filter, carrying out the multi-scale analysis of the sea ice image through an image pyramid, and obtaining sea ice images under different scales; s2, local feature extraction: extracting texture features of each local region by using a gray-level co-occurrence matrix; s3, performing enhanced texture feature processing on the extracted texture features through a directional filter and an enhancement algorithm to obtain final texture feature data; and S4, performing feature fusion on the obtained multi-dimensional features and the texture features to form a comprehensive feature image, and obtaining a classification result of the sea ice target. According to the sea ice target detection method, through multi-scale analysis, image representation is obtained under different scales by using an image pyramid transformation method, changes and expressions of texture features under different scales are captured, and the stability of sea ice target detection is improved.

The invention discloses an alternating current charging pile ordered charging power regulation and control method, and relates to the technical field of alternating current charging pile charging power regulation and control, and the method specifically comprises the following steps: obtaining the operation state information of each alternating current charging pile in real time, and generating a voltage fluctuation influence index and a load dynamic response index of each alternating current charging pile; constructing a power influence evaluation model for the generated voltage fluctuation influence index and the load dynamic response index of each alternating current charging pile, generating a power influence evaluation coefficient of each alternating current charging pile, performing analysis after generation, and evaluating the influence degree of the three-phase voltage phase difference fluctuation on the charging power of each alternating current charging pile. And the alternating current charging piles are divided into low-influence charging piles, medium-influence charging piles and high-influence charging piles according to the evaluation result. According to the method, the problem that the charging power is influenced by three-phase voltage phase difference fluctuation is solved, and the efficient and stable effects of dynamic evaluation and accurate regulation and control of the power of the charging pile are achieved.

The invention relates to the technical field of data processing, in particular to a shrimp group health condition assessment method based on deep clustering analysis, and the method comprises the steps: monitoring health parameters; determining a shrimp group dispersion grid; determining a temporary grid; identifying a non-shrimp group abnormal grid; calculating an actual health index; and adjusting the grid step length and evaluating health. According to the method, the health condition of the shrimp group can be accurately monitored and evaluated by combining physiological and behavior data such as real-time density, saturation and activeness, abnormal grids are eliminated step by step, more representative grids are screened out, evaluation deviation caused by environmental fluctuation or measurement errors is avoided, and the evaluation accuracy is improved. The dynamic monitoring of the health condition of the prawn group is realized, and the grid step length is automatically adjusted by comparing the ideal health index with the actual health index, so that the health assessment precision and reliability are improved, and the problem of inaccurate health assessment result caused by unstable environmental parameters is effectively solved.

The invention discloses a garbage truck real-time monitoring and remote scheduling method based on the Internet of Things, and the method comprises the following steps: S1, collecting the real-time data of the operation state of a garbage truck, and carrying out the preprocessing of the real-time data; s2, constructing a task prediction model based on an improved adaptive graph neural network, and generating a preliminary task distribution matrix and a path planning scheme; s3, optimizing the preliminary task allocation matrix and the path planning scheme by adopting an improved PPO algorithm; s4, the optimized task allocation matrix and the path planning scheme are sent to a garbage truck vehicle-mounted terminal, and the task completion rate and the path deviation value are calculated; s5, the operation state of the garbage truck is monitored in real time, and the monitoring result and the operation state data are transmitted to the cloud Internet of Things platform; and S6, dynamically updating the scheduling instruction. According to the method, the Internet of Things, the adaptive graph neural network and the improved PPO algorithm are fused, intelligent scheduling and real-time optimization of the garbage truck are realized, and the method has the advantages of high efficiency, stability and low-cost operation.

The invention discloses a semi-supervised seismic wave impedance inversion method and system based on Swin Transform iterative gradient correction, and belongs to the technical field of earthquakes in geophysical exploration. According to the method, the impedance gradient is iteratively corrected in the model space through the deep neural network, and complex nonlinear prior information is captured; and each anisotropic structure direction regularization item is established, and constraint is carried out along the direction of the stratum structure, so that the space consistency of the prediction result is ensured. Meanwhile, the USTNet network designed by the invention combines the advantages of the local feature capture capability of the convolutional neural network and the global modeling capability of Transform, and reduces the non-uniqueness of solutions. The network is used for inverting synthetic data and actual data, and the effectiveness of the method is verified.

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 relates to the technical field of equipment control, in particular to a self-adaptive switching method and system of a current transformer. The method comprises the following steps: collecting primary side parameters and secondary side parameters of a current transformer to obtain primary side collection parameters of the current transformer and secondary side collection parameters of the current transformer; performing current transformer state segmentation initialization according to the current transformer primary side acquisition parameters and the current transformer secondary side acquisition parameters so as to generate a current transformer operation state classification result; switching result judgment is conducted on the current transformers according to the current transformer operation state classification result, and when the current transformer operation state classification result is a current transformer critical state section, parallel buffering is conducted on the current transformers, and critical current transformer switching data are generated. According to the invention, through accurate state classification, a dynamic switching mechanism, time sequence optimization and visual display, the switching sensitivity and stability of the current transformer are improved.

The invention relates to the field of remote management of parking locks, in particular to an intelligent parking lock remote management method and system based on Bluetooth communication. The method comprises the following steps: acquiring an intelligent parking lock operation log and an intelligent parking lock real-time Bluetooth connection state parameter; according to the intelligent parking lock operation log, user personalized behavior habit evolution is carried out, user portrait modeling is carried out, and an unlocking behavior portrait of each user is constructed; based on the unlocking behavior portrait of each user, real-time user behavior abnormal deviation quantification and abnormal parking lock use decision making are carried out on the intelligent parking lock real-time Bluetooth connection state parameters, and an abnormal parking lock early warning strategy is generated; when the real-time user behavior abnormal deviation value does not exceed a preset behavior deviation threshold value, triangulation positioning distance measurement and user accurate positioning calculation are conducted on the intelligent parking lock real-time Bluetooth connection state parameters, and real-time position coordinates of the user are obtained. According to the invention, safe and efficient remote management of the parking lock is realized.

The invention relates to the technical field of power generation prediction data processing, and discloses a photovoltaic power station power generation prediction method based on multi-model fusion, which combines a photovoltaic physical model, a power generation statistical model and a power generation prediction model, not only considers a physical rule, but also combines statistical learning and machine learning to improve the stability and accuracy of prediction, and improves the prediction efficiency. And predicting the generating capacity of the photovoltaic power station by using environmental meteorological data, electrical parameters of the photovoltaic power station and historical power generation data. Accurate photovoltaic power generation power prediction is beneficial to better arrangement of power generation plans and resource scheduling of the power grid. The prediction precision and stability are improved through multi-model fusion, so that power grid dispatching is more accurate and efficient, and power grid instability factors caused by photovoltaic power generation fluctuation are reduced. Therefore, photovoltaic power generation can be connected to a power grid more stably, the power demand is met, and wide application and development of clean energy are promoted.

The invention relates to plasma gun automatic spraying equipment which comprises a workbench, an X-axis moving assembly, a Y-axis moving assembly and a rotating shaft assembly, the X-axis moving assembly, the Y-axis moving assembly and the rotating shaft assembly are located above the workbench, the rotating shaft assembly is connected with an ion gun fixing support, and the X-axis moving assembly, the Y-axis moving assembly and the rotating shaft assembly move independently or cooperatively. And the plasma gun on the ion gun fixing bracket is driven to spray the workpiece according to the set motion path. Through an advanced automatic control system and an intelligent adjusting function, the spraying process can be accurately controlled without manual intervention, the consistency of the production efficiency and the spraying quality is remarkably improved, meanwhile, the dependence on manual operation is reduced, and the equipment maintenance cost is reduced. The equipment disclosed by the invention has higher adaptability, can be automatically adjusted to adapt to different types of workpieces, has a fully automatic spraying process, has higher application and popularization values and is suitable for large-scale production.

The invention discloses a regional power grid wind power generation power prediction optimization method which comprises the following specific steps: firstly, collecting data such as wind speed and wind direction through meteorological sensors arranged in a wind power plant and around the wind power plant, and carrying out cleaning and de-noising preprocessing; secondly, decomposing the meteorological data into intrinsic mode function components with different frequencies by adopting variational mode decomposition, constructing a convolutional neural network to perform feature extraction on each component, and adaptively adjusting the size of a convolution kernel according to the periodic characteristics of the data; then, a bidirectional long-short-term memory network based on an attention mechanism is used for fusing features, time sequence information is captured, power prediction is achieved through a full-connection neural network, and the weight of the FCNN is initialized through a random forest regression model to improve the initial performance of the model; according to the method, the prediction precision is improved through multi-level characteristic decomposition and a self-adaptive mechanism, the real-time response capability is good, and the requirements of the regional power grid for the precision and timeliness of wind power generation power prediction are met.

The invention discloses a new energy station management system and method based on an intelligent power grid, and the system comprises a real-time monitoring module which is used for predicting the charging load fluctuation of an electric vehicle based on real-time monitoring data, a scheduling module which is used for adjusting the output of new energy power generation equipment according to the predicted load fluctuation, and an optimization module. The power distribution network management module is used for optimizing charging scheduling in a charging station and avoiding overload caused by peak charging at the same time, the dynamic adjustment module is used for dynamically adjusting the voltage and frequency of a power distribution network, and the real-time monitoring module is used for monitoring the state of the power distribution network in real time and dynamically updating a management strategy; according to the new energy station management system and method based on the smart power grid, the problems that in the prior art, due to instability and rapid change characteristics of electric vehicle charging loads, severe fluctuation of the loads can bring extra pressure to a power distribution network, the overload problem of the power distribution network is caused, and then the reliability and safety of the power grid are affected are solved.

The invention discloses a bionic climbing robot cluster formation input sharing control method, and particularly relates to the field of robots, and the method comprises the following steps: S1, a plurality of bionic climbing robots are connected through a wireless communication module to form a cluster, and a main robot is selected as a cluster control center; s2, all the bionic climbing robots obtain environment data through respective environment sensing modules and upload the data to the cluster control center in real time; s3, the cluster control center generates a control instruction according to the collected environment data in combination with a preset task target and a formation strategy; s4, the cluster control center broadcasts the generated control instruction to all the bionic climbing robots in the cluster through a wireless communication module; and S5, the bionic climbing robots execute corresponding actions according to the received control instructions, and cluster formation and collaborative operation are achieved. The method has the advantages that automatic obstacle avoidance and conflict adjustment can be achieved, and efficient operation and cooperation in a complex environment can be achieved.

The utility model relates to the technical field of two-plate machines, and discloses a two-plate machine band-type brake synchronous rotating shaft device which comprises a driving motor, a fixed plate and a movable plate, a second fixed block fixed plate is arranged on the side portion of the second fixed block movable plate, and accurate control over a band-type brake can be achieved through the driving motor, a rotating rod and a control motor in the device. The synchronous movement between the two plates of the mold is ensured, so that the precision and the quality of a formed product are improved, the possibility of part failure caused by abrasion or aging can be reduced through the cooperation of a movable groove, a driving lead screw and a limiting rod, the service life of the device is prolonged, and two sets of control band-type brakes are arranged on the movable plate, so that the production efficiency is improved. And synchronous movement is realized through a complex transmission mechanism, so that the stability of the device is enhanced, unstable factors in a mold movement process are reduced, the problem that synchronous operation is out of control is solved, and the production efficiency of the injection molding machine is improved.

The invention relates to the technical field of alloy processing, in particular to a long-service-life elastic titanium-copper alloy for a VCM elastic piece and a preparation method thereof.The preparation method comprises the following steps that firstly, raw materials including 55%-65% of titanium, 30%-25% of copper, 10%-8% of aluminum, 3%-1% of zinc and 2%-1% of magnesium are weighed according to the mass percent; 2, the raw materials are put into a vacuum furnace to be melted, and a liquid melt is obtained; thirdly, the melt is poured into a casting mold for casting, and a blank is obtained; according to the method, the change of the state of the blank can be responded in real time by calculating the influence of the rolling times and the thickness change of the blank on the rolling control parameters, so that the rolling process can be flexibly adjusted according to the actual situation, the adaptability during machining is improved, and the rolling speed is improved when the rolling times are gradually reduced along with the reduction of the rolling times of the blank. The rolling temperature, ratio, speed and pressure are reduced, excessive deformation and damage of blanks can be effectively prevented, and the quality of final products, namely the elastic life of the alloy, is improved.

The invention relates to an impurity detection system for high-purity gas and a detection control method thereof, and the impurity detection system for the high-purity gas comprises a detection path selection module, a busbar sample introduction module, a chromatography pretreatment module and a moisture detection pretreatment module, the detection path selection module is connected with confluence path units contained in the busbar sample introduction module, the chromatography pretreatment module comprises a plurality of chromatography pretreatment units, the confluence path units are respectively connected with the corresponding chromatography pretreatment units, each confluence path unit is connected with the moisture detection pretreatment module, and the moisture detection pretreatment module is connected with the detection path selection module. The detection path selection module is used for acquiring storage information in a high-purity gas storage tank, and performing corresponding chromatographic analysis steps and / or moisture detection analysis steps on the high-purity gas according to the storage information; the opportunity and parameters of gas entering the detection equipment are automatically adjusted according to the storage states of different steel cylinders, so that the accuracy of a detection result is improved.

The invention relates to an interaction control method and control system for an operating room display terminal. The control method comprises the following steps: carrying out coordinate division on display areas under different interfaces of the display terminal; different logic control instructions are given to the interaction operation of each area based on different interfaces, the interaction action of the display area is detected in real time, and control instruction judgment is carried out based on the action type and coordinates of the interaction action and the area to which the coordinates belong; therefore, the switching between the control interface and the full-screen display interface and the interactive operation under the two interfaces are realized, the complexity of operating room image control is simplified, the instability possibly caused by multi-device integration of the control terminal is reduced, and the reliability and usability of the system are improved.

The invention provides a methanol preheating and catalytic cracking system for a methanol engine, a ship and a control method, and belongs to the technical field of ship methanol engines. The problems that an existing methanol engine is difficult in cold start and methanol gasification are solved. The system comprises a methanol engine, a methanol storage tank, a water storage tank, a heat exchanger, an inlet temperature sensor and an outlet temperature sensor, one passage of a third three-way valve is connected with the waste gas output end of the methanol engine, the other passage of the third three-way valve is connected with another cooling water pipeline, and the methanol output end of the heat exchanger is connected with the methanol input end of the methanol engine. The methanol input end is connected with the methanol storage tank, a methanol pump is arranged on a methanol conveying pipeline between the methanol storage tank and the heat exchanger, two passages of the first three-way valve are arranged between the methanol conveying pipelines, and the inlet temperature sensor and the outlet temperature sensor are respectively arranged at a methanol inlet and a methanol outlet of the heat exchanger. The method is mainly used for starting the methanol engine.

The invention relates to the field of osteoarthritis drug treatment, in particular to application of RuO2 nano-enzyme in preparation of a drug for treating osteoarthritis, and particularly relates to knee osteoarthritis. The RuO2 nano-enzyme disclosed by the invention is prepared according to the following method: carrying out incubation deposition on Cu2O sphere nano-particles and RuCl3 to obtain Cu2O and RuO2, and etching Cu2O and RuO2 spheres to obtain the RuO2 nano-enzyme. The RuO2 nano-enzyme provided by the invention is a hollow RuO2 nano-enzyme sphere, and after the RuO2 nano-enzyme is horizontally used in an animal body for intervention, the hollow RuO2 nano-enzyme sphere can be used as a hollow RuO2 nano-enzyme sphere; the RuO2 nano-enzyme can effectively delay the development of knee osteoarthritis of a mouse knee osteoarthritis model caused by medial meniscus instability (DMM), improve the pain condition and movement function of a DMM modeling mouse, reduce the gonarthromeningitis degree and relieve knee joint cartilage degeneration and injury, and the effectiveness of the RuO2 nano-enzyme in delaying the knee osteoarthritis disease is further proved. The invention provides a novel method for treating knee osteoarthritis, and has strong practicability and application prospect.

The invention discloses a vibration control method for lifting construction of a target object, and the method comprises the steps: measuring the height of the target object at any moment, determining the mass of a damping box body, and determining the optimal frequency ratio and the optimal damping ratio of a TMD and a bearing structure based on the mass of the damping box body, the TMD comprises the damping box body and a lifting device used for lifting the damping box body, the bearing structure comprises a lifting target object and a distance measuring device for measuring the height of the target object, the suspension length of the damping box body is determined based on the height of the target object under the condition that the optimal frequency ratio and the optimal damping ratio serve as constraint conditions, and the damping box body is lifted according to the suspension length, so that the damping box body is lifted at the moment t; the suspension length of the damping box body is matched with the height of the target object so as to eliminate or weaken the vibration influence. In the construction process of lifting the target object, vibration, caused by wind loads, of the target object can be eliminated or weakened, and the construction safety is improved.

The invention discloses a multistage self-adaptive enhanced mixing device for a solid fuel ramjet engine, which belongs to the technical field of solid fuel ramjet engines and comprises a front flange, a rear flange, an afterburning chamber barrel, a mixing partition plate, a heat insulating layer, a pressure measuring seat and a temperature measuring seat. The front flange and the rear flange are welded with the afterburning chamber barrel, a water pressure test is carried out after welding is completed, the mixing partition plates are bonded to the inner wall of the afterburning chamber barrel through high-temperature-resistant glue, and the mixing partition plates are distributed in a multi-stage mode, are separated through a heat insulation layer and are bonded through the high-temperature-resistant glue. In order to verify the effect of the multi-stage self-adaptive enhanced mixing device for the solid fuel ramjet engine, a pressure measuring point and a temperature measuring point are arranged on an afterburning chamber barrel and a mixing partition plate; according to the multi-stage self-adaptive enhanced mixing device, the combustion efficiency of a solid fuel ramjet can be remarkably improved, and the efficiency is further improved along with the increase of the number of the mixing partition plates.

The invention discloses a PVG preparation method and device based on a vector light field generated by a spatial light modulator. The device comprises an illumination system, a phase design system and an exposure system. The lighting system is mainly formed by assembling a laser and a beam expander, and the phase design system is composed of two spatial light modulators (SLMs) which are named as a phase design system 1 and a phase design system 2 respectively. Wherein the lighting system and the phase design system 1 are sequenced on a center line, and the phase design system 2 and the exposure imaging system work on a vertical center line. According to the method, the flexibility of manufacturing the holographic waveguide of the polarization body is greatly improved, phase regulation and control are carried out through the programmable electric control element SLM to control the transverse period, and experimental errors caused by manual adjustment of a reflector in a traditional mode are reduced; meanwhile, the method is different from a traditional interference exposure method, and the whole system has high robustness; in addition, compared with a separated interference light path, one beam expander is used, and the problem of light beam consistency in the beam expanding process is avoided.

The invention provides a coastal high spot monitoring green tide recognition early warning method, medium and system, and belongs to the technical field of green tide recognition early warning, and the method comprises the steps: collecting a large number of high spot monitoring shot green tide images, carrying out the JSON data labeling of the green tide contour, converting the JSON labeling into a PNG mask image, and dividing the PNG mask image into a training set, a verification set and a test set. A deep learning green tide semantic segmentation algorithm is designed, PSPNet is selected and a backbone network is replaced by ShuffleNetV2, an ECA attention mechanism and multi-scale feature fusion are added, and the problem of class imbalance is solved in combination with FocalLoss and BCELoss. And training the model by using the training data set, and updating the weight through error back propagation. And evaluating the performance by using the verification set, adjusting the hyper-parameter optimization model, and selecting an optimal model. And inputting a to-be-recognized image into the model to obtain a green tide segmentation mask, and judging whether a threshold value is exceeded or abnormal change occurs according to a segmentation result.

The invention discloses a thermo-acoustic oscillation passive control axial grading combustion chamber which comprises a fuel gas distribution pipe, a head gas distribution cavity, a cooling end plate, a weak rotation micro-mixing on-duty nozzle, a direct injection micro-mixing main combustion nozzle, an axial grading nozzle, a flame tube, a flow guide bushing and an outer casing. Due to the combustion arrangement of head micro-mixing combustion and axial staged combustion, fuel gas distribution is optimized; the ignition / load-per-liter combustion stability of the combustion chamber is realized through the weak-rotation micro-mixing on-duty nozzle; the axial grading nozzle adopts a lobe structure to enhance fuel and air mixing; the cooling hole structure of the axial grading nozzle is optimized, and the effect of restraining thermo-acoustic oscillation is achieved; and the opening positions of the axial grading windward side and leeward side cooling holes are improved, so that flame lifting under the high-hydrogen-content fuel is realized. According to the axial grading combustion chamber arrangement mode, the residence time of smoke in a high-temperature area is shortened, NOx pollutant emission of the combustion chamber is reduced, and the requirement for efficient, stable and low-emission operation under the condition of high-hydrogen-content fuel is integrally met.

The utility model provides an automobile glass lifter motor support connecting structure which comprises a positioning assembly which comprises a first positioning plate and a second positioning plate, a limiting plate is arranged on one side of the positioning assembly, a hydraulic rod is arranged on one side of the first positioning plate, a sliding plate is fixedly installed at the telescopic end of the hydraulic rod, and a first clamping plate is rotatably installed at one end of the sliding plate. A second clamping plate is rotationally installed at one end of the first positioning plate. According to the motor support connecting structure of the automobile glass lifter, the hydraulic rod, the sliding wheel and other structures are arranged, when the hydraulic rod drives the sliding plate to move, the first clamping plate and the second clamping plate are indirectly driven to rotate, the position of glass is adjusted through movement of the first clamping plate and the second clamping plate, and the first clamping plate and the second clamping plate are further limited by embedding the sliding wheel into the second sliding groove; and meanwhile, when the first clamping plate and the second clamping plate move, the first clamping plate and the second clamping plate are positioned in real time through the clamping strips, so that the first clamping plate and the second clamping plate are more stable in the moving process, and the stability of glass lifting is effectively improved.