Clustering is an important tool in data-driven fault diagnosis to make use of unlabeled data. Collecting representative data for fault diagnosis is a difficult task since faults are rare events. In addition, using data collected from the field, e.g., logged operational data and data from different workshops about replaced components, can result in labelling uncertainties. A common approach for fault diagnosis of dynamic systems is to use residual-based features that filter out system dynamics while being sensitive to faults. The use of conventional clustering algorithms is complicated by that the distribution of residual data from one fault class varies for different realizations and system operating conditions. In this work, a clustering algorithm is proposed for residual data that clusters data by estimating fault signatures in residual space. The proposed clustering algorithm can be used on time-series data by clustering batches of data from the same fault scenario instead of clustering data sample-by-sample. The usefulness of the proposed clustering algorithm is illustrated using residual data from different fault scenarios collected from an internal combustion engine test bench. Copyright (C) 2022 The Authors.

Turbocharger testing is a time consuming process, and as rapid-prototyping technology advances, so must other areas in the development chain. As an example, in one study a compressor map took over 34 hours to measure. In this paper, an effort to combat the main bottleneck of turbocharger testing, namely the thermal inertia, is made. When changing operating point during the measurement process, several minutes can be required before the turbocharger components reach temperature steady state. In an earlier paper, a method based on non-linear trajectory optimization was developed that significantly reduced the testing time required to produce compressor performance maps. The time was reduced by a factor of over 60, compared to waiting for the system to reach steady state with constant inputs. However, the method required a model of the turbocharger. This paper extends the method with system identification and model predictive control (MPC). This is an important step in order to use the optimal control method when only geometric information of the turbocharger is known, such as new prototypes. To demonstrate the effectiveness of the combination of system identification, non-linear trajectory optimization and MPC, the control strategy is applied to a virtual gas-stand implemented as a Simulink model, based on data from a Mitsubishi TD04 turbocharger. The data that was used to create the model was originally collected at Saab Automobile in Trollhättan, 2011. The results show that system identification captures the turbocharger behavior. Trajectory optimization finds a set of time optimal input trajectories. MPC successfully tracks the generated references. Real time implementation of the Matlab/Simulink based algorithm is planned for experimental testing.

Turbocharger maps are used in design, evaluation and optimization of engine system operation to represent the turbo operation in different scenarios. To construct such a map, the turbo is tested in a gas flow test bench, called gas stand. Turbo testing is a time and resource consuming experimental process. The turbo is tested in a selected number of test points for different turbo rotational speeds, where the temperatures in the turbo have to be stationary when the measurements that constitute the map are acquired. In this paper, optimal control is used to find the most time efficient pattern of test conditions, and the optimal control strategy to traverse between them. A heat transfer model, describing the heat transfer between the compressor, bearing house, and turbine, is presented and validated against measured data. A direct collocation method is used to find time optimal control trajectories between the specified test points in the map. The method objective is to find the least time consuming control strategy which brings the turbo from one test point to the next, while ensuring thermal equilibrium at the final time. The results suggest that this method reduces turbocharger testing time with a factor higher than 60. The improvements can be further increased, with over 70 times, if a traveling salesman problem is solved to find the optimal route through the turbo map. The described method would be able to map a 43 points turbo map in 22 minutes, including a 5 minute warm-up phase. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.