41results about How to "Accurate performance assessment" patented technology

The invention discloses a turbine pump inducer cavitation flow numerical prediction method based on low-temperature fluid, and belongs to the technical field of impeller machinery. The implementationmethod comprises the following steps: establishing an inducer model and an inlet and outlet runner model according to actual working conditions, and dividing grids; Establishing a computational fluidmechanics model, carrying out rotation correction on the turbulence model, and carrying out thermodynamic correction on the cavitation model; Setting and calculating boundary conditions of the single-phase steady flow to obtain a single-phase result; Calculating two-phase steady flow by taking a single-phase result as an initial value, and then changing inlet pressure and outlet mass flow to obtain calculation results of different working conditions; Calculating unsteady flow of the two phases by taking the two-phase steady result as an initial value to obtain an unsteady calculation result; By analyzing a calculation result, accurate flow field characteristic distribution, inducer performance and radial force dynamic characteristics of the inducer under the action of a flow field are obtained, the inducer is assisted to be designed and optimized, and the experiment cost and time are saved.

The invention discloses a civil airplane APU performance assessment and fault pre-warning method, and relates to the technical field of equipment state monitoring. According to the method, based on QAR data of an airplane, a health state model of an important parameter in an APU is dug out by a machine learning technology; then the predicted value of the health state model is subtracted from the actual measured value of the APU state parameter, and thus the deviation value of the state parameter is acquired; then the deviation values of the several key parameters are fused into a health indexby using a regression model, and the health index is used for representing the degradation degree of APU performance, and fault pre-warning is timely emitted according to a set warning value. The method provided by the invention is applied to the civil airplane APU, particularly the APU that is faulted due to performance degradation, the problem of being lack of airplane APU performance assessmentmeans is solved, the performance state of the APU can be accurately assessed and the APU fault can be found at the early stage and the pre-warning can be timely emitted, and unplanned issuing eventscan be reduced; meanwhile, performance assessment data is from the existing civil airplane QAR data, and the method is relatively strong in practicability and operability.

The invention relates to a deep-towing multi-channel seismic towing cable attitude determination method and a processing terminal. The method comprises the steps of 1, picking up direct wave travel time, sea surface reflected wave travel time and seabed reflected wave travel time of each detector for each shot gather; 2, obtaining the average propagation speed of seismic waves in seawater; 3, calculating the horizontal distance and the height difference of each detector relative to the seismic source; and 4, solving the three-dimensional coordinates of each detector according to the corresponding equation set to obtain the three-dimensional coordinates of each detector, thereby equivalently obtaining the attitude of the towing cable. According to the method, the three-dimensional spatial form of the towing cable can be obtained without two-dimensional hypothesis. Meanwhile, the relative position between the detector and the seismic source is determined by utilizing the travel time of the direct wave, the seabed reflected wave and the sea surface reflected wave, and the accuracy of the relative position is improved.

The invention provides a method and device for verifying network performance statistic data, relates to the field of wireless network quality, and solve the problem that accuracy of data generated randomly by an existing network cannot be found. The method for verifying the network performance statistic data comprises the following steps: obtaining the network performance statistic data in a preset time granularity; obtaining raw performance statistic data of the same-type performance indexes in multiple kinds of pre-stored raw performance statistic data, and mapping the obtained raw network performance statistic data of the same-type performance indexes into corresponding mapping results according to a pre-established mapping relationship between the raw network performance statistic data and a verification algorithm; and verifying whether the obtained network performance statistic data are consistent with the mapping results. So the accuracy of the data is known, and, in this way, accuracy of a performance index can be assessed by utilizing multiple data sources of the same-type performance indexes.

The embodiment of the invention provides a neighbor measurement control method, system and equipment. The neighbor measurement control method comprises the following steps of: judging whether a terminal is at a power-on state and has received an MDT configuration message sent by a base station, and forcing the terminal to always perform neighbor measurement in an MDT measuring process when the terminal is at a power-on state and has received the MDT configuration message sent by the base station. By adopting the technical scheme provided by the embodiment of the invention, complete neighbor measurement information can be acquired, and therefore accurate performance assessment for surrounding cells neighboring the current serving cell according to the complete neighbor measurement information in the follow-up process.

The invention relates to the technical field of mmWave multi-cell interference suppression, and discloses a mmWave multi-cell interference suppression method for minimizing base station transmitting power in combination with user association and beam width selection. The method comprises the following steps: combining a Gaussian antenna model with an mmWave dual-ray channel model as a system model, under the system model, converting the problem of minimizing the base station transmitting power into two methods of user association and optimal beam width selection, firstly, selecting a user withSINR greater than a threshold after pairing of base stations and users under a distributed framework, selecting, by the user, the base station with the smallest transmitting power among the neighboring base stations for association, and then selecting the optimal beam width by using a particle swarm optimization method. The invention designs a user association and beam width selection method, which reduces the computational complexity and reduces the inter-cell interference while ensuring that the base station transmission power is minimized.

The invention discloses an LTE (long term evolution) system performance determining method and an LTE system performance determining system. The LTE system performance determining method includes: determining configuration parameter combination corresponding to current ground feature types according to preset corresponding relation between the ground feature types and the configuration parameter combination, and initializing the determined configuration parameter combination; executing the following operations to each subframe of each radio frame in the LET system: updating positions of all users and corresponding timers in the LTE system according to the initialized configuration parameter combination; executing downlink evaluation procedures when the current subframe is a downlink subframe and a downlink is evaluate currently; and executing uplink evaluation procedures when the current subframe is an uplink subframe and an uplink is evaluated currently.

The embodiment of the invention provides a benchmark test method and device of a supervised learning algorithm in a distributed environment. The method comprises the steps that a first benchmark test result which is determined according to output data in a benchmark test is obtained; a distributed performance index in the benchmark test is obtained, and the distributed performance index is determined as a second benchmark test result; the first benchmark test result and the second benchmark test result are merged to obtain a benchmark test total result. According to a perfect scheme used for solving a benchmark test problem of the supervised learning algorithm in the distributed environment, the benchmark test method and device can assist technical staff to conduct accurate and quick assessment on the performance of the supervised learning algorithm.

The invention provides a reactor three-dimensional power probability distribution monitoring method based on an information fusion theory. The reactor three-dimensional power probability distributionmonitoring method comprises the following steps: evaluating uncertainty of measurement information; determining adopted detector information based on the monitored dynamic or steady-state requirements; estimating a reactor power distribution state based on a Bayesian method; performing probability estimation of reactor core burn-up distribution; updating the probability of reactor core burnup distribution, and estimating the performance of the fuel rod in real time. According to the method, a detection information fusion method with different angles and different precisions is explored, basedon the Bayesian framework, information loss is reduced as much as possible, the matching problem of different angles and different precision measurements is solved, spatial distribution information with higher precision is obtained, probability distribution of power values of any point in a reactor core space is provided, and the probability distribution of the burnup value of any space point in the reactor core is further provided.

The invention discloses a performance evaluation method for a universal processor. The method comprises the following steps: dividing a test program into a plurality of execution stages; a section ofinstruction sequence is selected from each execution stage to serve as a test sample, and calculating the weight; running a test program on a real processor chip platform, and generating a check pointfor each test sample by adopting a check point creation tool; recovering and executing a certain number of instructions on the simulation verification platform by using check points recovered and generated by a check point recovery tool on the test verification platform, and collecting performance data; and calculating the performance data of the whole test program according to the weight of eachtest sample and the performance data of the corresponding check point. The performance can be evaluated quickly and accurately, and various simulation verification platforms and flexible and fine-grained performance evaluation are supported.

The embodiment of the invention provides a performance evaluation method for a multi-stage task system (PMS), which comprises the following steps of: on the basis of deep learning feature representation, fusing polymorphic feature data of the PMS on the basis of Graph feature input of a PMS damage evolution mechanism; according to task stage requirements of the PMS and a causal mechanism of Graph,applying factor graph segmentation to obtain an appropriate distributed evaluation framework; based on attention operator design of state & position contrast coding, acquiring a minimum Graph neighborhood module in distributed evaluation in a self-adaptive mode; and further focusing the target neighborhood, generating a GRU adaptive deep network capable of describing PMS stage correlation, and mining neighborhood information of the minimum evaluation sub-graph. By applying the scheme provided by the embodiment of the invention, the interpretability of the corresponding deep learning module isexpected to be increased, the complexity of the evaluation algorithm is greatly reduced, and the generalization ability of the model is improved, so that the performance of the PMS is accurately evaluated.

The invention discloses a method and device for detecting the transverse control function of a vehicle ADAS system and a storage medium, and the method comprises the steps: obtaining the actual diameter of a wheel, and obtaining the transverse operation data of a vehicle in real time; the transverse operation data at least comprise wheel image diameters acquired by an image acquisition device arranged at the side part of the vehicle and image distances between the side part of the vehicle and corresponding lane lines; according to a preset proportion coefficient, the actual wheel diameter, the wheel image diameter and the image distance, the actual distance between the side portion of the vehicle and the corresponding lane line is calculated, and the proportion coefficient reflects the proportion relation between the image distance and the actual distance; and processing the actual distance, and detecting a transverse control function associated with the actual distance according to a processing result. According to the detection method and device for the transverse control function of the vehicle ADAS and the storage medium provided by the embodiment of the invention, the redundancy of data processing is reduced, and the accuracy of performance detection and analysis of the transverse control function of the vehicle is ensured.

The invention discloses a lightning early warning method based on radar data, and the method comprises the following steps: S1, detecting radar data, classifying the radar data, and removing repeated data; s2, extracting historical thunder and lightning data and thunder and lightning monitoring data of the radar monitoring area; s3, lightning simulation is carried out according to radar monitoring data and historical data; s4, performing grade judgment on the simulated thunder and lightning, and performing early warning in different modes according to different grades; s5, monitoring thunder and lightning occurrence time and thunder and lightning grades after early warning, and recording monitoring data; and S6, comparing the monitored thunder and lightning occurrence time and thunder and lightning level with early warning data, performing difference degree calculation, and evaluating early warning accuracy. Thunder and lightning early warning can be effectively performed, and evaluation of early warning accuracy is facilitated.

The invention is suitable for the technical field of equipment evaluation, and provides an efficiency evaluation system for communication navigation equipment. The system comprises a data query module, a data processing module and an efficiency evaluation module; the data query module is used for querying historical communication data of the communication navigation equipment according to the query condition to obtain evaluation data of the communication navigation equipment; the data processing module is used for grouping the evaluation data according to attributes of the evaluation data; and the efficiency evaluation module is used for inputting each group of evaluation data into a preset multi-level efficiency evaluation model for analysis to obtain an efficiency evaluation result of the communication navigation equipment. According to the system, efficiency evaluation can be carried out on the communication navigation equipment of the helicopter more objectively and accurately.

The invention discloses a multi-cycle and multi-constraint fused aviation emergency rescue efficiency evaluation method, which comprises the following steps of: firstly, disassembling an aviation emergency rescue task into a set of a plurality of OREAD cycles to obtain five indexes, then determining weights of the five indexes according to task training requirements, and according to a task completion condition, obtaining a task time consumption coefficient and a task success rate, then determining a circulation probability and a circulation difficulty coefficient according to a circulation occurrence probability and difficulty, finally obtaining evaluation values of five indexes, and calculating and obtaining a single-item efficiency evaluation result and a team comprehensive efficiency evaluation result. The task is disassembled into the set of OREAD cycles, and mapping from a task execution process to an evaluation index is realized; two constraints of a task time consumption coefficient and a task success rate are introduced, and two constraints of a cyclic probability and a cyclic difficulty coefficient are added, so that efficiency evaluation is more accurate; a single index is independent from comprehensive evaluation, two-way evaluation is achieved, and evaluation and training are more targeted.