84results about How to "Calculation results are stable" patented technology

The invention discloses a neural network method for performing short-term earthquake prediction by using earthquake parameters, which is a method for predicting earthquakes by using regression analysis and neural network combined technology and comprises: (1) determining a region to be studied; 2) collecting and preprocessing earthquake catalogues; 3) acquiring the earthquake parameters and time varying parameters by using the earthquake catalogues; 4) sorting information; 5) performing regression analysis; 6) performing neural network simulation; and 7) computing earthquake magnitudes of predicted earthquakes. The method has the advantages that: regional short-term earthquake prediction can be realized and the prediction time is no more than 6 months; earthquake magnitude quantization prediction can be realized, the computing result is stable and the prediction precision is high; the maneuverability and practicality are high; and the regional short-term earthquake prediction result can be used to provide service for 'disaster prevention and reduction', and the social and economic benefits are obvious.

The invention discloses a hierarchical planning-based urban conventional bus line network designing method. Hierarchical division is carried out on the line network according to a passenger travel mode, and according to a hierarchical planning result, optional lines in each hierarchy are searched sequentially, and demand allocation is carried out; with a demand allocation structure as a basis, an urban conventional bus line network designing model with a departure frequency as a variable, with the minimization of the cost of the passenger and an operator as an objective function and with restrictive conditions such as a line length, the departure frequency, a transport capability and a line number as constraints is built; and a simulated annealing algorithm is adopted to solve the model, the line network structure and the line departure frequency can be determined at the same time finally, and one-time networking of the hierarchical planning-based bus line network designing model is realized. Traffic zones can be corresponding to multiple departure or arrival sites, which is more consistent with the actual setting condition of sites around the traffic zones, and the calculation result of the model is stable.

The invention discloses a three-dimensional mark point matching method for point cloud registration and a three-dimensional scanning method, and belongs to the field of three-dimensional measurement technologies. The three-dimensional mark point matching method comprises the following steps: acquiring three-dimensional coordinate sets P and Q of two visual-angle mark points through a binocular camera; calculating energy of candidate matching points, which are in regard to a point p in the set P, in the set Q; selecting a certain updating strategy according to the 'uniqueness' constraint and the 'spatial feature unchange' constraint; if the matching points found out are all highest in energy and meet a threshold value condition and a restraint condition, ending iteration, otherwise, returning to iteration until the matching points meeting the conditions are found; calculating Euclid transformation according to the matching points, and performing splicing of multi-view point cloud. Compared with other mark point splicing methods, the splicing method has the advantages that the operation is simpler, the number of iteration times is small, 1-2 times of iteration are performed generally, the splicing time is greatly shortened, and a splicing result is stable and reliable under a condition of enough overlapping regions; the three-dimensional mark point matching method can be widely applied to multi-view point cloud data splicing acquired by various sets of measurement equipment.

The invention discloses a design method for a conventional urban public transit network. The method includes the steps that a known urban road network is simplified; then, a design model of the conventional urban public transit network is established with the departure frequency as a variable, the minimum passenger and operator cost as an object function and the minimum departure frequency, the vehicle capacity and other limiting conditions as constraints; finally, the optimal public transit network design and the corresponding departure frequency with the minimum total cost can be obtained by adopting a simulated annealing algorithm for solving. The dual purposes of public transit route design and operation dispatching are achieved in a one-time networking mode, route adjustment is performed based on an obtained optimized network, and therefore the public transit network is more reasonable in design, high in quality and more stable in structure and has higher practicability.

The invention relates to a method for estimating the medium quality factor by utilizing the peak instantaneous frequency of the envelope of the seismic signal, which includes the following steps: 1) dividing the medium into several thin plates with the thickness of the distance between two adjacent geophones in the well; 2) for the i-th A small thin plate, use the constant phase wavelet to approach the direct wave received at its top, and determine its modulation frequency σi and energy attenuation factor δi, and calculate the ratio ηi=σi / 2πδi; 3) Calculate the travel of the seismic wave on the i-th small thin plate Time 4) Calculate the instantaneous frequency of the received signal at the upper and lower interfaces of the i-th thin plate; 5) Calculate the change Δfp(i)=fp(0)-fp(τi) of the envelope peak instantaneous frequency in the i-th small thin plate; 6 ) Substituting σi, δi, τi and fp(0)-fp(τi) into the analytical relationship between the envelope peak instantaneous frequency and the quality factor Q value to calculate the quality factor Q value of the i-th small thin plate: 7) Repeat steps From 2) to 6), calculate the quality factors of other small thin plates in turn; 8) use the quality factor Q value to predict the oil and gas of the reservoir.

The invention discloses a novel, highly-efficient and dynamic multi-runway airport entering-flight multi-efficiency optimization ordering method aimed at providing a safe, scientific and feasible entering scheme for entering flights at a terminal area. By defining a multi-performance runway distribution evaluation function, runways and entry air lines are allocated and determined; delay distribution is optimized and flight landing time is determined according to a flight comprehensive priority so as to realize airport runway flow uniform distribution, reduce flight entry flight time and flight delay, raise spatial domain resource utilization efficiency, reduce control work load and further raise the operation efficiency of the entering flights in a multi-runway airport.

The present invention relates to an improved algorithm of water energy utilization improvement rate and cascade hydropower station dispatching benefit evaluation system, comprising the following steps: according to the monthly average water level of the actual operation of the reservoir in the past ten years, the dead water level of the reservoir design, the normal high water level of the reservoir design, and the reservoir dispatching diagram Data, calculate the monthly assessment index of the reservoir water level control coefficient, and inversely find the water level control target; calculate the actual power generation; calculate the assessment power generation according to the initial water level, interval flow, water level control target, and dispatching rules of the actual operating water level control coefficient of the reservoir; calculate the storage capacity Differential power; calculate the improvement rate of water energy utilization based on actual power generation, assessment power generation, and storage capacity difference power. The invention can improve the stability and adaptability of the electricity calculation results of the cascade hydropower station group assessment, and can avoid the situation that the water level will drop to the stagnant water level when the inflow of water is less at the beginning of the year, according to the dispatching chart. to good guidance.

When the heart beats, ventricular ejection is performed to genertate a pulse wave which is transmitted to the whole body at a certain velocity. The pulse wave velocity (PWV) refers to the velocity of the pulse wave which is transmitted from one position of an artery to another position of the artery. The parameter is always used in the diagnosis of vascular diseases and particularly used to judge the degree of vascular sclerosis. When the degree of vascular sclerosis increases, the arterial compliance is reduced and the PWV is increased. The current noninvasive measurement method of the PWV is to divide the length of a blood vessel by the conduction time of the pulse wave in the blood vessel, thus how to correctly identify the initial beating point becomes a main problem for calculating the PWV. The computation method of the invention utilizes wavelet transform to extract the initial beating point of the pulse wave. The method comprises the following steps: the multiresolution decomposition method is utilized to decompose the pulse wave, the a'trous algorithm is used to ensure that the decomposition result has the same length of the original signal and the detailed and general information of the pulse wave can be obtained; and the result of wavelet transform is compared with the pulse wave, and an obvious peak appears at the initial beating point of the pulse wave in the decomposition result of the pulse wave, thus the position of the initial beating point can be obtained through the comparison of the decomposition result and the pulse wave.

The invention relates to a virtual plant canopy photosynthesis effective radiation distribution simulating method of a multi-layer data structure. The method includes the following steps of establishing a virtual plant model and a virtual plane scene high in fidelity and complexity, collecting plant canopy photosynthesis effective radiation intensity values through a related astronomical parameter method or a field measurement method, sequentially establishing the three-layer data structure with a virtual plane scene single-tree geometrical model, a scene and a terrain through a data structure starting method, establishing a direct solar radiation and sky scattering PAR ray vector plane, sequentially tracing all ray vectors on the ray plane through a Monte Carlo ray tracing algorithm and a turtle shell algorithm, traversing the three-layer data structure with the scene, the single tree and the terrain, judging whether rays intersect with triangular patches or not, counting the sky visible rates of all the triangular patches, and calculating direct solar radiation PAR and sky scattering PAR of all the triangular patches to obtain a virtual plane canopy PAR three-dimensional space distribution simulating result.

The invention discloses a spectral data dimension reduction method for optimizing kernel independent components based on a sparrow search algorithm. The spectral data dimension reduction method comprises the following steps: acquiring near infrared spectral data of a tested sample; selecting a kernel function, determining a kernel parameter optimization problem and constructing a fitness function;setting initial parameters of the sparrow population, and generating an initial position matrix and a fitness matrix of the sparrow population; updating the optimal position and the optimal fitness value of the sparrow population according to the predation and anti-predation behaviors of the sparrows; applying the optimal kernel parameter matrix meeting the convergence condition to a kernel function; performing spheroidization decomposition preprocessing on the spectral data; and solving an unmixing matrix in independent component analysis to realize effective dimensionality reduction of spectral data. According to the method, blindness of kernel function parameter selection in a spectral data dimension reduction process by applying a kernel independent component analysis method is avoided, effective dimension reduction of near-infrared spectrum high-dimensional nonlinear data is realized, and the method is excellent in performance, high in practicability, stable in calculation process and easy to realize.

The invention discloses a cross-basin multi-scale calculation method based on a second-order nonlinear constitutive model. This second-order nonlinear constitutive model, starting from the model of moment method and without simplification, directly adopts the Eu closure method to close the left end of the stress and heat flux transport equation, and then introduces Rayleigh method into the solution of the second-order constitutive model. Onsager dissipation function reduces the nonlinear algebraic equations to an equation about dissipation function R, and the steepest descent method is used tosolve the R equation of monoatomic and diatomic gases, and the corresponding stresses, heat flux and additional volume stresses of gases are obtained. It makes up for the shortcomings of NS equationin solving rarefied flow and cross-basin problems, and provides an effective solution for transitional flow, rarefied flow and multi-scale flow. In order to shorten the convergence time and improve the computational efficiency, a variety of iterative methods are used in the numerical solution of the method.

The invention discloses a day-ahead reactive power optimization method based on a branch-bound method and a primal-dual interior point method and belongs to the reactive power optimization technology field of a power system. The method is characterized in that constraints of discrete variables in each period, adjusting gears in adjacent period and whole-day operation times are considered, and primary problems are branched and combined with branched branch-cutting strategies according to the periods, so that all discrete variables gradually approach to discrete values and meet the coupling constraints among the period; the primary problems are decomposed into a series of single-period reactive power optimization sub-problems only containing continuous variables after being branched according to the periods, the sub-problems are solved by adopting the primal-dual interior point method, and the calculated amount is reduced. The day-ahead reactive power optimization method is excellent in convergence, high in reliability and steady in calculation result.

The invention discloses a power grid new energy consumption capability evaluation calculation method, device and system based on a multi-scene generation technology, and the method comprises the steps: determining a power grid range and a calculation period of power grid new energy consumption capability evaluation calculation needing to be carried out, and other calculation boundaries; generatinga large number of combined scenes containing multiple new energy power plants through free permutation and combination among the fields; adopting a scene reduction technology of a backward reductionmethod, merging similar scenes, reducing the number of combined scenes, and generating typical combined scenes; carrying out the similar time period merging analysis, and reducing the number of time periods entering optimization; for each typical combination scene, establishing an SCUC dimension reduction model after time period merging, and solving to obtain a unit combination result; establishing an all-time SCED model based on the unit combination result and solving, and finally obtaining a new energy consumption result under each typical combination scene, thereby finishing power grid newenergy consumption capability evaluation calculation based on a multi-scene generation technology. According to the invention, on the premise of ensuring the safety and stability of the system, the effectiveness and referenceability of the consumption capability evaluation result are improved.

The invention discloses a coordinated dynamic reactive power optimization method for a phase modifier and a hierarchical-structure UHV DC receiving end converter station. The coordinated dynamic reactive power optimization method comprises the main steps of: 1) establishing a hierarchical-structure UHV DC receiving end converter station dynamic reactive power optimization model M; 2) and calculating the hierarchical-structure UHV DC receiving end converter station dynamic reactive power optimization model M, so as to optimize the hierarchical-structure UHV DC receiving end converter station. The coordinated dynamic reactive power optimization method adopts a hybrid solution algorithm based on a nonlinear interior point method and a dynamic programming method, comprises three sub-problems of global optimization based on the nonlinear interior point method, discrete control variable optimization based on the dynamic programming method and continuous control variable optimization based onthe nonlinear interior point method, and has the advantages of high solution efficiency and stable calculation result.

The invention relates to the field of computer image processing, and discloses a lane detecting and partitioning method based on traffic intersection videos. The method includes the steps that Hough conversion is used for detecting video images of a traffic intersection; multi-dimensional feature vectors are extracted; a K-means algorithm is used for achieving clustering. The lane detecting and partitioning method has the advantages that good adaptability, practicability and stability are achieved, the anti-interference capacity is good, operation steps and computation steps are simplified, consumption of computing resources is reduced, besides, a stable detection result can be further obtained, and the lane detecting and partitioning method is adaptive to lane analytical application of various intelligent transportation systems and has high application value.

The invention discloses a downhole microseismic three-component rotating method and belongs to the field of downhole microseismic three-component signal processing. The method includes: firstly, inputting an X-component and a Y-component, then solving extreme values to obtain an X-direction azimuth and a new R-component and a new T-component; secondly, inputting a Z-component and the R-component prior to solving extreme values to obtain a Z-direction azimuth and a new P-component and a new S-component; finally, taking the P-component, the S-component and the T-component as rotating results of the original X-component, Y-component and Z-component. By the method, accurate rotation of the downhole microseismic X-component, Y-component and Z-component can be realized, the new three components P, S and T uniform in amplitude are obtained, and information such as wave propagation and the azimuth theta of the X-component and the azimuth phi of the Z-component can be obtained. In addition, the method is small in influence of noise and stable in computed results, and input data are unlimited to P-waves or S-waves.

The invention discloses a system for measuring a displacement field in a reconstructed material component based on CT. The system comprises an experiment platform, a CT machine main body, a to-be-measured material component, iron sand particles, a substance identification and image processing computer, a control computer matched with a CT machine, a CT machine display control platform and a data analysis computer, wherein the CT machine is used for acquiring CT scanning data of the to-be-measured material component; the substance identification and image processing computer is used for carrying out three-dimensional reconstruction on the CT scanning data of the to-be-measured material component to obtain a three-dimensional position distribution diagram of iron sand particles of each moment; and the data analysis computer is used for carrying out image processing on the three-dimensional position distribution diagram of the iron sand particles, carrying out whole-field matching on positions of all the iron sand particles at different phases, and determining the position change of all the iron sand particles at the different moments to form the whole displacement field. The measurement system is simple to operate and has high precision; the influences on the mechanical property of the to-be-measured material component are small and the measurement of the whole displacement field in the material component is realized.

The invention discloses a rapid lithium battery capacity degradation early warning method. The method comprises the following steps: obtaining a lithium battery degradation curve of a lithium batterycharge-discharge cycle after data preprocessing, the lithium battery degradation curve comprising a starting point Q1 and an ending point Q2; determining a characteristic included angle of the lithiumbattery degradation curve according to the bending degree of the lithium battery degradation curve; comparing the characteristic included angle of the lithium battery degradation curve with a characteristic included angle alarm threshold value and a characteristic included angle diving threshold value; determining whether rapid lithium battery capacity degradation occurs or not according to a comparison result; and triggering rapid lithium battery capacity degradation alarm when determining that the rapid lithium battery capacity degradation has occurred.

The invention relates to a measuring and calculating method of the water level control improved efficiency of cascade hydropower stations. The method includes following steps: calculating a water level control target interval of a reservoir; calculating a practical generating capacity; calculating an examination electric quantity; determining a value of the responsible abandoned water power; calculating the storage capacity differential electric quantity; and calculating the water level control improved efficiency according to the practical generating capacity, the examination electric quantity, the responsible abandoned water power, and the storage capacity differential electric quantity. The invention also relates to an optimization dispatching benefit evaluation system of basin cascade hydropower stations in order to realize the measuring and calculating method of the water level control improved efficiency of the cascade hydropower stations, and the system specifically comprises a client layer, a service layer and a data storage layer. According to the method and the system, the problem of large influence of the water level at the beginning of the year on a hydropower utilization increase rate traditional algorithm is solved, and the unreasonable condition that the dispatching of the power stations often operate in a dead water level according to a design dispatching chart when the water level is low at the beginning of the year and the inflow is low is greatly improved.

The invention relates to a trademark image retrieval method and system, a medium and equipment. The trademark image retrieval method comprises the following steps: according to all similar group numbers of each trademark classification number, carrying out library division on a feature library containing feature vectors of trademark images; clustering the feature vectors of the trademark images ineach sub-library according to the determined number of the clustering centers, and dividing a sub-library for the feature vectors belonging to the same clustering cluster in each sub-library; determining a sub-library to which the to-be-queried trademark belongs according to the classification number of the to-be-queried trademark and the trademark image; calculating the distance between the feature vector of the trademark image of the trademark to be inquired and the feature vector of the trademark image in the subordinate sub-library; and outputting the trademarks corresponding to the feature vectors of the first N trademark images with the smallest distance as retrieval results. According to the method, the search range during online retrieval can be greatly reduced, and especially when the vector dimension is relatively high, the retrieval efficiency can be remarkably improved.

The invention belongs to the field of direct trading for large power users and particularly relates to a direct trading method considering influence factors of grid losses and using a centralized matching method for cross-district and cross-province power users. The direct trading method includes the steps of (1) the establishment of a mathematical model of direct electricity purchasing trade considering the grid losses for cross-district and cross-province power users; (2) a measuring method of the influence factors of the grid losses; (3) ordering rules of direct purchasing trade considering the grid losses for cross-district and cross-province large users; (4) a matching process of direct trading considering the grid losses for cross-district and cross-province large users. The direct trading method has the advantages of considering the influence of grid losses on the premise that the ordering rules of direct purchasing trade for cross-district and cross-province large users comply with the principle of 'price priority and energy conservation and environmental protection priority', allowing trading methods to be further elaborated, and being more beneficial for preserving the trade fairness of the power market.

The invention discloses a rapid lithium battery capacity degradation identification method and device. The method comprises the following steps: acquiring a lithium battery degradation curve of a lithium battery charge-discharge cycle after data preprocessing, the lithium battery degradation curve comprising a starting point Q1 and an ending point Q2; determining a characteristic included angle ofthe lithium battery degradation curve according to the bending degree of the lithium battery degradation curve; comparing the characteristic included angle of the lithium battery degradation curve with a characteristic included angle diving threshold value; and identifying whether the lithium battery belongs to a rapid lithium battery capacity degradation sample or not according to a comparison result.

The invention discloses a vehicle total mass dynamic measuring method based on power balance. The vehicle total mass dynamic measuring method comprises the steps of: firstly, calculating vehicle driving power and vehicle acceleration corresponding to each moment according to driving parameters of a vehicle in an actual driving state within a period of time or at different moments within a certainmileage; secondly, screening data of the vehicle driving power and the vehicle acceleration; thirdly, subjecting the data obtained by screening to regression calculation to obtain measurement model parameters; and finally, calculating vehicle total mass in the actual driving state through acquiring reference model parameters and vehicle total mass of the vehicle in a no-load driving state and combining with the measurement model parameters. Therefore, the vehicle total mass dynamic measuring method does not need to additionally arrange any sensor, is low in cost, and has more precise and morestable calculation results compared with the method of calculating the vehicle total mass based on single-point data.

The invention discloses an ore rock block degree measurement method based on improved FERRET. According to the method, target objects are subjected to labeling processing after the contours of the target objects are smoothed, the area, mass center and primary axis direction of each two-dimensional target object are solved based on 0-, 1- and 2-order moments of inertia, then a secondary axis direction which passes through the mass center and is vertical to the primary axis direction is solved according to the primary axis direction, the minimal external rectangle of the target object is solved based on the primary axis direction and the secondary axis direction, the length and width of the external rectangle are measured and regarded as the length and width of a measured object, the length-width ratio is regarded as a first shape parameter of the measured object, and the ratio of the area of the measured object to the area of the external rectangle of the measured object is regarded as a second shape parameter of the measured object. Compared with the prior art, the method disclosed by the invention is not affected by the rotation of objects, and two different shape parameters are added.

The invention provides an edge calculation task allocation method based on a branch and bound method, and belongs to the field of edge calculation. According to the method, the total energy consumption of task allocation is minimized on the premise of considering constraints such as task completion time and DAG parallel system reliability requirements. Firstly, the optimization problem is relaxed,and a temporary solution is obtained through an interior point method; taking a discrete value of a decision variable of a first task in the temporary solution and meeting the constraint that one task can only be executed in one edge server; wherein only one decision variable is taken as 1 and the other decision variables are taken as 0, traversing the value taking condition from the first edge server to the last edge server, continuously solving the remaining tasks by adopting an interior point method, calculating the energy consumption values of different solutions, taking the solution withthe minimum energy consumption value, and repeating the step until the last task. The method has the advantages of good convergence, small calculated amount and the like.

The invention discloses an expressway front vehicle detection method. The method includes the following steps that: multi-scale transformation is performed on acquired original images of expressway front vehicles, so that contracted images can be obtained; edge detection is performed on the contracted images, so that detection result images of vehicle contours can be obtained; the detection result images of the vehicle contours are numbered, so that numbered vehicles can be obtained; the numbered images are scanned, and the external rectangle of each numbered vehicle is calculated, and the parameters of the external rectangle of each numbered vehicle can be calculated according to the obtained external rectangle. The expressway front vehicle detection method has the advantages of stable calculation results as well as being concise and effective, and can effectively reflect the driving conditions of the front vehicles through the parameters of the external rectangles.

The invention discloses a safety constraint unit combination rapid calculation method and system based on model dimension reduction. The method comprises the steps as follows: performing calculation to obtain an active power output result of each unit and a power flow result of each power transmission section according to a preset safety constraint economic dispatching relaxation model; dividing generator units into must start and stop units and buffer units according to the active power output result of each generator unit; dividing the power transmission sections into valid sections and invalid sections according to the power flow result of each power transmission section; obtaining a safety constraint unit combination dimension reduction model by taking the start-stop state of the buffer units as a combination decision variable and considering the limit value constraint of the valid power transmission sections; and performing calculation to obtain a unit combination result accordingto the safety constraint unit combination dimension reduction model. By identifying invalid shaping variables and network constraints, the feasible region is reduced, and the calculation efficiency is improved.

The invention discloses an electric power spot market clearing calculation method based on multi-strategy fusion dimension reduction, and the method comprises the steps: firstly obtaining a power gridrange and a calculation period for carrying out the clearing calculation of an electric power spot market, and determining a calculation boundary; then carrying out merging analysis of similar time periods, and reducing the number of time periods for entering clearing calculation; establishing and solving a relaxation safety constraint unit combination model, carrying out invalid start-stop variable identification and effective safety constraint identification according to a calculation result, and reducing the number of start-stop variables and safety constraints entering clearing calculation; establishing a time period merged security constraint unit combination dimension reduction clearing model, and quickly solving to obtain a unit combination result; and carrying out full-time-periodeconomic dispatching and safety check analysis based on a unit combination result, adding the out-of-limit time period and the monitoring section into a unit combination dimension reduction clearingmodel, and finally obtaining an electric power spot market clearing result through iterative computation.

The invention provides a geomagnetic Kp index inversion method and system, an electronic device and a storage medium. The method comprises the following steps: removing a background magnetic field ofa geomagnetic field from geomagnetic field observation data to obtain geomagnetic field disturbance amplitude values DFi of an observation day at different moments; calculating an accumulated geomagnetic disturbance amplitude value DTj of each time period of the observation day according to the geomagnetic disturbance amplitude values DFi; according to the accumulated geomagnetic disturbance amplitude value, inversing the geomagnetic Kp index by a formula Kp = Aj[lg(DTj) + Bj], wherein different time periods in one day are corresponding to different Aj and Bj, and j is a time period serial number. According to the method, the geomagnetic field disturbance amplitude value serves as the calculation basis, the geomagnetic Kp index is inverted by adopting the accumulated geomagnetic disturbance amplitude value, and the calculation result is stable. According to the method, a statistical relationship can be directly established between a K index observation value and a Kp index of a singlegeomagnetic observation station, and real-time calculation can be realized. Through the inversion system of the geomagnetic Kp index, the electronic device and the storage medium, the inversion methodof the geomagnetic Kp index can be realized, and the same beneficial effects are achieved.

The invention relates to a lithium battery capacity diving turning point identification method based on a geometric feature fusion decision, and the method comprises the following steps: 1), obtaining a capacity recession curve of a to-be-identified lithium battery online, and carrying out smooth denoising and biaxial normalization processing to obtain a normalized lithium battery capacity recession curve; 2) connecting a starting point C1 and an ending point C2 on the normalized lithium battery capacity recession curve to form a completely linear aging reference line C1C2; 3) calculating the degree of deviation of points on the normalized lithium battery capacity recession curve from the aging reference line, comparing the maximum deviation value with a capacity diving early warning threshold beta, and judging whether capacity diving occurs or not in real time; and 4) triggering lithium battery capacity diving early warning when capacity diving occurs, and taking the point corresponding to the maximum deviation degree as the identified lithium ion battery capacity diving turning point. Compared with the prior art, the method has the advantages of accurate identification, automatic batch processing and the like.