This article presents an overview of the principle, engineering implementation and applications of spectroscopic measurements technology, such as NIR (near infrared), FTIR (Fourier transform infrared), Raman and NMR (nuclear magnetic resonance). The discussions are focused on oil sands industry in which the process majorly consists of the mining, slurry hydro-transport, extraction, froth treatment and upgrading plants.

Steam quality is a critical process variable in the once through steam generator (OTSG) operation. However, the lack of online measurement and closed loop control of steam quality limits the efficient operation of OTSGs. To resolve this problem, this paper presents a model predictive control (MPC) solution based on soft sensor measurement for OTSG steam quality. Bad status handling strategy is outlined to ensure reliable estimation and control results. Instrumentation reliability issue is also considered in the MPC design. Successful application results have demonstrated effectiveness of the developed control strategy.

A definition for the reliability of inferential sensor predictions is provided. A data-driven Bayesian framework for real-time performance assessment of inferential sensors is proposed. The main focus is on characterizing the effect of operating space on the reliability of inferential sensor predictions. A holistic, quantitative measure of the reliability of the inferential sensor predictions is introduced. A methodology is provided to define objective prior probabilities over plausible classes of reliability based on the total misclassification cost. The real-time performance assessment of multi-model inferential sensors is also discussed. The application of the method does not depend on the identification techniques employed for model development. Furthermore, on-line implementation of the method is computationally efficient. The effectiveness of the method is demonstrated through simulation and industrial case studies.

A data-driven Bayesian framework for real-time performance assessment of inferential sensors is proposed. The application of the proposed Bayesian solution does not depend on the identification techniques employed for inferential model development. The effectiveness of the proposed method is demonstrated through a simulation case study.

An alarm is a notification of an abnormal event that requires operator action. In the process industries, owing to inefficiencies in configuring alarms, an abnormal event is often not annunciated using a single alarm. Chattering alarms are the most common form of nuisance alarms. Chattering is a condition where an alarm is annunciated excessively within a short span of time. Quantification of alarm chatter is important to evaluate the performance of various alarm generating algorithms. In this work, the concept of run length is applied to alarm data and an index is proposed to quantify the degree of alarm chatter based on run length distributions. Prominent features of the proposed index are demonstrated using real industrial alarm data.

Oil sands development is both a costly and technically complex business with potential concerns in land use, water consumption and greenhouse gas emissions. Therefore, it is of practical interest to further investigate novel techniques to improve profitability while diligently maintaining environmental compliance. Our approach for finding solutions to achieve this objective is to develop innovative strategies for advanced monitoring, optimisation and control of plant operations. Development of reliable process models is a key requirement for investigating the behaviour of complex systems. Such descriptive models can help to improve analysis, simulation, optimisation, design, control and operation of process systems at both micro and macro levels. This paper presents a summary of some of the successful applications focussed on development and implementation of inferential process models, also known as soft sensors for oil sands processes.

Separation cells used in primary extraction in the oil sands industry are integral components in the overall process of bitumen extraction. Good regulation of the interface level between the bitumen froth and the middlings in these cells can result in a significant improvement in bitumen recovery and heavily influence process economics. This paper details a case study application of identification and the design of a model based predictive controller for the separation cell process. Model predictive control (MPC) using linear models is designed, implemented and tested in real time on the industrial separation cell. The test results show that the MPC scheme provides significant benefits over current operation which uses a PID controller. The benefits include significant reduction in the variance of the interface level and underflow pump movement, resulting in higher bitumen recovery, smoother operations downstream and pump energy savings.

The major impediment in controlling the interface between Bitumen-froth and Middlings in separation cells is the lack of safe and reliable sensors for interface level detection. This work describes a novel sensor for this problem using computer vision techniques on video frames captured from a sight glass camera. A simple edge detection method combined with state-space model based particle filtering is used to estimate the interface level and its quality. Industrial results show that the algorithm is robust to lighting changes and process abnormalities. Highly improved control performance results when the sensor estimates are used for feedback control.

I Dynamic Modeling through Subspace Identification.- System Identification: Conventional Approach.- Open-loop Subspace Identification.- Closed-loop Subspace Identification.- Identification of Dynamic Matrix and Noise Model Using Closed-loop Data.- II Predictive Control.- Model Predictive Control: Conventional Approach.- Data-driven Subspace Approach to Predictive Control.- III Control Performance Monitoring.- Control Loop Performance Assessment: Conventional Approach.- State-of-the-art MPC Performance Monitoring.- Subspace Approach to MIMO Feedback Control Performance Assessment.- Prediction Error Approach to Feedback Control Performance Assessment.- Performance Assessment with LQG-benchmark from Closed-loop Data.

Bicoherence or Bispectrum analysis is emerging as a new powerful technique in signal processing, especially in areas where traditional linear spectral analysis provides insufficient information. It is most effective in analyzing systems with non-linear coupling between frequencies. Faults in rotating machineries leave their signature on the vibration signal sensors and generally manifest themselves as a non-linear transformation in the vibration signal. Bicoherence analysis detects and quantifies the presence of non-linearity in the signal and thus indicates the severity of the fault in the machine. This paper demonstrates the use of bicoherence analysis on both simulated and rig-generated vibration data from a rub-effected rotor-stator system, and shows the application of bicoherence analysis on industrial data from final tailing pumps to detect impeller wear in an oil-sands plant. Keywords: Rotating Machinery, Fault Detection, Vibration, Bispectrum Analysis