Memory versus non-linearity in reservoirs

Verstraeten, David; Dambre, Joni; Dutoit, Xavier; Schrauwen, Benjamin

doi:10.1109/ijcnn.2010.5596492

Cited by 78 publications

(55 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The setting of parameters is not very critical, and was chosen to give best generalization performance. Note that a decrease in the input scaling counteracts an increase in the spectral radius to avoid a possible loss of memory capacity and of performance [37].…”

Section: Methodsmentioning

confidence: 99%

“…Although it does not guarantee the ESP, in practice it has been empirically observed that this criterium works well and often produces analog sigmoid ESNs with ESP for any input. It is important to note that spectral radius closer to unity as well as larger input scaling makes the reservoir more non-linear, which has a deterioration impact on the memory capacity as side-effect [37]. This scaling of matrices is important because it influences greatly the reservoir dynamics [38] and, in this way, must be chosen according to the task at hand empirically, analyzing the behavior of the reservoir states over time, or by grid searching.…”

Section: Esn Modelmentioning

confidence: 99%

See 1 more Smart Citation

Echo State Networks for data-driven downhole pressure estimation in gas-lift oil wells

Antonelo¹,

Camponogara²,

Foss

2017

Neural Networks

View full text Add to dashboard Cite

Process measurements are of vital importance for monitoring and control of industrial plants. When we consider offshore oil production platforms, wells that require gas-lift technology to yield oil production from low pressure oil reservoirs can become unstable under some conditions. This undesirable phenomenon is usually called slugging flow, and can be identified by an oscillatory behavior of the downhole pressure measurement. Given the importance of this measurement and the unreliability of the related sensor, this work aims at designing data-driven soft-sensors for downhole pressure estimation in two contexts: one for speeding up first-principle model simulation of a vertical riser model; and another for estimating the downhole pressure using real-world data from an oil well from Petrobras based only on topside platform measurements. Both tasks are tackled by employing Echo State Networks (ESN) as an efficient technique for training Recurrent Neural Networks. We show that a single ESN is capable of robustly modeling both the slugging flow behavior and a steady state based only on a square wave input signal representing the production choke opening in the vertical riser. Besides, we compare the performance of a standard network to the performance of a multiple timescale hierarchical architecture in the second task and show that the latter architecture performs better in modeling both large irregular transients and more commonly occurring small oscillations.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Esn Modelmentioning

confidence: 99%

Echo State Networks for data-driven downhole pressure estimation in gas-lift oil wells

Antonelo¹,

Camponogara²,

Foss

2017

Neural Networks

View full text Add to dashboard Cite

show abstract

“…Considering a normalized input signal u[n], the effect of input scaling υ r i on the reservoir dynamics is such that, the larger the scaling, the closer to saturation the reservoir states will be, since the reservoir state is shifted torwards the nonlinear area of the tanh activation function. Spectral radius closer to unity as well as larger input scaling makes the reservoir more non-linear, which has a deterioration impact on the memory capacity as side-effect [38].…”

Section: Methodsmentioning

confidence: 99%

On Learning Navigation Behaviors for Small Mobile Robots With Reservoir Computing Architectures

Antonelo

Schrauwen

2015

IEEE Trans. Neural Netw. Learning Syst.

Self Cite

View full text Add to dashboard Cite

Abstract-This work proposes a general Reservoir Computing (RC) learning framework which can be used to learn navigation behaviors for mobile robots in simple and complex unknown, partially observable environments. RC provides an efficient way to train recurrent neural networks by letting the recurrent part of the network (called reservoir) fixed while only a linear readout output layer is trained. The proposed RC framework builds upon the notion of navigation attractor or behavior which can be embedded in the high-dimensional space of the reservoir after learning. The learning of multiple behaviors is possible because the dynamic robot behavior, consisting of a sensorymotor sequence, can be linearly discriminated in the highdimensional nonlinear space of the dynamic reservoir. Three learning approaches for navigation behaviors are shown in this paper. The first approach learns multiple behaviors based on examples of navigation behaviors generated by a supervisor, while the second approach learns goal-directed navigation behaviors based only on rewards. The third approach learns complex goaldirected behaviors, in a supervised way, using an hierarchical architecture whose internal predictions of contextual switches guide the sequence of basic navigation behaviors towards the goal.

show abstract

“…It is important to note that spectral radius closer to unity as well as larger input scaling makes the reservoir more non-linear, which has a deterioration impact on the memory capacity as side-effect [13]. The scaling of these nontrainable weights is a parameter which should be chosen according to the task at hand empirically, analyzing the behavior of the reservoir state over time, or by grid searching over parameter ranges.…”

Section: Esn Modelmentioning

confidence: 99%

An Echo State Network-Based Soft Sensor of Downhole Pressure for a Gas-Lift Oil Well

Antonelo

Camponogara

2015

Engineering Applications of Neural Networks

View full text Add to dashboard Cite

Abstract.1 Soft sensor technology has been increasingly used in industry. Its importance is magnified when the process variable to be estimated is key to control and monitoring processes and the respective sensor either has a high probability of failure or is unreliable due to harsh environment conditions. This is the case for permanent downhole gauge (PDG) sensors in the oil and gas industry, which measure pressure and temperature in deepwater oil wells. In this paper, historical data obtained from an actual offshore oil well is used to build a black box model that estimates the PDG downhole pressure from platform variables, using Echo State Networks (ESNs), which are a class of recurrent networks with powerful modeling capabilities. These networks, differently from other neural networks models used by most soft sensors in literature, can model the nonlinear dynamical properties present in the noisy real-world data by using a two-layer structure with efficient training: a recurrent nonlinear layer with fixed randomly generated weights and a linear adaptive readout output layer. Experimental results show that ESNs are a promising technique to model soft sensors in an industrial setting.

show abstract

Memory versus non-linearity in reservoirs

Cited by 78 publications

References 14 publications

Echo State Networks for data-driven downhole pressure estimation in gas-lift oil wells

Echo State Networks for data-driven downhole pressure estimation in gas-lift oil wells

On Learning Navigation Behaviors for Small Mobile Robots With Reservoir Computing Architectures

An Echo State Network-Based Soft Sensor of Downhole Pressure for a Gas-Lift Oil Well

Contact Info

Product

Resources

About