Vijay Manikandan Janakiraman

Nonlinear identification of a gasoline HCCI engine using neural networks coupled with principal component analysis

Homogeneous charge compression ignition (HCCI) is a futuristic combustion technology that operates with high efficiency and reduced emissions. HCCI combustion is characterized by complex nonlinear dynamics which necessitates the use of a predictive model in controller design. Developing a physics based model for HCCI involves significant development times and associated costs arising from developing simulation models and calibration. In this paper, a neural networks (NN) based methodology is reported where black box type models are developed to predict HCCI combustion behavior during transient operation. The NN based approach can be considered a low cost and quick alternative to the traditional physics based modeling. A multi-input single-output model was developed each for indicated net mean effective pressure, combustion phasing, maximum in-cylinder pressure rise rate and equivalent air-fuel ratio. The two popular architectures namely multi-layer perceptron (MLP) and radial basis network (RBN) models were compared with respect to design, prediction performance and overall applicability to the transient HCCI modeling problem. A principal component analysis (PCA) is done as a pre-processing step to reduce input dimension thereby reducing memory requirements of the models. Also, PCA reduces the cross-validation time required to identify optimal model hyper-parameters. On comparing the model predictions with the experimental data, it was shown that neural networks can be a powerful approach for non-linear identification of a complex combustion system like the HCCI engine.

An ELM based predictive control method for HCCI engines

We formulate and develop a control method for homogeneous charge compression ignition (HCCI) engines using model predictive control (MPC) and models learned from operational data. An HCCI engine is a highly efficient but complex combustion system that operates with a high fuel efficiency and reduced emissions compared to the present technology. HCCI control is a nonlinear, multi-input multi-output problem with state and actuator constraints which makes controller design a challenging task. In this paper, we propose an MPC approach where the constraints are elegantly included in the control problem along with optimality in control. We develop the engine models using experimental data so that the complexity and time involved in the modeling process can be reduced. An Extreme Learning Machine (ELM) is used to capture the engine dynamic behavior and is used by the MPC controller to evaluate control actions. We also used a simplified quadratic programming making use of the convexity of the MPC problem so that the algorithm can be implemented on the engine control unit that is limited in memory. The working and effectiveness of the proposed MPC methodology has been analyzed in simulation using a nonlinear HCCI engine model. The controller tracks several reference signals taking into account the constraints defined by HCCI states, actuators and operational limits.

Stochastic gradient based extreme learning machines for stable online learning of advanced combustion engines

We propose and develop SG-ELM, a stable online learning algorithm based on stochastic gradients and Extreme Learning Machines (ELM). We propose SG-ELM particularly for systems that are required to be stable during learning; i.e., the estimated model parameters remain bounded during learning. We use a Lyapunov approach to prove both asymptotic stability of estimation error and boundedness in the model parameters suitable for identification of nonlinear dynamic systems. Using the Lyapunov approach, we determine an upper bound for the learning rate of SG-ELM. The SG-ELM algorithm not only guarantees a stable learning but also reduces the computational demand compared to the recursive least squares based OS-ELM algorithm (Liang et al., 2006). In order to demonstrate the working of SG-ELM on a real-world problem, an advanced combustion engine identification is considered. The algorithm is applied to two case studies: An online regression learning for system identification of a Homogeneous Charge Compression Ignition (HCCI) Engine and an online classification learning (with class imbalance) for identifying the dynamic operating envelope. The case studies demonstrate that the accuracy of the proposed SG-ELM is comparable to that of the OS-ELM approach but adds stability and a reduction in computational effort.

Identification of the Dynamic Operating Envelope of HCCI Engines Using Class Imbalance Learning

Homogeneous charge compression ignition (HCCI) is a futuristic automotive engine technology that can significantly improve fuel economy and reduce emissions. HCCI engine operation is constrained by combustion instabilities, such as knock, ringing, misfires, high-variability combustion, and so on, and it becomes important to identify the operating envelope defined by these constraints for use in engine diagnostics and controller design. HCCI combustion is dominated by complex nonlinear dynamics, and a first-principle-based dynamic modeling of the operating envelope becomes intractable. In this paper, a machine learning approach is presented to identify the stable operating envelope of HCCI combustion, by learning directly from the experimental data. Stability is defined using thresholds on combustion features obtained from engine in-cylinder pressure measurements. This paper considers instabilities arising from engine misfire and high-variability combustion. A gasoline HCCI engine is used for generating stable and unstable data observations. Owing to an imbalance in class proportions in the data set, the models are developed both based on resampling the data set (by undersampling and oversampling) and based on a cost-sensitive learning method (by overweighting the minority class relative to the majority class observations). Support vector machines (SVMs) and recently developed extreme learning machines (ELM) are utilized for developing dynamic classifiers. The results compared against linear classification methods show that cost-sensitive nonlinear ELM and SVM classification algorithms are well suited for the problem. However, the SVM envelope model requires about 80% more parameters for an accuracy improvement of 3% compared with the ELM envelope model indicating that ELM models may be computationally suitable for the engine application. The proposed modeling approach shows that HCCI engine misfires and high-variability combustion can be predicted ahead of time, given the present values of available sensor measurements, making the models suitable for engine diagnostics and control applications.

A System Identification Framework for Modeling Complex Combustion Dynamics Using Support Vector Machines

Machine Learning is being widely applied to problems that are difficult to model using fundamental building blocks. However, the application of machine learning in powertrain modeling is not common because existing powertrain systems have been simple enough to model using simple physics. Also, black box models are yet to demonstrate sufficient robustness and stability features for widespread powertrain applications. However, with emergence of advanced technologies and complex systems in the automotive industry, obtaining a good physical model in a short time becomes a challenge and it becomes important to study alternatives. In this chapter, support vector machines (SVM) are used to obtain identification models for a gasoline homogeneous charge compression ignition (HCCI) engine. A machine learning framework is discussed that addresses several challenges for identification of the considered system that is nonlinear and whose region of stable operation is very narrow.

Discovery of Precursors to Adverse Events using Time Series Data.

We develop an algorithm for automatic discovery of precursors in time series data (ADOPT). In a time series setting, a precursor may be considered as any event that precedes and increases the likelihood of an adverse event. In a multivariate time series data, there are exponential number of events which makes a brute force search intractable. ADOPT works by breaking down the problem into two steps (1) inferring a model of the nominal time series (data without adverse event) by considering the nominal data to be generated by a hidden expert and (2) using the expert’s model as a benchmark to evaluate the adverse time series to identify suboptimal events as precursors. For step (1), we use a Markov Decision Process (MDP) framework where value functions and Bellman’s optimality are used to infer the expert’s actions. For step (2), we define a precursor score to evaluate a given instant of a time series by comparing its utility with that of the expert. Thus, the search for precursors is transformed to a search for sub-optimal action sequences in ADOPT. As an application case study, we use ADOPT to discover precursors to go-around events in commercial flights using real

Anomaly detection in aviation data using extreme learning machines

We develop fast anomaly detection algorithms using extreme learning machines (ELM) to discover operationally significant anomalies in large aviation data sets. Anomaly detection (aka one-class classification or outlier detection) is an active area of research to identify safety risks in aviation. Aviation data is characterized by high dimensionality, heterogeneity (continuous and categorical variables), multimodality and temporality. To address these challenges, NASA Ames has developed several anomaly detection algorithms including MKAD, the present state of the art [1]. MKADs computational complexity is quadratic with respect to the number of training examples which makes it time consuming (and sometimes infeasible) for mining very large data sets. In this paper, we utilize ELMs fast training and good generalization properties to develop scalable anomaly detection algorithms for very large data sets. We adapt unsupervised ELM algorithms such as the autoencoder and embedding models to perform anomaly detection. The unsupervised models capture the nominal data distribution and by choosing a desired strength of detection that defines the upper bound of outliers in the training data, the anomaly decision boundary is determined. The autoencoder model detects anomalies as the ones that have a large reconstruction error while the embedding model detects anomalies as the ones that lie outside a hypersphere in the embedded space. The proposed algorithms are applied to a real aviation safety benchmark problem and the results show that the ELM based algorithms are comparable to MKAD in detection while training is made faster by two orders of magnitude.

A lyapunov based stable online learning algorithm for nonlinear dynamical systems using extreme learning machines

Extreme Learning Machine (ELM) is a promising learning scheme for nonlinear classification and regression problems and has shown its effectiveness in the machine learning literature. ELM represents a class of generalized single hidden layer feed-forward networks (SLFNs) whose hidden layer parameters are assigned randomly resulting in an extremely fast learning speed along with superior generalization performance. It is well known that the online sequential learning algorithm (OS-ELM) based on recursive least squares [1] might result in ill-conditioning of the Hessian matrix and hence instability in the parameter estimation. To address this issue, the stability theory of Lyapunov is utilized to develop an online learning algorithm for temporal data from dynamic systems and time series. The developed algorithm results in parameter estimation that is asymptotically stable leading to boundedness in model states. Simulations results of the developed algorithm compared against online sequential ELM (OS-ELM) and the offline batch learning ELM (O-ELM) show that the Lyapunov ELM algorithm can perform online learning at reduced computation, comparable accuracy and with a guarantee on the boundedness of the estimated system.

Finding Precursors to Anomalous Drop in Airspeed During a Flight's Takeoff

Aerodynamic stall based loss of control in flight is a major cause of fatal flight accidents. In a typical takeoff, a flights airspeed continues to increase as it gains altitude. However, in some cases, the airspeed may drop immediately after takeoff and when left uncorrected, the flight gets close to a stall condition which is extremely risky. The takeoff is a high workload period for the flight crew involving frequent monitoring, control and communication with the ground control tower. Although there exists secondary safety systems and specialized recovery maneuvers, current technology is reactive; often based on simple threshold detection and does not provide the crew with sufficient lead time. Further, with increasing complexity of automation, the crew may not be aware of the true states of the automation to take corrective actions in time. At NASA, we aim to develop decision support tools by mining historic flight data to proactively identify and manage high risk situations encountered in flight. In this paper, we present our work on finding precursors to the anomalous drop-in-airspeed (ADA) event using the ADOPT (Automatic Discovery of Precursors in Time series) algorithm. ADOPT works by converting the precursor discovery problem into a search for sub-optimal decision making in the time series data, which is modeled using reinforcement learning. We give insights about the flight data, feature selection, ADOPT modeling and results on precursor discovery. Some improvements to ADOPT algorithm are implemented that reduces its computational complexity and enables forecasting of the adverse event. Using ADOPT analysis, we have identified some interesting precursor patterns that were validated to be operationally significant by subject matter experts. The performance of ADOPT is evaluated by using the precursor scores as features to predict the drop in airspeed events.

Explaining Aviation Safety Incidents Using Deep Temporal Multiple Instance Learning

Although aviation accidents are rare, safety incidents occur more frequently and require a careful analysis to detect and mitigate risks in a timely manner. Analyzing safety incidents using operational data and producing event-based explanations is invaluable to airline companies as well as to governing organizations such as the Federal Aviation Administration (FAA) in the United States. However, this task is challenging because of the complexity involved in mining multi-dimensional heterogeneous time series data, the lack of time-step-wise annotation of events in a flight, and the lack of scalable tools to perform analysis over a large number of events. In this work, we propose a precursor mining algorithm that identifies events in the multidimensional time series that are correlated with the safety incident. Precursors are valuable to systems health and safety monitoring and in explaining and forecasting safety incidents. Current methods suffer from poor scalability to high dimensional time series data and are inefficient in capturing temporal behavior. We propose an approach by combining multiple-instance learning (MIL) and deep recurrent neural networks (DRNN) to take advantage of MILs ability to learn using weakly supervised data and DRNNs ability to model temporal behavior. We describe the algorithm, the data, the intuition behind taking a MIL approach, and a comparative analysis of the proposed algorithm with baseline models. We also discuss the application to a real-world aviation safety problem using data from a commercial airline company and discuss the models abilities and shortcomings, with some final remarks about possible deployment directions.