There are two kinds of unbalance vibrations—force vibration and displacement vibration due to the existence of unbalance excitation in active magnetic bearings(AMB)system.And two unbalance compensation methods—close...There are two kinds of unbalance vibrations—force vibration and displacement vibration due to the existence of unbalance excitation in active magnetic bearings(AMB)system.And two unbalance compensation methods—closed-loop feedback and open loop feed-forward are presented to reduce the force vibration.The transfer function order of the control system directly influencing the system stability will be increased when the closed-loop method is adopted,which makes the real-time compensation not easily achieved.While the open loop method would not increase the primary transfer function order,it provides conditions for real-time compensation.But the real-time compensation signals are not easy to be obtained in the open loop method.To implement real-time force compensation,a new method is proposed to reduce the force vibration caused by the rotor unbalance on the basis of AMB active control.The method realizes real-time and on-line force auto-compensation based on H∞controller and one novel feed-forward compensation controller,which makes the rotor rotate around its inertia axis.The time-variable feed-forward compensatory signal is provided by a modified adaptive variable step-size least mean square(VSLMS)algorithm.And the relevant least mean square(LMS)algorithm parameters are used to solve the H∞controller weighting functions.The simulation of the new method to compensate some frequency-variable and sinusoidal signals is completed by MATLAB programming,and real-time compensation is implemented in the actual AMB experimental system.The simulation and experiment results show that the compensation scheme can improve the robust stability and the anti-interference ability of the whole AMB system by using H∞controller to achieve close-loop control,and then real-time force unbalance compensation is implemented.The proposed research provides a new control strategy containing real-time algorithm and H∞controller for the force compensation of AMB system.And the stability of the control system is finally improved.展开更多
Since the dead zone phenomenon occurs in electro-hydraulic servo system, the output of the system corresponding to a sinusoidal input contains higher harmonic besides the fundamental input, which causes harmonic disto...Since the dead zone phenomenon occurs in electro-hydraulic servo system, the output of the system corresponding to a sinusoidal input contains higher harmonic besides the fundamental input, which causes harmonic distortion of the output signal. The method for harmonic cancellation based on adaptive filter is proposed. The task is accomplished by generating reference signals with frequency that should be eliminated from the output. The reference inputs are weighted by the adaptive filter in such a way that it closely matches the harmonic. The output of the adaptive filter is a harmonic replica and is injected to the fundamental signal such that the output harmonic is cancelled leaving the desired signal alone, and the total harmonic distortion （THD） is greatly reduced. The weights of filter are adjusted on-line according to the control error by using least-mean-square （LMS） algorithm. Simulation results performed with a hydraulic system demonstrate the efficiency and validity of the proposed adaptive feed-forward compensator （AFC） control scheme展开更多
A dynamic velocity feed-forward compensation （RBF-NN） dynamic model identification was presented for control （DVFCC） approach with RBF neural network the adaptive trajectory tracking of industrial robots. The prop...A dynamic velocity feed-forward compensation （RBF-NN） dynamic model identification was presented for control （DVFCC） approach with RBF neural network the adaptive trajectory tracking of industrial robots. The proposed control approach combined the advantages of traditional feedback closed-loop position control and computed torque control based on inverse dynamic model. The feed-forward compensator used a nominal robot dynamics as accurate dynamic model and on-line identification with RBF-NN as uncertain part to improve dynamic modeling accu- racy. The proposed compensation was applied as velocity feed-forward by an inverse velocity controller that can con- vert torque signal into velocity in the standard industrial controller. Then, the need for a torque control interface was avoided in the real-time dynamic control of industrial robot. The simulations and experiments were carried out on a gas cutting manipulator. The results show that the proposed control approach can reduce steady-state error, suppress overshoot and enhance tracking accuracy and efficiency in joint space and Cartesian space, especially under high- speed condition.展开更多
In order to improve the steady state performance,dynamic response and power factor of traditional power factor correction(PFC)digital control method and reduce the harmonic distortion of input current,a double closed ...In order to improve the steady state performance,dynamic response and power factor of traditional power factor correction(PFC)digital control method and reduce the harmonic distortion of input current,a double closed loop active power factorcorrection(APFC)control method with feed-forward is proposed.Firstly,the small signal model of Boost PFC control systemis built and the system transfer function is deduced,and then the parameters of the main device with Boost topology is estimated.By means of the feed-forward,the system can quickly respond to the change in input voltage.Furthermore,the use ofvoltage loop and current loop can achieve input current and output voltage regulation Simulink modeling shows that this methodcan effectively control the output voltage in case of input voltage largely fluctuating,improve the system dynamic response abilityand input power factor,and reduce the input current harmonic distortion展开更多
This paper describes the application of principal component analysis (PCA) and artificial neural network (ANN) to predict the air pollutant index (API) within the seven selected Malaysian air monitoring stations in th...This paper describes the application of principal component analysis (PCA) and artificial neural network (ANN) to predict the air pollutant index (API) within the seven selected Malaysian air monitoring stations in the southern region of Peninsular Malaysia based on seven years database (2005-2011). Feed-forward ANN was used as a prediction method. The feed-forward ANN analysis demonstrated that the rotated principal component scores (RPCs) were the best input parameters to predict API. From the 4 RPCs, only 10 (CO, O3, PM10, NO2, CH4, NmHC, THC, wind direction, humidity and ambient temp) out of 12 prediction variables were the most significant parameters to predict API. The results proved that the ANN method can be applied successfully as tools for decision making and problem solving for better atmospheric management.展开更多
In data stream management systems (DSMSs), how to maintain the quality of queries is a difficult problem because both the processing cost and data arrival rates are highly unpredictable. When the system is overloaded,...In data stream management systems (DSMSs), how to maintain the quality of queries is a difficult problem because both the processing cost and data arrival rates are highly unpredictable. When the system is overloaded, quality degrades significantly and thus load shedding becomes necessary. Unlike processing overloading in the general way which is only by a feedback control (FB) loop to obtain a good and stable performance over data streams, a feedback plus feed-forward control (FFC) strategy is introduced in DSMSs, which have a good quality of service (QoS) in the aspects of miss ratio and processing delay. In this paper, a quality adaptation framework is proposed, in which the control-theory-based techniques are leveraged to adjust the application behavior with the considerations of the current system status. Compared to previous solutions, the FFC strategy achieves a good quality with a waste of fewer resources.展开更多
Feed-forward gene transcriptional regulatory networks, as a set of common signal motifs, are widely distributed in the biological systems. In this paper, the noise characteristics and propagation mechanism of various ...Feed-forward gene transcriptional regulatory networks, as a set of common signal motifs, are widely distributed in the biological systems. In this paper, the noise characteristics and propagation mechanism of various feed-forward gene transcriptional regulatory loops are investigated, including （i） coherent feed-forward loops with AND-gate, （ii） coherent feed-forward loops with OR-gate logic, and （iii） incoherent feed-forward loops with AND-gate logic. By introducing logarithmic gain coefficient and using linear noise approximation, the theoretical formulas of noise decomposition are derived and the theoretical results are verified by Gillespie simulation. From the theoretical and numerical results of noise decomposition algorithm, three general characteristics about noise transmission in these different kinds of feed-forward loops are observed, i） The two-step noise propagation of upstream factor is negative in the incoherent feed-forward loops with AND-gate logic, that is, upstream factor can indirectly suppress the noise of downstream factors, ii） The one-step propagation noise of upstream factor is non-monotonic in the coherent feed-forward loops with OR-gate logic, iii） When the branch of the feed-forward loop is negatively controlled, the total noise of the downstream factor monotonically increases for each of all feed-forward loops. These findings are robust to variations of model parameters. These observations reveal the universal rules of noise propagation in the feed-forward loops, and may contribute to our understanding of design principle of gene circuits.展开更多
The coagulation bath system of carbon fiber precursor is a complicated and multivariable coupling system. Based on the model of industrial production,the full dynamic decoupling control of the coagulation bath system ...The coagulation bath system of carbon fiber precursor is a complicated and multivariable coupling system. Based on the model of industrial production,the full dynamic decoupling control of the coagulation bath system of carbon fiber precursor is achieved in combination with multivariable feed-forward-like decoupling and proportional-integral-differential( PID) control. Compared with the conventional PID decoupling control,the experiment results show that the proposed method has a better control effect. The use of the controller can achieve complete decoupling of three parameters from coagulation bath system. The method should have great applications.展开更多
In the oil industry, the productivity of oil wells depends on the performance of the sub-surface equipment system. These systems often have problems stemming from sand, corrosion, internal pressure variation, or other...In the oil industry, the productivity of oil wells depends on the performance of the sub-surface equipment system. These systems often have problems stemming from sand, corrosion, internal pressure variation, or other factors. In order to ensure high equipment performance and avoid high-cost losses, it is essential to identify the source of possible failures in the early stage. However, this requires additional maintenance fees and human power. Moreover, the losses caused by these problems may lead to interruptions in the whole production process. In order to minimize maintenance costs, in this paper, we introduce a model for predicting equipment failure based on processing the historical data collected from multiple sensors. The state of the system is predicted by a Feed-Forward Neural Network (FFNN) with an SGD and Backpropagation algorithm is applied in the training process. Our model’s primary goal is to identify potential malfunctions at an early stage to ensure the production process’ continued high performance. We also evaluated the effectiveness of our model against other solutions currently available in the industry. The results of our study show that the FFNN can attain an accuracy score of 97% on the given dataset, which exceeds the performance of the models provided.展开更多
A kind of second order algorithm--recursive approximate Newton algorithm was given by Karayiannis. The algorithm was simplified when it was formulated. Especially, the simplification to matrix Hessian was very reluct...A kind of second order algorithm--recursive approximate Newton algorithm was given by Karayiannis. The algorithm was simplified when it was formulated. Especially, the simplification to matrix Hessian was very reluctant, which led to the loss of valuable information and affected performance of the algorithm to certain extent. For multi layer feed forward neural networks, the second order back propagation recursive algorithm based generalized cost criteria was proposed. It is proved that it is equivalent to Newton recursive algorithm and has a second order convergent rate. The performance and application prospect are analyzed. Lots of simulation experiments indicate that the calculation of the new algorithm is almost equivalent to the recursive least square multiple algorithm. The algorithm and selection of networks parameters are significant and the performance is more excellent than BP algorithm and the second order learning algorithm that was given by Karayiannis.展开更多
We propose a dual feed-forward neural network(DFNN)model,consisting of a cavity parameter feature expander(CPFE)and a dynamic process predictor(DPP),for predicting the complex nonlinear dynamics of mode-locked fiber l...We propose a dual feed-forward neural network(DFNN)model,consisting of a cavity parameter feature expander(CPFE)and a dynamic process predictor(DPP),for predicting the complex nonlinear dynamics of mode-locked fiber lasers.The output of the CPFE,following layer normalization,is combined with the pulse complex electric field amplitude and then fed into the DPP to predict the dynamics.The pulse evolution process from the detuned steady state to the steady state under different cavity configurations is rapidly calculated.The predicted results of the proposed DFNN are consistent with the numerical split-step Fourier method(SSFM).The simulation speed has been greatly improved with low computational complexity,which is approximately 152 times faster than the SSFM and 4 times faster than the long short-term memory recurrent neural network(LSTM)model.The findings provide a new low computational complexity and efficient machine learning approach to model the complex nonlinear dynamics of mode-locked lasers.展开更多
A feed-forward Common-Mode （CM） charge control circuit for a high-speed Charge-Domain （CO） pipelined Analog-to-Digital Converter （ADC） is presented herein. This study aims at solving the problem whereby the prec...A feed-forward Common-Mode （CM） charge control circuit for a high-speed Charge-Domain （CO） pipelined Analog-to-Digital Converter （ADC） is presented herein. This study aims at solving the problem whereby the precision of CD pipelined ADCs is restricted by the variation in input CM charge, which can compensate for CM charge errors caused by a variation in CM charge input in real time. Based on the feed-forward CM charge control circuit, a 12-bit 250-MS/s CD pipelined ADC is designed and realized using a 1P6M 0.18-μm CMOS process. The ADC achieved a Spurious Free Dynamic Range （SFDR） of 78.1 dB and a Signal-to-Noise-and-Distortion Ratio （SNDR） of 64.6 dB for a 20.1-MHz input; a SFDR of 74.9 dB and SNDR of 62.0 dB were achieved for a 239.9-MHz input at full sampling rate. The variation in signal-to-noise ratio was less than 3 dB over a 0-1.2 V input CM voltage range. The power consumption of the prototype ADC is only 85 mW at 1.8 V supply, and it occupies an active die area of 2.24 mm^2.展开更多
Feed-forward loops(FFLs)are thought to be one of the most common and important classes of transcriptional network motifs involved in various diseases.Enhancers are cis-regulatory elements that positively regulate prot...Feed-forward loops(FFLs)are thought to be one of the most common and important classes of transcriptional network motifs involved in various diseases.Enhancers are cis-regulatory elements that positively regulate protein-coding genes or microRNAs(miRNAs)by recruiting DNA-binding transcription factors(TFs).However,a comprehensive resource to identify,store,and analyze the FFLs of typical enhancer and super-enhancer FFLs is not currently available.Here,we present EnhFFL,an online database to provide a data resource for users to browse and search typical enhancer and super-enhancer FFLs.The current database covers 46280/7000 TFenhancer-miRNA FFLs,9997/236 enhancer-miRNA-gene FFLs,3561164/3193182 TF-enhancer-gene FFLs,and 1259/235 TF-enhancer feed-back loops(FBLs)across 91 tissues/cell lines of human and mouse,respectively.Users can browse loops by selecting species,types of tissue/cell line,and types of FFLs.EnhFFL supports searching elements including name/ID,genomic location,and the conservation of miRNA target genes.We also developed tools for users to screen customized FFLs using the threshold of q value as well as the confidence score of miRNA target genes.Disease and functional enrichment analysis showed that master miRNAs that are widely engaged in FFLs including TF-enhancer-miRNAs and enhancer-miRNA-genes are significantly involved in tumorigenesis.Database URL:http://lcbb.swjtu.edu.cn/EnhFFL/.展开更多
This paper proposes a new Deep Feed-forward Neural Network(DFNN)approach for damage detection in functionally graded carbon nanotube-reinforced composite(FG-CNTRC)plates.In the proposed approach,the DFNN model is deve...This paper proposes a new Deep Feed-forward Neural Network(DFNN)approach for damage detection in functionally graded carbon nanotube-reinforced composite(FG-CNTRC)plates.In the proposed approach,the DFNN model is developed based on a data set containing 20000 samples of damage scenarios,obtained via finite element(FE)simulation,of the FG-CNTRC plates.The elemental modal kinetic energy(MKE)values,calculated from natural frequencies and translational nodal displacements of the structures,are utilized as input of the DFNN model while the damage locations and corresponding severities are considered as output.The state-of-the art Exponential Linear Units(ELU)activation function and the Adamax algorithm are employed to train the DFNN model.Additionally,in order to enhance the performance of the DFNN model,the mini-batch and early-stopping techniques are applied to the training process.A trial-and-error procedure is implemented to determine suitable parameters of the network such as the number of hidden layers and the number of neurons in each layer.The accuracy and capability of the proposed DFNN model are illustrated through two distinct configurations of the CNT-fibers constituting the FG-CNTRC plates including uniform distribution(UD)and functionally graded-V distribution(FG-VD).Furthermore,the performance and stability of the DFNN model with the consideration of noise effects on the input data are also investigated.Obtained results indicate that the proposed DFNN model is able to give sufficiently accurate damage detection outcomes for the FG-CNTRC plates for both cases of noise-free and noise-influenced data.展开更多
The harmonic disturbance in the background grid is a problem that must be considered in the design of a gridconnected inverter.However,the full feed-forward method cannot completely suppress the harmonic disturbance i...The harmonic disturbance in the background grid is a problem that must be considered in the design of a gridconnected inverter.However,the full feed-forward method cannot completely suppress the harmonic disturbance in theory and is sensitive to noise.To tackle these problems,a fractional full feed-forward method of grid voltage is proposed in this paper.First,the mathematical model of the full feed-forward method is deduced,and the differences with the theoretical solution,which can suppress all harmonics,are analyzed.Then,the parameter equation,the harmonic suppression performance,stability analysis and the implementation process of this method are given.Compared with the full feed-forward method,the proposed method not only further improves the harmonic suppression performance,but also reduces the order of the mathematical model of the differential term in the feed-forward loop.In addition,the proposed method can be used to flexibly design feed-forward coefficients by selecting the order of suppressed harmonics.Finally,the proposed method is validated by a hardware-in-the-loop experiment on a MT real-time control platform NI PXIE-1071.展开更多
基金supported by National Natural Science Foundation of China(Grant No.50437010)National Hi-tech Research and Development Program of China(863Program,Grant No.2006AA05Z205)Project of Six Talented Peak of Jiangsu Province,China(Grant No.07-D-013)
文摘There are two kinds of unbalance vibrations—force vibration and displacement vibration due to the existence of unbalance excitation in active magnetic bearings(AMB)system.And two unbalance compensation methods—closed-loop feedback and open loop feed-forward are presented to reduce the force vibration.The transfer function order of the control system directly influencing the system stability will be increased when the closed-loop method is adopted,which makes the real-time compensation not easily achieved.While the open loop method would not increase the primary transfer function order,it provides conditions for real-time compensation.But the real-time compensation signals are not easy to be obtained in the open loop method.To implement real-time force compensation,a new method is proposed to reduce the force vibration caused by the rotor unbalance on the basis of AMB active control.The method realizes real-time and on-line force auto-compensation based on H∞controller and one novel feed-forward compensation controller,which makes the rotor rotate around its inertia axis.The time-variable feed-forward compensatory signal is provided by a modified adaptive variable step-size least mean square(VSLMS)algorithm.And the relevant least mean square(LMS)algorithm parameters are used to solve the H∞controller weighting functions.The simulation of the new method to compensate some frequency-variable and sinusoidal signals is completed by MATLAB programming,and real-time compensation is implemented in the actual AMB experimental system.The simulation and experiment results show that the compensation scheme can improve the robust stability and the anti-interference ability of the whole AMB system by using H∞controller to achieve close-loop control,and then real-time force unbalance compensation is implemented.The proposed research provides a new control strategy containing real-time algorithm and H∞controller for the force compensation of AMB system.And the stability of the control system is finally improved.
文摘Since the dead zone phenomenon occurs in electro-hydraulic servo system, the output of the system corresponding to a sinusoidal input contains higher harmonic besides the fundamental input, which causes harmonic distortion of the output signal. The method for harmonic cancellation based on adaptive filter is proposed. The task is accomplished by generating reference signals with frequency that should be eliminated from the output. The reference inputs are weighted by the adaptive filter in such a way that it closely matches the harmonic. The output of the adaptive filter is a harmonic replica and is injected to the fundamental signal such that the output harmonic is cancelled leaving the desired signal alone, and the total harmonic distortion （THD） is greatly reduced. The weights of filter are adjusted on-line according to the control error by using least-mean-square （LMS） algorithm. Simulation results performed with a hydraulic system demonstrate the efficiency and validity of the proposed adaptive feed-forward compensator （AFC） control scheme
文摘A dynamic velocity feed-forward compensation （RBF-NN） dynamic model identification was presented for control （DVFCC） approach with RBF neural network the adaptive trajectory tracking of industrial robots. The proposed control approach combined the advantages of traditional feedback closed-loop position control and computed torque control based on inverse dynamic model. The feed-forward compensator used a nominal robot dynamics as accurate dynamic model and on-line identification with RBF-NN as uncertain part to improve dynamic modeling accu- racy. The proposed compensation was applied as velocity feed-forward by an inverse velocity controller that can con- vert torque signal into velocity in the standard industrial controller. Then, the need for a torque control interface was avoided in the real-time dynamic control of industrial robot. The simulations and experiments were carried out on a gas cutting manipulator. The results show that the proposed control approach can reduce steady-state error, suppress overshoot and enhance tracking accuracy and efficiency in joint space and Cartesian space, especially under high- speed condition.
基金National Natural Science Foundation of China(No.61261029)
文摘In order to improve the steady state performance,dynamic response and power factor of traditional power factor correction(PFC)digital control method and reduce the harmonic distortion of input current,a double closed loop active power factorcorrection(APFC)control method with feed-forward is proposed.Firstly,the small signal model of Boost PFC control systemis built and the system transfer function is deduced,and then the parameters of the main device with Boost topology is estimated.By means of the feed-forward,the system can quickly respond to the change in input voltage.Furthermore,the use ofvoltage loop and current loop can achieve input current and output voltage regulation Simulink modeling shows that this methodcan effectively control the output voltage in case of input voltage largely fluctuating,improve the system dynamic response abilityand input power factor,and reduce the input current harmonic distortion
文摘This paper describes the application of principal component analysis (PCA) and artificial neural network (ANN) to predict the air pollutant index (API) within the seven selected Malaysian air monitoring stations in the southern region of Peninsular Malaysia based on seven years database (2005-2011). Feed-forward ANN was used as a prediction method. The feed-forward ANN analysis demonstrated that the rotated principal component scores (RPCs) were the best input parameters to predict API. From the 4 RPCs, only 10 (CO, O3, PM10, NO2, CH4, NmHC, THC, wind direction, humidity and ambient temp) out of 12 prediction variables were the most significant parameters to predict API. The results proved that the ANN method can be applied successfully as tools for decision making and problem solving for better atmospheric management.
基金Supported by the National Key R&D Program of China(2016YFC1401900)the National Science Foundation of China(61173029,61672144)
文摘In data stream management systems (DSMSs), how to maintain the quality of queries is a difficult problem because both the processing cost and data arrival rates are highly unpredictable. When the system is overloaded, quality degrades significantly and thus load shedding becomes necessary. Unlike processing overloading in the general way which is only by a feedback control (FB) loop to obtain a good and stable performance over data streams, a feedback plus feed-forward control (FFC) strategy is introduced in DSMSs, which have a good quality of service (QoS) in the aspects of miss ratio and processing delay. In this paper, a quality adaptation framework is proposed, in which the control-theory-based techniques are leveraged to adjust the application behavior with the considerations of the current system status. Compared to previous solutions, the FFC strategy achieves a good quality with a waste of fewer resources.
基金Project supported by the Fundamental Research Funds for the Central Universities,China(Grant Nos.2662015QC041 and 2662014BQ069)the Huazhong Agricultural University Scientific&Technological Self-innovation Foundation,China(Grant No.2015RC021)the National Natural Science Foundation of China(Grant Nos.11675060,91730301,11547244,and 11474117)
文摘Feed-forward gene transcriptional regulatory networks, as a set of common signal motifs, are widely distributed in the biological systems. In this paper, the noise characteristics and propagation mechanism of various feed-forward gene transcriptional regulatory loops are investigated, including （i） coherent feed-forward loops with AND-gate, （ii） coherent feed-forward loops with OR-gate logic, and （iii） incoherent feed-forward loops with AND-gate logic. By introducing logarithmic gain coefficient and using linear noise approximation, the theoretical formulas of noise decomposition are derived and the theoretical results are verified by Gillespie simulation. From the theoretical and numerical results of noise decomposition algorithm, three general characteristics about noise transmission in these different kinds of feed-forward loops are observed, i） The two-step noise propagation of upstream factor is negative in the incoherent feed-forward loops with AND-gate logic, that is, upstream factor can indirectly suppress the noise of downstream factors, ii） The one-step propagation noise of upstream factor is non-monotonic in the coherent feed-forward loops with OR-gate logic, iii） When the branch of the feed-forward loop is negatively controlled, the total noise of the downstream factor monotonically increases for each of all feed-forward loops. These findings are robust to variations of model parameters. These observations reveal the universal rules of noise propagation in the feed-forward loops, and may contribute to our understanding of design principle of gene circuits.
基金the Key Project of the National Nature Science Foundation of China(No.61134009)Program for Changjiang Scholars and Innovation Research Team in University from the Ministry of Education,China(No.IRT1220)+1 种基金Specialized Research Fund for Shanghai Leading Talents,Project of the Shanghai Committee of Science and Technology,China(No.13JC1407500)the Fundamental Research Funds for the Central Universities,China(No.2232012A3-04)
文摘The coagulation bath system of carbon fiber precursor is a complicated and multivariable coupling system. Based on the model of industrial production,the full dynamic decoupling control of the coagulation bath system of carbon fiber precursor is achieved in combination with multivariable feed-forward-like decoupling and proportional-integral-differential( PID) control. Compared with the conventional PID decoupling control,the experiment results show that the proposed method has a better control effect. The use of the controller can achieve complete decoupling of three parameters from coagulation bath system. The method should have great applications.
文摘In the oil industry, the productivity of oil wells depends on the performance of the sub-surface equipment system. These systems often have problems stemming from sand, corrosion, internal pressure variation, or other factors. In order to ensure high equipment performance and avoid high-cost losses, it is essential to identify the source of possible failures in the early stage. However, this requires additional maintenance fees and human power. Moreover, the losses caused by these problems may lead to interruptions in the whole production process. In order to minimize maintenance costs, in this paper, we introduce a model for predicting equipment failure based on processing the historical data collected from multiple sensors. The state of the system is predicted by a Feed-Forward Neural Network (FFNN) with an SGD and Backpropagation algorithm is applied in the training process. Our model’s primary goal is to identify potential malfunctions at an early stage to ensure the production process’ continued high performance. We also evaluated the effectiveness of our model against other solutions currently available in the industry. The results of our study show that the FFNN can attain an accuracy score of 97% on the given dataset, which exceeds the performance of the models provided.
文摘A kind of second order algorithm--recursive approximate Newton algorithm was given by Karayiannis. The algorithm was simplified when it was formulated. Especially, the simplification to matrix Hessian was very reluctant, which led to the loss of valuable information and affected performance of the algorithm to certain extent. For multi layer feed forward neural networks, the second order back propagation recursive algorithm based generalized cost criteria was proposed. It is proved that it is equivalent to Newton recursive algorithm and has a second order convergent rate. The performance and application prospect are analyzed. Lots of simulation experiments indicate that the calculation of the new algorithm is almost equivalent to the recursive least square multiple algorithm. The algorithm and selection of networks parameters are significant and the performance is more excellent than BP algorithm and the second order learning algorithm that was given by Karayiannis.
基金supported by the National Natural Science Foundation of China(No.62203473)the Project of State Key Laboratory of Precision Manufacturing for Extreme Service Performance,Central South University(No.ZZYJKT2023-15)the Hunan Provincial Natural Science Foundation(No.2023JJ40778).
文摘We propose a dual feed-forward neural network(DFNN)model,consisting of a cavity parameter feature expander(CPFE)and a dynamic process predictor(DPP),for predicting the complex nonlinear dynamics of mode-locked fiber lasers.The output of the CPFE,following layer normalization,is combined with the pulse complex electric field amplitude and then fed into the DPP to predict the dynamics.The pulse evolution process from the detuned steady state to the steady state under different cavity configurations is rapidly calculated.The predicted results of the proposed DFNN are consistent with the numerical split-step Fourier method(SSFM).The simulation speed has been greatly improved with low computational complexity,which is approximately 152 times faster than the SSFM and 4 times faster than the long short-term memory recurrent neural network(LSTM)model.The findings provide a new low computational complexity and efficient machine learning approach to model the complex nonlinear dynamics of mode-locked lasers.
基金supported by National Natural Science Foundation of China under grant No.61704161Key Project of Natural Science of Anhui Provincial Department of Education under grant No.KJ2017A396
文摘A feed-forward Common-Mode （CM） charge control circuit for a high-speed Charge-Domain （CO） pipelined Analog-to-Digital Converter （ADC） is presented herein. This study aims at solving the problem whereby the precision of CD pipelined ADCs is restricted by the variation in input CM charge, which can compensate for CM charge errors caused by a variation in CM charge input in real time. Based on the feed-forward CM charge control circuit, a 12-bit 250-MS/s CD pipelined ADC is designed and realized using a 1P6M 0.18-μm CMOS process. The ADC achieved a Spurious Free Dynamic Range （SFDR） of 78.1 dB and a Signal-to-Noise-and-Distortion Ratio （SNDR） of 64.6 dB for a 20.1-MHz input; a SFDR of 74.9 dB and SNDR of 62.0 dB were achieved for a 239.9-MHz input at full sampling rate. The variation in signal-to-noise ratio was less than 3 dB over a 0-1.2 V input CM voltage range. The power consumption of the prototype ADC is only 85 mW at 1.8 V supply, and it occupies an active die area of 2.24 mm^2.
基金The research was supported by the Basic Research Cultivation Support Programof Fundamental Research Funds for the Central Universities,the National Natural Science Foundation of China(Grant No.32071343)the Sichuan Science and Technology Program(Grant No.2021YJ0048).
文摘Feed-forward loops(FFLs)are thought to be one of the most common and important classes of transcriptional network motifs involved in various diseases.Enhancers are cis-regulatory elements that positively regulate protein-coding genes or microRNAs(miRNAs)by recruiting DNA-binding transcription factors(TFs).However,a comprehensive resource to identify,store,and analyze the FFLs of typical enhancer and super-enhancer FFLs is not currently available.Here,we present EnhFFL,an online database to provide a data resource for users to browse and search typical enhancer and super-enhancer FFLs.The current database covers 46280/7000 TFenhancer-miRNA FFLs,9997/236 enhancer-miRNA-gene FFLs,3561164/3193182 TF-enhancer-gene FFLs,and 1259/235 TF-enhancer feed-back loops(FBLs)across 91 tissues/cell lines of human and mouse,respectively.Users can browse loops by selecting species,types of tissue/cell line,and types of FFLs.EnhFFL supports searching elements including name/ID,genomic location,and the conservation of miRNA target genes.We also developed tools for users to screen customized FFLs using the threshold of q value as well as the confidence score of miRNA target genes.Disease and functional enrichment analysis showed that master miRNAs that are widely engaged in FFLs including TF-enhancer-miRNAs and enhancer-miRNA-genes are significantly involved in tumorigenesis.Database URL:http://lcbb.swjtu.edu.cn/EnhFFL/.
基金This research was funded by Vietnam National Foundation for Science and Technology Development(NAFOSTED)under No.107.02-2019.330.
文摘This paper proposes a new Deep Feed-forward Neural Network(DFNN)approach for damage detection in functionally graded carbon nanotube-reinforced composite(FG-CNTRC)plates.In the proposed approach,the DFNN model is developed based on a data set containing 20000 samples of damage scenarios,obtained via finite element(FE)simulation,of the FG-CNTRC plates.The elemental modal kinetic energy(MKE)values,calculated from natural frequencies and translational nodal displacements of the structures,are utilized as input of the DFNN model while the damage locations and corresponding severities are considered as output.The state-of-the art Exponential Linear Units(ELU)activation function and the Adamax algorithm are employed to train the DFNN model.Additionally,in order to enhance the performance of the DFNN model,the mini-batch and early-stopping techniques are applied to the training process.A trial-and-error procedure is implemented to determine suitable parameters of the network such as the number of hidden layers and the number of neurons in each layer.The accuracy and capability of the proposed DFNN model are illustrated through two distinct configurations of the CNT-fibers constituting the FG-CNTRC plates including uniform distribution(UD)and functionally graded-V distribution(FG-VD).Furthermore,the performance and stability of the DFNN model with the consideration of noise effects on the input data are also investigated.Obtained results indicate that the proposed DFNN model is able to give sufficiently accurate damage detection outcomes for the FG-CNTRC plates for both cases of noise-free and noise-influenced data.
基金supported in part by the National Natural Science Foundation of China(61703375).
文摘The harmonic disturbance in the background grid is a problem that must be considered in the design of a gridconnected inverter.However,the full feed-forward method cannot completely suppress the harmonic disturbance in theory and is sensitive to noise.To tackle these problems,a fractional full feed-forward method of grid voltage is proposed in this paper.First,the mathematical model of the full feed-forward method is deduced,and the differences with the theoretical solution,which can suppress all harmonics,are analyzed.Then,the parameter equation,the harmonic suppression performance,stability analysis and the implementation process of this method are given.Compared with the full feed-forward method,the proposed method not only further improves the harmonic suppression performance,but also reduces the order of the mathematical model of the differential term in the feed-forward loop.In addition,the proposed method can be used to flexibly design feed-forward coefficients by selecting the order of suppressed harmonics.Finally,the proposed method is validated by a hardware-in-the-loop experiment on a MT real-time control platform NI PXIE-1071.
