245 results about "Exponential smoothing" patented technology

A control method for fire-fighting false alarm of an energy storage system comprises the following steps of: calculating a fluctuation period, a derivative and a probability density function by using fast Fourier transform and the like by acquiring data streams such as current, voltage, power and heat energy of the energy storage system, determining a system state and predicting a related trend. And predicting power by using an autoregressive moving average algorithm and a support vector machine, and setting a threshold value according to a probability density function to judge abnormity. For heat energy data, through derivative calculation and signal analysis, states are identified and early warning is carried out in combination with models such as a deep convolutional adversarial neural network and a support vector machine. The normal range of monitoring data is determined by means of the long-short-term memory neural network, the change trend and physical quantity relation of abnormal data are analyzed and predicted through a cubic exponential smoothing method and Gaussian process regression, potential risks are judged through weighted fusion and a self-adaptive Simpson algorithm, fire-fighting false alarms are effectively reduced, fire risks are accurately recognized, and the safety of an energy storage system is guaranteed.

The invention discloses a temperature collecting, monitoring and alarming method for palm fire fighting, relates to the technical field of temperature monitoring for fire fighting, and is used for solving the problem of inaccurate temperature monitoring and prediction. According to the method, environmental factor data are synchronously collected through temperature monitoring points, the correlation between each factor and the temperature is calculated after denoising, multi-factor characteristics are extracted to generate an output characteristic vector of temperature prediction, the output characteristic vector is constructed into a time sequence matrix to be input into a long-short-term memory network, and temperature trend characteristics are extracted; the method comprises the following steps: predicting a future trend and comparing the future trend with a safety threshold to judge whether early warning is triggered or not, meanwhile, identifying temperature changes of different frequencies by adopting fast Fourier transform, a weighted moving average method and an exponential smoothing method through short-term, medium-term and long-term scale temperature trend analysis, setting a three-level early warning mechanism and dynamically adjusting the threshold, and finally, carrying out early warning. Based on early warning feedback data, a Q-learning algorithm is adopted to optimize a threshold value and an alarm strategy, a self-adaptive learning closed loop is formed, and the early fire early warning capability is improved.

The invention relates to the technical field of data processing and analysis, in particular to a temperature monitoring method and system for a # imgabs0 # vaporizer. The method comprises the following steps: calculating an optimal coefficient; according to the optimal coefficient, using an exponential smoothing method to predict the temperature; the step of calculating the optimal coefficient comprises the following steps: acquiring temperature data at a historical moment, and selecting a preset number of data to construct an adjacent data set at the current moment; calculating a local fluctuation degree and a trend influence factor at the current moment according to the temperature data in the adjacent data set, and taking a product of the local fluctuation degree and the trend influence factor as an adjustment coefficient; and setting an initial smoothing coefficient, and adjusting the initial smoothing coefficient through the adjusting coefficient to obtain an optimal coefficient. The method has the effect of improving the reaction capacity of the vaporizer to the real-time temperature change in the temperature prediction process.

The invention relates to the technical field of distributed data processing, in particular to a hydraulic engineering distributed flood prevention prediction system and method.The system comprises a real-time monitoring module for calculating the mean value and variance of data of a water level sensor every five minutes, and an anomaly detection module for comparing the mean value difference with a threshold value, recognizing a sudden change point and sending the sudden change point to a cloud server; the window adjustment module adjusts the collection frequency according to the abnormal result, the trend analysis module predicts the flood trend by applying moving average and exponential smoothing, and the early warning generation module evaluates the water level and generates an early warning signal. According to the method, the mean value and the variance of the water level monitoring data are calculated in real time and compared with the threshold values, so that the recognition capability of the abrupt change point is improved, the data collection frequency and the window size are adjusted according to the anomaly detection result, the adaptability and efficiency of data collection are enhanced, and the water level trend is predicted through the moving average and exponential smoothing technology; and in combination with a geographic information system, the affected area is accurately evaluated, so that the early warning accuracy and the decision support effectiveness are remarkably improved.

The invention relates to the technical field of pavement disease automatic detection, and particularly discloses a pavement disease automatic detection method and system based on deep learning, and the system comprises a municipal pavement data collection module, a pavement image quality analysis module, a pavement image re-collection module, a pavement disease evaluation feedback module and a database. Flight parameters (speed and height) are dynamically adjusted through an environmental quality index (EQI), the image acquisition quality is ensured, the detection precision in a complex environment is improved, a multi-disease coupling risk model is constructed, region weight and traffic volume dynamic weight are integrated, a disease risk level is output, and finally a disease development trend is predicted through a one-time exponential smoothing method. The system achieves the dual functions of current situation evaluation and trend early warning, predicts the disease evolution trend based on historical data, remarkably improves the scientificity and timeliness of municipal road risk management, reduces the maintenance cost, and prevents safety accidents.

The invention discloses an enterprise electricity consumption analysis and prediction method based on data mining, and relates to the technical field of DC power supply monitoring, historical electricity consumption data is acquired, and electricity consumption characteristic analysis and mining are performed from multiple directions according to the historical electricity consumption data to obtain electricity consumption rule information and seasonal change characteristics; analyzing the scale and speed increase of the agent electricity purchasing users and the annual change trend of the agent electricity purchasing quantity to obtain development and operation information of agent electricity purchasing; based on historical electric quantity data, constructing exponential smoothing, ARIMA, SARIMA and LSTM neural network multiple time sequence prediction models, and performing comparison to select an optimal model; through accurate power consumption prediction, the enterprise is helped to optimize power resource configuration, plan power purchase work in advance and reduce the power cost of the enterprise, and especially under the condition that the power market fluctuates greatly, the enterprise can better understand the market dynamic condition and formulate a corresponding power purchase strategy by predicting the scale and speed increase of the proxy power purchase user.

The invention discloses a backup task scheduling method based on load prediction and an improved SAC algorithm, and relates to the technical field of data backup and intelligent scheduling. Aiming at the problems of response lagging, insufficient key task guarantee and low strategy exploration efficiency of a traditional scheduling method under a dynamic load, a task weight monitoring matrix is constructed, server performance, a network state and data importance are fused, and a task risk coefficient is calculated in combination with an exponential smoothing algorithm; carrying out sequence decomposition on a service load by adopting an Autoformer model, identifying a periodic mode through a weighted autocorrelation mechanism, and predicting a future load change trend; on this basis, an improved SAC scheduling framework is designed, a task weight guidance and temperature coefficient adaptive adjustment mechanism is introduced, and the task starting opportunity and execution duration are optimized in combination with an elastic time window. The invention provides an intelligent scheduling method with strong adaptive capability, which can autonomously optimize the task scheduling strategy according to the dynamic change of the system and effectively reduce the peak load of the server and the task backlog risk.

The invention is suitable for the technical field of big data, and provides an enterprise business income prediction method based on multiple time sequence models, which comprises the steps of data collection, data preprocessing, feature engineering, model selection and training, model evaluation optimization, model fusion prediction and result display. Wherein the models are five models including an XGBoost model, an SARIMAX model, a Prophet model, an LSTM model and an exponential smoothing model, the performance index of each model is calculated, the corresponding weight is distributed, and finally weighted average calculation is carried out to obtain the final prediction result of the business income of the enterprise. Through model fusion and optimization, the method has the characteristics of robustness enhancement, stability improvement, performance optimization and the like, the advantages of each model are fully utilized, the prediction accuracy can be remarkably improved, and reliable data support is provided for decision makers to make decisions.

The invention relates to the technical field of wireless communication, in particular to a network security communication method and system applied to an unmanned equipment formation and a medium. The method comprises the following steps: firstly, according to a wind speed, a wind direction and a difference condition between an actual spatial position and a set spatial position of unmanned equipment, obtaining a self-noise representation degree at a to-be-analyzed moment; according to the noise interference degree, the initial smoothing coefficient of the to-be-analyzed moment is adjusted, and the adjusted smoothing coefficient of the unmanned equipment at the to-be-analyzed moment is obtained; filtering the actual spatial position by using the adjusted smoothing coefficient at each sampling moment to obtain filtered actual spatial position data; and performing communication transmission based on the filtered actual spatial position data. According to the method, the smoothing coefficient of each sampling moment is reasonably set, and exponential smoothing filtering is performed on the actual spatial position data of the unmanned equipment, so that the filtering effect is improved, the filtering data accuracy in wireless communication is improved, and the wireless communication safety is guaranteed.

The invention discloses an intelligent campus management system and method, and the method comprises the following steps: S1, collecting campus management data, and transmitting the campus management data to a cloud computing platform; s2, preprocessing the campus management data; s3, analyzing data by using an improved gating cycle unit and an exponential smoothing method; s4, performing curriculum scheduling optimization and classroom and laboratory resource allocation by adopting a super-gravitational search optimization algorithm; s5, automatically adjusting the air conditioner, the illumination, the access control authority and the monitoring strategy; s6, real-time data transmission is carried out, and the equipment response speed is optimized based on distributed calculation; s7, constructing a data sharing platform based on a block chain technology, and setting a data access permission through an intelligent contract; and S8, optimizing the campus management scheme based on reinforcement learning. The intelligent campus scheduling optimization and trend prediction system integrates the Internet of Things, artificial intelligence, a super gravitational search optimization algorithm, an improved gating loop unit and an exponential smoothing method, realizes intelligent campus scheduling optimization and trend prediction, and has the advantages of efficient optimization and intelligent management.

The invention discloses an electricity-gas integrated energy system distributed dynamic state estimation method based on long short-term memory neural network load prediction. According to the method, a Kalman filtering algorithm is taken as a framework, a state space model of a gas network is established based on a finite difference model, and a state space model of a power grid is established based on a two-parameter exponential smoothing method; predicting the gas network load node flow based on the long short-term memory neural network, and providing flow prediction information for the gas network prediction step; and a distributed estimation algorithm is adopted, and the filtering steps of the power grid and the gas grid are iteratively corrected through information interaction of the boundary coupling elements, so that a collaborative state estimation result of the electricity-gas interconnected integrated energy system is obtained. According to the method, the flow information of the gas network load node can be accurately predicted, the operation state of the electricity-gas interconnection integrated energy system can be dynamically tracked in real time, the information barrier problem between the electricity subsystem and the gas subsystem can be effectively solved, and support is provided for dynamic monitoring and operation control of the electricity-gas interconnection integrated energy system.

The invention relates to the technical field of data communication, and discloses an intelligent agent high-speed bus implementation method and system based on an FPGA and a dynamic smoothing technology. The method comprises the following steps: constructing a single handshake message through a lightweight protocol, adjusting network indexes by adopting an exponential smoothing method, constructing a pipeline processing unit in an FPGA card by utilizing a VHDL, realizing key updating and encryption processing, constructing a three-layer adaptive structure connection subsystem, and optimizing algorithm parameters and resource allocation according to a running state. And a high-performance bus communication system is formed. On the premise of ensuring the communication security and reliability, the network transmission delay is remarkably reduced, and the data throughput is improved, so that the communication system can meet the requirements of emergency linkage and large data volume transmission in a scene (such as a rail transit station-level system) with a high real-time requirement.

The invention discloses a bird sound recognition method based on multi-model integration. The method comprises the following steps: (1) obtaining a bird sound signal and preprocessing data; (2) extracting key features from the bird sound data by combining Mel-frequency cepstrum coefficient, short-time Fourier transform and decibel conversion feature extraction methods; (3) a plurality of improved pre-training convolutional neural network models are used for training; each convolutional neural network model has different structures or parameter settings so as to capture different features of bird sounds; (4) designing an integration module by using an exponential smoothing method and a weighted smoothing method, and integrating a plurality of trained convolutional neural network models; output results of all the models are fused; (5) carrying out comprehensive performance evaluation on the integrated model; according to the method, the overall accuracy of bird sound recognition is remarkably improved.

The invention discloses a reservoir landslide displacement interval prediction method fusing BiLSTM and Transformer, and the method comprises the following steps: obtaining landslide multi-source monitoring data, and reconstructing displacement time series data into trend term displacement and periodic term displacement through employing empirical mode decomposition; performing interval prediction on trend term displacement by adopting a Holt two-parameter exponential smoothing method and a kernel density estimation method; after a trend term displacement point prediction result is obtained, interval prediction is carried out by adopting a kernel density estimation method, and a trend term displacement interval prediction result is obtained; a BiLSTM-Transformer model and a kernel density estimation method are adopted to carry out interval prediction on the displacement of the periodic term; and finally, superposing the trend term displacement interval prediction result and the periodic term displacement interval prediction result to obtain an accumulated displacement interval prediction result. According to the method, the accuracy and the efficiency of slope displacement prediction are improved by capturing the change before and after the time sequence data through the BiLSTM, a Transform multi-head attention mechanism and parallel calculation.

The invention relates to the technical field of data processing, in particular to a smart city building fire monitoring and early warning system based on multi-modal data, and the system comprises a data collection module which is used for obtaining a temperature data sequence and a smoke data sequence in any building; the data analysis module is used for constructing a self-adaptive window according to data fluctuation in the smoke data sequence, and obtaining a trend value according to a data change trend in the self-adaptive window; the threshold acquisition module is used for obtaining a predicted value of a future moment according to the data in the adaptive window by using an exponential smoothing algorithm, and adjusting a preset initial early warning value according to a difference between an actual value and the predicted value of the future moment and a trend value to obtain an adaptive early warning value of the future moment; and the abnormity early warning module is used for carrying out fire early warning on the building according to the temperature data corresponding to each piece of smoke data in the adaptive window if the actual value at the future moment is greater than the adaptive early warning value, so that the accuracy of fire early warning on the building is improved.

The invention discloses a method for predicting the remaining service life of a proton exchange membrane fuel cell based on an improved HHO, and the method comprises the following steps: 1), building a PEMFC original data set, and carrying out the data denoising processing through employing a wavelet adaptive threshold (WAD); 2) performing smoothing processing on the data by using an exponential smoothing method (ES) and performing secondary denoising; 3) constructing a Harris eagle algorithm (HHO) model based on improved Logistic chaotic mapping and arc tangent decline strategy optimization so as to improve the optimization performance; 4) using the improved HHO to optimize the weight and bias terms of the long short-term memory network (LSTM), and using a mean square error (MSE) to find out the optimal weight and bias parameters for prediction; and 5) carrying out secondary prediction on the prediction result by using a self-attention mechanism (Self-Attention) to obtain a final prediction result. The feasibility and superiority of the method are proved, the iteration speed of the model can be greatly improved, the training time is shortened, and the accuracy is greatly improved.

The invention relates to the technical field of industrial automation control, and discloses a chemical production-oriented multi-stage cooperative intelligent control and equipment operation and maintenance data management platform, which comprises a multi-source data acquisition module for acquiring multi-source heterogeneous data and performing time alignment processing and quality verification; the working condition identification and feature extraction module is used for obtaining equipment operation state data by adopting weighted fusion and principal component analysis; the equipment health assessment module is used for calculating an equipment health degree score and predicting a degradation trend through exponential smoothing; the self-adaptive control module is used for designing a multivariable coordination controller and dynamically generating operation constraints of the multivariable coordination controller according to the equipment health degree score; the layered optimization control module is used for establishing an interlayer feedback mechanism; the collaborative decision-making module generates a collaborative decision-making trigger signal according to the equipment health degree information and generates a collaborative decision-making scheme; according to the invention, by establishing a collaborative decision-making mechanism, a decision-making scheme considering the production task completion degree, the equipment health degree and the maintenance cost is generated.

The invention belongs to the technical field of data processing, and particularly relates to a power supply load prediction method and system for a new energy vehicle, and the method comprises the steps: obtaining real-time operation parameters and a historical power supply load data set of the new energy vehicle; for any moment, determining the load fluctuation sensitivity according to the offset of the power supply load relative to the historical power supply load in the same period and the difference of the power supply load relative to the previous moment; determining the load demand potential by combining the difference between the residual electric quantity of the on-line controlled vehicle and the environment temperature; and determining a prediction weight based on the two indexes, extracting a long or short period data set in combination with a magnitude relationship between the prediction weight and a preset weight threshold, and realizing prediction by using an exponential smoothing method. According to the method, the problems of response lag and insufficient precision of a traditional prediction algorithm when a charging scene is highly dynamic are effectively solved, and high-precision real-time tracking of unsteady state load characteristics is realized.

The invention discloses an electric charge real-time settlement and intelligent deduction method for an intelligent electric meter, and belongs to the technical field of intelligent power grids, and the method comprises the steps: obtaining current waveform data through high-frequency sampling, generating a carrying feature vector through differential exponential smoothing, and achieving the dynamic classification of equipment load types based on the spectral characteristics and a lightweight convolutional neural network. Generating a dynamic rate parameter in combination with the power grid peak regulation demand index and the historical credit score of the user, and constructing a mixed charging bill; a cross-chain transaction message is formed by analyzing the hash value and embedding the hash value into a block chain timestamp, and a three-section settlement smart contract is activated; and monitoring load fluctuation data after fee deduction, generating an abnormal fluctuation mark, and performing optimization through federated learning. According to the method, comprehensive optimization of high-precision real-time charging, tamper-proof trusted transaction and adaptive anomaly monitoring is realized, and the peak regulation efficiency and the power utilization fairness of the power grid are remarkably improved.

The invention relates to a humanoid robot whole body motion control method and system based on guide learning and remapping data. According to the method, guide learning is adopted as a motion control algorithm framework of the humanoid robot; the method comprises the following steps: firstly, acquiring human body weight mapping data, and performing coordinate transformation on the human body weight mapping data to serve as a feedforward action in guide learning; constructing a neural network, inputting an observation vector of the robot into the neural network, carrying out reinforcement learning on the neural network, outputting rotation angle data, needing to be finely adjusted, of each joint of the robot, carrying out exponential smoothing processing on the rotation angle data, and outputting a smoothed signal as a feedback signal; and the feedforward action and the feedback signal are combined through PD control to generate a final rotation angle required by each joint of the robot, and then the rotation angle is converted into an output torque of a motor and output, so that whole-body motion control of the humanoid robot is completed. Compared with the prior art, the method has the advantages that the action accuracy and the real-time fine adjustment requirement are both considered, and the control robustness is high.

The embodiment of the invention provides an end-to-end speech synthesis method, device and equipment based on multi-modal emotion fusion. The method is applied to the technical field of speech synthesis. Extracting acoustic features through an acoustic encoder; the interaction weights of the text features and the acoustic features are dynamically distributed by using a double attention mechanism, and the collaboration of the emotion information and the voice information is enhanced; generating an emotion intensity curve by using a bidirectional long-short-term memory network, a time convolutional network, a self-attention mechanism and an exponential smoothing technology; inputting the fused features and the emotion intensity curve into a variance adapter for feature enhancement and expansion, and outputting the fused features and the emotion intensity curve to a Mel decoder for parallel decoding to obtain a Mel spectrum of the synthesized speech; the vocoder is used to map the Mel spectrum features into the sound waveform, and the final emotional voice is generated, so that the emotional expressive force and naturalness of the generated voice are improved, and the emotional intensity can be adjusted according to different application scenes.

The invention discloses a maximum correlation entropy adaptive dynamic estimation method containing a distributed photovoltaic power distribution network, and belongs to the technical field of power systems. The invention discloses a maximum correlation entropy adaptive dynamic estimation method for a distributed photovoltaic power distribution network, and the method comprises the steps: building a two-stage three-phase distributed photovoltaic physical model containing a feedback control link, and forming a photovoltaic system state equation containing a differential link; establishing a power distribution network state equation based on a Holt's two-parameter exponential smoothing method; the kernel width parameter of the maximum correlation entropy is adjusted in real time according to the measurement error statistical characteristics; a maximum correlation entropy criterion is adopted as a target function, and a robust Kalman gain updating formula is obtained in combination with a Gaussian kernel function and an adaptive weight matrix; and integrating the robust Kalman gain updating formula into the integral framework of the volume Kalman filtering to complete the dynamic state estimation of the power distribution network. By adopting the maximum correlation entropy adaptive dynamic estimation method containing the distributed photovoltaic power distribution network, the problem that an existing dynamic state estimation method is insufficient in estimation precision in the distributed photovoltaic and non-Gaussian noise power distribution network can be solved.

The invention relates to the technical field of data monitoring, in particular to a method and system for monitoring the running state of a numerical control machine tool for machining hardware. The method comprises the following steps: forming a first data set by a plurality of similar historical hardware parameter data time sequence sequences of hardware, and forming a second data set by a parameter data time sequence sequence of the hardware and a similar historical hardware parameter data time sequence sequence; calculating the deviation degree of the hardware; recording an area defined by the fitting curve of the parameter data time sequence and the coordinate axis as a first feature of the hardware corresponding to the parameter data time sequence; calculating the machining stability of the hardware; weighting the deviation degree of the hardware through the machining stability of the hardware to obtain a smoothness parameter of the hardware; the smoothing parameters of the hardware are used in the exponential smoothing algorithm, the numerical control machine tool operation state monitoring result corresponding to the hardware is obtained, and the accuracy of numerical control machine tool operation state monitoring is effectively improved.

The invention discloses a low-voltage reactive power compensation method and system for a power distribution device. The method comprises the steps that S10, reactive load data of all phases of the low-voltage side of the power distribution device are collected; s20, constructing a time sequence model of the reactive load of each phase of the low-voltage side of the power distribution device based on the reactive load data of each phase of a plurality of time nodes; s30, on the basis of the time sequence model, reactive load data of each phase of the next time node is predicted through exponential smoothing; and S40, based on the reactive load data of each phase of the next time node, determining the grouping switching condition of the reactive compensation capacitor of each phase of the power distribution device. According to the method, the reactive load data in the future is predicted based on the existing data through the exponential smoothing method, so that the adaptability to load fluctuation is improved, the response to emergency situations is enhanced, and the timeliness is higher.

The invention relates to a real-time overrun early warning method based on gas concentration characteristic change, and belongs to the technical field of coal mine safety monitoring and early warning. The method comprises the following steps: collecting gas concentration time sequence data in real time through gas sensor networks deployed on a working face and a tunneling face; grading concentration intervals, wherein the concentration intervals comprise a threshold interval, an alarm color and an alarm score; the method comprises the following steps of: preprocessing collected gas concentration time sequence data, and then respectively calculating three alarm scores by adopting an amplification change bandwidth judgment method, an exponential smoothing difference method and a sliding window Z score method to correspond to alarm scores in concentration interval grading; voting fusion: at the same sampling moment, synchronously aligning the calculated three alarm scores, and then determining final early warning information according to a voting fusion rule; and outputting early warning information in a visualized manner. According to the method, the concentration interval dynamic grading and voting fusion strategy is combined, refined recognition and grading early warning of the gas risk situation are achieved, and the timeliness and accuracy of alarm are remarkably improved.

The invention discloses an AUV (Autonomous Underwater Vehicle) pose detection method and system based on underwater light vision, a medium and a program product, and belongs to the field of underwater robots. The method comprises the following steps: judging the turbidity of a water body by analyzing the blue channel characteristics of an image, dynamically adjusting the graying weight, and generating a preprocessed image by adopting blue channel enhancement; constructing triple constraint conditions of contour number, spacing and area, and combining a variable step size strategy to realize threshold adaptive optimization; wick pixel coordinates are extracted through contour circularity screening, a leak detection grading mechanism is established, and feature point compensation is performed based on exponential smoothing filtering prediction and historical frame data; and finally, performing P3P pose calculation according to confidence levels. The method can be suitable for various water areas, blue light source feature extraction and feature point matching tasks are well achieved, and the stability and accuracy of AUV pose detection are improved.

The invention discloses an online car-hailing driver fatigue driving pre-judgment method and system. The invention relates to the technical field of online car hailing. And collecting driver departure data O and driver order receiving data M which are arranged according to a time sequence. O = [O1, O2,..., ON]; m = [M1, M2,..., MN]; o1, O2,..., ON and M1, M2,..., MN are historical data of driver departure data O and driver order receiving data M respectively. Executing exponential smoothing according to the trend and the seasonality of the driver departure data O and the driver order receiving data M in the time sequence; by adopting the Holt-Wittes seasonal model, the method can accurately capture the periodicity and tendency characteristics of the driving and order receiving behaviors of the driver, and significantly improves the prediction accuracy. The automatic data processing and prediction process greatly shortens the prediction period, improves the operation efficiency, and enables the platform to respond to the potential risk of fatigue driving of the driver more quickly.

The invention provides an optical access network PON port fault prediction and operation and maintenance support method and system, and the method comprises the steps: collecting real-time multi-index sequences of optical power, bit error rate, temperature, frame loss rate, business load and the like, automatically extracting a long-term trend based on variational mode decomposition, constructing a residual sequence, generating a state transition map through employing a Markov transition field algorithm, and carrying out the fault prediction and operation and maintenance support of the PON port. And future operation state distribution prediction is realized through the LSTM neural network. And after an abnormal threshold is dynamically adjusted in combination with quantile regression, a time sequence prediction result and a threshold response are fused to output a comprehensive fault probability score, and an early warning signal is triggered through exponential smoothing and continuity check, so that the progressive fault detection accuracy and advancement can be improved, and an efficient and steady prediction and early warning basis is provided for intelligent operation and maintenance of the PON port.

The invention relates to a bearing residual life prediction model training method and device, a medium and a prediction method, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring a bearing vibration signal; carrying out adaptive frequency processing on the bearing vibration signal to obtain a fusion feature; performing exponential smoothing processing on the bearing vibration signal to obtain exponential smoothing output; performing haar wavelet feature enhancement processing on the bearing vibration signal to obtain haar wavelet feature enhancement output; dividing the bearing vibration signals subjected to adaptive frequency processing, exponential smoothing processing and Haar wavelet feature enhancement processing to obtain a training set; and training the initial bearing residual life prediction model by using the obtained training set to obtain a target bearing life prediction model. The model trained by the training method can predict the residual life of the bearing more accurately.

The invention discloses a rotary table angle data filtering method based on an adaptive cut-off frequency filter, relates to the technical field of signal data filtering, and aims to solve the problem of numerical value jumping caused by high-frequency noise interference on rotary table angle data. The method comprises the following steps: firstly, setting a minimum cut-off frequency and a speed coefficient for an adaptive cut-off frequency filter, obtaining rotary table angle data, calculating the change speed of filtered standard data, and dynamically adjusting the cut-off frequency of the filter; filtering the data for multiple times by using an exponential smoothing formula; and finally, updating the dynamic cut-off frequency according to a filtering result to obtain filter parameters. The filter is a low-pass filter, has the advantages of being easy and convenient to adjust parameters, high in adaptability and high in calculation efficiency, can effectively suppress noise, and meanwhile keeps sensitivity to fast change data. Experiments prove that the method has good performance in different speed data processing, and is suitable for an inertial navigation equipment turntable system.