Canadian Journal of Chemical Engineering最新文献

This paper proposes a fault-tolerant predictive tracking control strategy for the industrial process based on a high-order fully actuated (HOFA) system method. First, a novel system representation method is employed to model the industrial process as a HOFA system. Subsequently, a fault-compensated HOFA predictive fault-tolerant control scheme is introduced, which includes two components: HOFA feedback stabilization and HOFA model predictive tracking control. Within this framework, an incremental predictive model is developed to replace the reduced-order prediction model by employing a Diophantine equation. The cost function, which incorporates tracking performance, is subsequently minimized using multi-step output prediction. Additionally, sufficient conditions for the stability and tracking performance of the closed-loop HOFA system are derived. The advantage of this approach lies in its ability to reduce system dimensionality while effectively eliminating the impact of faults on system stability through the introduction of an observer-based compensation concept. This ensures stable operation under fault conditions, or even operation unaffected by faults. Finally, the effectiveness and reliability of the proposed method are validated through a case study involving a nonlinear reactor.

Oil extraction methods are categorized into three main stages: primary, secondary, and tertiary enhanced oil recovery (EOR). In the tertiary stage, techniques such as chemical injection, thermal injection, and dissolved gas injection are employed, with nanoparticles providing innovative solutions. Following primary and secondary recovery processes, more than 50% of the total oil volume remains trapped in reservoirs, highlighting the significance of EOR. Nanoparticles, ranging from 1 to 100 nanometres, enhance EOR through mechanisms such as permeability control, interfacial tension reduction, and mass transfer improvement. Among the nanoparticles studied, silica nanoparticles have shown extensive potential due to their stability and ability to alter reservoir wettability. These nanoparticles, along with others such as magnesium oxide, aluminium oxide, zinc oxide, and iron oxide, can increase the recovery factor by up to 20% by altering wettability, decreasing interfacial tension, and improving mobility control. The application of nanotechnology in the oil industry spans from exploration to refining, enhancing processes with nanomaterials such as solid compounds, complex fluids, and nanoparticle mixtures. Challenges include the high cost of chemicals and environmental concerns. The use of nanoparticles, particularly silica nanoparticles, in EOR demonstrates significant potential for improving oil extraction methods; however, it faces challenges in maximizing oil recovery while minimizing negative environmental impacts. Future research should focus on the application of nanotechnology in EOR to develop methods that are both effective and environmentally sustainable. Balancing efficiency and environmental responsibility are essential for advancing toward a cleaner and more efficient oil industry.

The formation of hydrates provides a safe and efficient solution for the storage, transportation, and distribution of natural gas. The rapid generation of natural gas hydrates is one of the current research orientations. This paper focuses on studying the effects of different water saturations and salt concentrations on the generation kinetics and morphology of methane hydrates in complex systems, and fully considers the changes in the randomness characteristics of hydrate nucleation caused by these two factors. The experimental results show that with the increase of water saturation, the induction time of hydrate formation first decreases and then increases, while the randomness of the induction period gradually decreases and the distribution becomes more concentrated. Among them, a water saturation of 70% is relatively favourable for the formation of hydrates, with a shorter and more concentrated induction period. In addition, the magnitude of the salt concentration can affect the nucleation of hydrates. As the salt concentration increases, its effect on hydrate nucleation changes from promotion to inhibition. Therefore, there exists an optimal salt concentration range for promoting hydrate nucleation, but this optimal promotion range is different in systems with different water saturations. At a water saturation of 70%, the promotion range of the salt concentration on hydrate nucleation is larger. Therefore, water saturation and salt concentration have a coupling effect on the formation of hydrates. This study explains the reasons for the inconsistent effects of salt concentration on hydrate formation at present, and provides unique insights into the mechanism of hydrate formation.

Energy storage, exploiting the latent heat of phase change materials, offers an efficient method to store and release heat. Initial developments of phase change materials and their containment targeted performance over environmental impact. This article presents a bio-sourced, fully compostable, and biodegradable composite phase change material made from hardwood charcoal and beeswax. This material combines effective heat transfer performance with a significant focus on sustainable end-of-life waste management, especially when compared to traditional methods that rely on synthetic or heavily modified bio-sourced phase change materials. We measured a beeswax uptake within charcoal's porous structure of up to 61.2%. This resulted in a composite with a latent heat enthalpy of 77.83 J g⁻¹, with cyclability tested up to 20 cycles without any reduction in performance. The process was scaled from millimetre-scale particles up to centimetre-scale particles with similar adsorption capacities. The Laplace–Young equation confirmed that adsorption of beeswax is mostly driven by a capillary pressure, of at least 2.2 × 10⁴ N m⁻², caused by the naturally occurring porous structure of charcoal. These beeswax-loaded charcoal particles could find applications for heat recovery in HVAC systems or in packed bed heat storage systems.

Enhancing the efficiency of alkaline water electrolysis is critical for large-scale green hydrogen production, yet accurately predicting hydrogen-in-oxygen concentrations remains a significant challenge due to the complex nonlinear interactions between electrochemical and fluid dynamic parameters. This study employs machine learning to improve the predictive accuracy of hydrogen-in-oxygen levels under varying trans-diaphragm fluid flow conditions, addressing a gap in existing modelling approaches that rely primarily on theoretical or empirical methods. Five artificial neural network models were developed using experimental data from a 0.6 m single-stack electrolyzer operating with an electrolyte. The models were trained and tested on 132 experimental data points, with 75% allocated for training and 25% for testing. The number of neurons in the hidden layer of the network models developed with a single hidden layer and the TanSig activation function was optimized by analyzing the performance of different network models. The models achieved exceptional predictive accuracy, with mean squared errors below 1.47E-02, correlation coefficients exceeding 0.989, and margin of deviation within ±0.82% across all test cases. These findings confirm the capability of machine learning-based predictive modelling to enhance electrolysis optimization, reduce experimental costs, and support the scalable deployment of green hydrogen production. The novel integration of machine learning in trans-diaphragm fluid flow analysis advances predictive modelling beyond conventional techniques, offering a robust approach for industrial-scale electrolysis system enhancement. This study primarily aims to develop accurate predictive models for hydrogen-in-oxygen concentrations under varying trans-diaphragm flow conditions, addressing a critical gap in monitoring and controlling alkaline water electrolysis systems.

Aiming at the question of information loss between layers when mining spatiotemporal features of process data and whether pseudo-labels are generated for unlabelled data, this paper proposes the dynamically scaling spatio-temporal semi-supervised adaptive networks based soft sensor for industrial process (DSST-SSAN). In order to extract the local temporal correlation features and decrease the information loss between layers, the dynamic scaled spatio-temporal feature module is constructed, the local prediction models between the current input and the hidden layer features are built in each hidden layer of the long short-term memory (LSTM) network respectively, the prediction deviations of multiple local models are calculated and the dynamic scaled factors are constructed to update the corresponding hidden layer features. The spatial features are extracted in parallel using graph attention network (GAT), and the spatio-temporal features are obtained by fusion to establish a soft sensor model. To address the lack of modelling labelling data, a semi-supervised thresholding mechanism is proposed to filter the pseudo-labelled data for adaptive data accumulation. The threshold is constructed using the likelihood root mean square of the root mean square error (RMSE) and mean absolute error (MAE) of the labelled data, which can determine whether the unlabelled data need to generate pseudo-labels and perform modelling data accumulation and thus update the model. The effectiveness of the proposed method is confirmed by simulation experiments on two industrial cases, debutane tower and sulphur recovery.

This study explores a novel approach to the geometric optimization of coiled flow inverters (CFIs) aimed at enhancing biodiesel production efficiency. By simulating nine distinct CFI geometries using advanced computational fluid dynamics (CFD) and genetic algorithms (GA), this research introduces innovative methods for optimizing fluid flow characteristics. The CFD model yielded essential hydrodynamic data and friction factors, while oil conversion percentages for biodiesel were derived from existing literature. The integration of CFD results with experimental data significantly informed the GA optimization process, marking a key advancement in the field. Two new correlations were developed to predict friction factors and oil conversion percentages based on the coil length-to-diameter ratio, Reynolds number, and the number of 90° bends. This study uniquely identifies optimal geometries through a GA-based multi-objective approach, effectively balancing oil conversion and friction factor. Additionally, it delves into the trade-offs between improving oil conversion and the resultant increase in pressure drop, highlighting the intricate complexities of fluid flow in CFIs and their implications for biodiesel production efficiency.

As industrial-scale dense gas–solid fluidized bed separators have been gradually employed in dry coal beneficiation, it is urgent to develop feasible and efficient methods to evaluate and quantify the bed density stability, which directly influences the coal separation efficiency. The instantaneous signals recorded with the pressure sensor and optical fibre probe (OFP) is utilized to analyze the nonlinear characteristics and evaluate the complexity and instability of the fluidization process for dense gas solid fluidization. It is verified that the hidden chaotic characteristics of pressure drop signals can be retrieved with multi-dimensional reconstruction of attractor. Based on the reconstructed attractor, the Shannon entropy and Kolmogorov entropy are investigated and estimated under different bed heights and air flows. The results indicate that the differential pressure signal and the optical fibre signal are typical chaotic signals, effectively representing the complexity of fluid dynamics in the local measurement space. Nonlinear bubble behaviour is the primary cause of the increased rate of information loss in chaotic signals, which severely affects the stability of fluidization quality in the bed. The spatial distribution of the chaos index corroborates the internal circulation pattern within the fluidized bed, which can feasibly characterize and quantify the nonlinear characteristics and instability of gas solid fluidization.

Poly (methyl methacrylate) (PMMA) is a thermoplastic with outstanding tensile strength, UV resistance, and a high level of transparency that has been used widely for optical applications such as glazing in the automobile industry. Mechanical recycling, the most widespread method, degrades the physical properties and prevents reusing PMMA in transparent applications. Thermal depolymerization to recover methyl methacrylate (MMA) monomer is becoming an alternative route for PMMA recycling. In this paper, the thermal depolymerization process of impact-modified PMMA in a micro fluidized bed reactor was investigated. The pyrolysis was conducted over aluminium oxide ($��{\mathrm{Al}}_2�O_3�$) and fluid-cracking catalyst (FCC) as catalytic beds; sand and SiC as inert beds at temperatures below $��{400}^{°}�C�$. A mechanical recycling process was also simulated using sequential injection moulding to investigate its impact on the properties of PMMA. After 5 cycles of injection moulding, the impact strength and optical properties of PMMA were severely diminished due to an increase in free volume and partial thermal degradation. Regarding PMMA pyrolysis, $��{\mathrm{Al}}_2�O_3�$ demonstrated limited cracking ability with a maximum MMA yield of 46%, as opposed to FCC, which over-cracked both PMMA and MMA into coke and unwanted products. In contrast, non-catalytic beds exhibited higher activity for MMA recovery, with SiC yielding the highest amount of 92% at $��{380}^{°}�C�$.

Solvent-assisted recovery methods have gained attention as an alternative to conventional thermal methods which have significant environmental drawbacks and high greenhouse gas emissions. Solvent-assisted recovery methods can lead to various phase equilibria, such as vapour–liquid equilibrium (VLE) and liquid–liquid equilibrium (LLE), under specific thermodynamic conditions. A thorough understanding of this phenomenon is essential for effectively designing and optimizing these processes. Despite the availability of both VLE and LLE for mixtures of solvent-bitumen systems, LLE data is scarcer, particularly at conditions close to the solvent's saturation pressure. This study reports on LLE measurements of n-butane–bitumen mixtures at temperatures ranging from 295 to 372 K. The pressures were set just above the solvent's saturation pressure at each temperature to maintain LLE conditions. We examined both the phase behaviour and thermophysical properties of the mixture. The thermophysical properties of the light phase, including density and viscosity, were recorded. The bubble point pressure, n-butane content in the light phase, heavy phase onset, and detailed characterization of both the light and heavy cuts were also conducted. The experimental thermophysical data were then modelled using the Peng–Robinson equation of state (PR-EoS) and modified Pederson model. In conclusion, these findings enhance our understanding of the LLE behaviour of n-butane–bitumen systems under conditions near the solvent's saturation pressure and find plications in solvent-aided in situ bitumen recovery processes.