An automatic well test interpretation method for radial composite reservoirs based on convolutional neural network(CNN) is proposed, and its effectiveness and accuracy are verified by actual field data. In this paper,...An automatic well test interpretation method for radial composite reservoirs based on convolutional neural network(CNN) is proposed, and its effectiveness and accuracy are verified by actual field data. In this paper, based on the data transformed by logarithm function and the loss function of mean square error(MSE), the optimal CNN is obtained by reducing the loss function to optimize the network with "dropout" method to avoid over fitting. The trained optimal network can be directly used to interpret the buildup or drawdown pressure data of the well in the radial composite reservoir, that is, the log-log plot of the given measured pressure variation and its derivative data are input into the network, the outputs are corresponding reservoir parameters(mobility ratio, storativity ratio, dimensionless composite radius, and dimensionless group characterizing well storage and skin effects), which realizes the automatic initial fitting of well test interpretation parameters. The method is verified with field measured data of Daqing Oilfield. The research shows that the method has high interpretation accuracy, and it is superior to the analytical method and the least square method.展开更多
Automatic speaker recognition(ASR)systems are the field of Human-machine interaction and scientists have been using feature extraction and feature matching methods to analyze and synthesize these signals.One of the mo...Automatic speaker recognition(ASR)systems are the field of Human-machine interaction and scientists have been using feature extraction and feature matching methods to analyze and synthesize these signals.One of the most commonly used methods for feature extraction is Mel Frequency Cepstral Coefficients(MFCCs).Recent researches show that MFCCs are successful in processing the voice signal with high accuracies.MFCCs represents a sequence of voice signal-specific features.This experimental analysis is proposed to distinguish Turkish speakers by extracting the MFCCs from the speech recordings.Since the human perception of sound is not linear,after the filterbank step in theMFCC method,we converted the obtained log filterbanks into decibel(dB)features-based spectrograms without applying the Discrete Cosine Transform(DCT).A new dataset was created with converted spectrogram into a 2-D array.Several learning algorithms were implementedwith a 10-fold cross-validationmethod to detect the speaker.The highest accuracy of 90.2%was achieved using Multi-layer Perceptron(MLP)with tanh activation function.The most important output of this study is the inclusion of human voice as a new feature set.展开更多
At present,artificial intelligence computing platforms are usually based on cloud hosts for services,which have the characteristics of fast training speed and a wide variety of model types.However,the online models of...At present,artificial intelligence computing platforms are usually based on cloud hosts for services,which have the characteristics of fast training speed and a wide variety of model types.However,the online models of such platforms mostly adopt the form of downloading model files,which is difficult to integrate into traditional software system systems.In response to existing problems,this paper takes the relevant theoretical technologies of next-generation intelligent computing platforms as the development framework,and conducts research on the diversity of multi-level intelligent computing requirements,by implementing a universal algorithm model construction and automatic integration mechanism;Build a multi domain and multi-level application algorithm library for different application scenarios;Design a personalized algorithm recommendation based on knowledge reasoning and object-oriented approach,and build an emerging intelligent computing platform for analyzing and understanding real-world data,meeting the needs of complex engineering application software such as heavy backend,light frontend,loose coupling,microservices,etc.,providing theoretical and technical support for innovative big data services and applications with diverse computing requirements.展开更多
A 'Human-Imitating Intelligent Control Theory' with 'generalized reduction' and 'Human Imitating' concepts as its kernel is proposed. And a world puzzlein the control circles is solved successf...A 'Human-Imitating Intelligent Control Theory' with 'generalized reduction' and 'Human Imitating' concepts as its kernel is proposed. And a world puzzlein the control circles is solved successfully based on this theory. The puzzle is thewell-known 'triple inverted pendulum control' using a SINGLE motor. A human-imitating intelligent technique to control inverted pendulum is here described. The success. ful experimental results show that our control objective can be achieved without a precise mathematical model of the plant. Finally, general principles of designing complexautomatic control systems based on the human-imitating intelligent control theory areconcluded.展开更多
The present research employs artificial intelligence to come up with an automatic solution for the modulation's classification of various radio signal varieties.As a result,the work we performed involved selecting...The present research employs artificial intelligence to come up with an automatic solution for the modulation's classification of various radio signal varieties.As a result,the work we performed involved selecting the database required for supervised deep learning,evaluating the performance of current techniques on unprocessed communication signals,and suggesting a deep learning networkbased method that would enable the classification of modulation types with the best possible ratio between computation time and accuracy.We started by examining the automatic classification models that are currently in usage.In light of the difficulty of forecasting in low Signal Noise Ratio(SNR)situations,we suggested an ensemble learning strategy based on adjusted Res Net and Transformer Neural Network,which is effective at extracting multi-scale features from the raw I/Q sequence data.Finally,we produced an architecture that is simple to use and apply to communication signals.The architecture of this solution is strong and optimal,enabling it to determine the type of modulation with up to 95%accuracy automatically.展开更多
基金Supported by the National Science and Technology Major Project(2017ZX05009005-002)
文摘An automatic well test interpretation method for radial composite reservoirs based on convolutional neural network(CNN) is proposed, and its effectiveness and accuracy are verified by actual field data. In this paper, based on the data transformed by logarithm function and the loss function of mean square error(MSE), the optimal CNN is obtained by reducing the loss function to optimize the network with "dropout" method to avoid over fitting. The trained optimal network can be directly used to interpret the buildup or drawdown pressure data of the well in the radial composite reservoir, that is, the log-log plot of the given measured pressure variation and its derivative data are input into the network, the outputs are corresponding reservoir parameters(mobility ratio, storativity ratio, dimensionless composite radius, and dimensionless group characterizing well storage and skin effects), which realizes the automatic initial fitting of well test interpretation parameters. The method is verified with field measured data of Daqing Oilfield. The research shows that the method has high interpretation accuracy, and it is superior to the analytical method and the least square method.
基金This work was supported by the GRRC program of Gyeonggi province.[GRRC-Gachon2020(B04),Development of AI-based Healthcare Devices].
文摘Automatic speaker recognition(ASR)systems are the field of Human-machine interaction and scientists have been using feature extraction and feature matching methods to analyze and synthesize these signals.One of the most commonly used methods for feature extraction is Mel Frequency Cepstral Coefficients(MFCCs).Recent researches show that MFCCs are successful in processing the voice signal with high accuracies.MFCCs represents a sequence of voice signal-specific features.This experimental analysis is proposed to distinguish Turkish speakers by extracting the MFCCs from the speech recordings.Since the human perception of sound is not linear,after the filterbank step in theMFCC method,we converted the obtained log filterbanks into decibel(dB)features-based spectrograms without applying the Discrete Cosine Transform(DCT).A new dataset was created with converted spectrogram into a 2-D array.Several learning algorithms were implementedwith a 10-fold cross-validationmethod to detect the speaker.The highest accuracy of 90.2%was achieved using Multi-layer Perceptron(MLP)with tanh activation function.The most important output of this study is the inclusion of human voice as a new feature set.
文摘At present,artificial intelligence computing platforms are usually based on cloud hosts for services,which have the characteristics of fast training speed and a wide variety of model types.However,the online models of such platforms mostly adopt the form of downloading model files,which is difficult to integrate into traditional software system systems.In response to existing problems,this paper takes the relevant theoretical technologies of next-generation intelligent computing platforms as the development framework,and conducts research on the diversity of multi-level intelligent computing requirements,by implementing a universal algorithm model construction and automatic integration mechanism;Build a multi domain and multi-level application algorithm library for different application scenarios;Design a personalized algorithm recommendation based on knowledge reasoning and object-oriented approach,and build an emerging intelligent computing platform for analyzing and understanding real-world data,meeting the needs of complex engineering application software such as heavy backend,light frontend,loose coupling,microservices,etc.,providing theoretical and technical support for innovative big data services and applications with diverse computing requirements.
文摘A 'Human-Imitating Intelligent Control Theory' with 'generalized reduction' and 'Human Imitating' concepts as its kernel is proposed. And a world puzzlein the control circles is solved successfully based on this theory. The puzzle is thewell-known 'triple inverted pendulum control' using a SINGLE motor. A human-imitating intelligent technique to control inverted pendulum is here described. The success. ful experimental results show that our control objective can be achieved without a precise mathematical model of the plant. Finally, general principles of designing complexautomatic control systems based on the human-imitating intelligent control theory areconcluded.
文摘The present research employs artificial intelligence to come up with an automatic solution for the modulation's classification of various radio signal varieties.As a result,the work we performed involved selecting the database required for supervised deep learning,evaluating the performance of current techniques on unprocessed communication signals,and suggesting a deep learning networkbased method that would enable the classification of modulation types with the best possible ratio between computation time and accuracy.We started by examining the automatic classification models that are currently in usage.In light of the difficulty of forecasting in low Signal Noise Ratio(SNR)situations,we suggested an ensemble learning strategy based on adjusted Res Net and Transformer Neural Network,which is effective at extracting multi-scale features from the raw I/Q sequence data.Finally,we produced an architecture that is simple to use and apply to communication signals.The architecture of this solution is strong and optimal,enabling it to determine the type of modulation with up to 95%accuracy automatically.
