Job Safety Analysis (JSA) is critical for proactively identifying workplace hazards, assessing their potential consequences, and implementing effective control measures. However, traditional JSA methods can be inefficient and prone to errors, particularly in complex industrial environments. This paper introduces AutoJSA, a knowledge-enhanced framework that leverages large language models (LLMs) to automate and optimize the JSA process. We collected 73 high-quality JSA reports from a chemical engineering company and divided the JSA workflow into three key tasks: hazard identification, consequence identification, and control measure generation. Two approaches - fine-tuning and retrieval-augmented generation (RAG) - were employed on a base LLM (GLM-4-9B-Chat) to adapt it for these domain-specific tasks. Experimental results demonstrate that both fine-tuning and RAG significantly improve task performance relative to the unmodified model, with fine-tuning generally providing larger gains. W... [more]

Hybrid modelling is widely employed in chemical engineering to generate highly accurate predictions. Such an approach merges first-principle modelling with machine learning techniques to identify and model the epistemic uncertainty from experimental data. Despite its advantages, this still requires cross-domain competencies that are difficult to find in the chemical industry and high human involvement. The possibility of automating the identification and training model would be significantly beneficial for the widespread adoption of hybrid modelling methodology within the chemical industry. This work presents a novel algorithm for the automatic identification of hybrid models (HMs) starting from the first-principle representation of the system, described by differential equation sets. The methodology formulates the problem as mixed-integer programming, identifying the equation running under uncertainty, identifying the machine learning model hyperparameters, and training the latter. Th... [more]

This study aims at developing a holistic approach to establish robust mathematical models of integrated and interactive multistep processes, while systematically identifying the corresponding design space and acceptable operating region (AOR). The overall objective is to reduce the experimentation costs, enhance accuracy of integrated metathetical models, and deliver built-in quality assurance based on a new Quality by Digital Design (QbDD) paradigm. This methodology starts with the construction of a set of model candidates for different unit operations, based on the prior knowledge and inherent assumptions. Several model candidates of the integrated multistep process are considered. A model discrimination based on model prediction performance reveals the best integrated model for the multistep process. In the next step, the estimability analysis and model-based design of experiment (MBDoE) are implemented to deliver information-rich data and systematically refine the integrated model.... [more]

Design space identification (DSId) and flexibility analysis are critical in process systems engineering, enabling efficient design of operating conditions. For bioprocess, these tasks are often hindered by the absence of reliable mechanistic models and limited experimental data. This paper presents an algorithm to address these challenges in bioprocesses. The methodology begins by constructing a Gaussian process (GP) model to predict key performance indicators (KPIs) from process inputs. Leveraging the probabilistic nature of GP predictions, we perform probabilistic design space identification (PDSId), characterizing each input point by its probability of feasibility which is the likelihood that constraints imposed on KPIs are satisfied. To visualize and analyse the feasibility space, contours at varying probability levels are identified using alpha shapes, which define deterministic boundaries corresponding to different confidence levels. This enables the quantification of volumetric... [more]

Endpoint indicators provide a concise representation of environmental impacts by aggregating multiple midpoint indicators into a single value. Traditional endpoint weighting systems, however, are often limited by biases introduced through panel reviews and a lack of robustness in scientific process models. Additionally, they typically fail to account for the preferences of key stakeholders, including industry, government, and the public. This work addresses these limitations by developing an endpoint indicator that incorporates stakeholder preferences and minimizes dissatisfaction. A multi-stakeholder optimization framework was formulated to achieve this goal, employing distance, downside risk, and conditional value at risk as objective functions. Stakeholder preferences were derived from emissions data for industry, federal spending on environmental issues for government, and public surveys for societal input. Results highlight regional variations in midpoint indicator weightings acro... [more]

This paper presents state-of-the-art object detection and object identification algorithms for digitizing and interpreting chemical engineering diagrams, including, Block Flow Diagrams (BFDs), Process Flow Diagrams (PFDs), and Piping and Instrumentation Diagrams (P&IDs), using computer vision techniques. These diagrams are essential for visualizing plant processes and equipment but are often stored as image-based PDFs, making manual digitization/interpretation labor-intensive and error-prone. The proposed algorithm automates tasks such as detecting unit operations and identifying them using a set of rule-based and predefined approaches including edge and contour-based rules, spatial arrangement rules, and geometric rules. This method avoids data requirements and computational requirements of deep learning approaches, offering a scalable and efficient solution for preliminary extraction of complex process information.

This study introduces a novel methodology that combines Gaussian process regression (GPR) with alpha shape design space reconstruction to visualize multi-dimensional design spaces. The proposed GPR surrogate approach incorporates a kernel optimization step, employing a greedy tree search strategy to identify the optimal combinatorial kernel from a selection of base kernels. This approach efficiently evaluates design spaces around specific points of interest, enabling alpha shape reconstruction. The methodology's adaptability is demonstrated through its application to both lower-dimensional (2D and 3D) cases and more complex, higher-dimensional systems (up to 7D), showcasing its scalability and versatility. Its effectiveness is further validated by its ability to generate accurate surrogate models from limited data. Overall, this study presents a robust framework that leverages GPR surrogate modeling and alpha shape reconstruction to facilitate design space evaluation in complex, multid... [more]

Addressing challenges in process design and optimisation, especially with complex models and data uncertainties, requires effective tools for model development, selection, and identification. Techniques such as Model-based Design of Experiments (MBDoE) help support this task by screening and discriminating between models and, eventually, calibrating them. Open-source and user-friendly Python packages have implemented some model identification techniques. However, the need for a tool that can couple with various model simulators and account for the steps of model identification as well as physical constraints of systems in design of experiments remains unmet. In that light, we present the python package MIDDOE (Model-(based) Identification, Discrimination, and Design of Experiments) to address this gap. It integrates rival models screening, parameter estimation, uncertainty analysis, and MBDoE techniques, while adapting to various process constraints. These functionalities are demonstra... [more]

Predicting processes’ future behavior based on past data is vital for automatic control and dynamic optimization in engineering. Recent advances in deep learning, particularly Artificial Neural Networks, have improved predictions in various engineering fields. Recurrent Neural Networks (RNNs) are well-suited for time series data, as they naturally evolve through dynamic systems with recurrent updates. Despite their high predictive power, RNNs may underperform if their training ignores the model's future application. In Model Predictive Control, for example, the model evolves over time using only current information, relying on its own predictions at later steps. A model trained for one-step-ahead predictions may fail when tasked with multi-step-ahead forecasting in autoregressive mode. This study explores deep recurrent neural network models for predicting critical operational time series of a large-scale Electric Submersible Pump system. We present an innovative training approach, fra... [more]

Nucleophilic aromatic substitutions (SNAr) are key chemical transformations in pharmaceutical and agrochemical synthesis, yet their complex mechanisms (concerted or two-step) complicate kinetic model identification. Accurate kinetic models for SNAr are essential for scale-up, optimization, and control of the reaction process, but conventional methods struggle with mechanism uncertainty driven by substrates, nucleophiles, and reaction conditions, with data collection being difficult due to its source-intensive nature. We address this using DoE-SINDy, a data-driven framework for generative modelling without complete theoretical understanding. A benchmark study on the SNAr reaction of 2,4-difluoronitrobenzene with morpholine in ethanol was conducted, incorporating parallel and consecutive side-product formation. Ground-truth kinetic models validated in prior studies were used to generate in-silico data under varying noise levels and sampling intervals. DoE-SINDy successfully identified th... [more]

N4-methylcytosine (4mC) is a critical epigenetic modification that plays a pivotal role in the regulation of a multitude of biological processes, including gene expression, DNA replication, and cellular differentiation. Traditional experimental methods for detecting DNA N4-methylcytosine sites are time-consuming, labor-intensive, and costly, making them unsuitable for large-scale or high-throughput research. Computational methods for identifying DNA N4-methylcytosine sites enable the rapid and cost-effective analysis of DNA 4mC sites across entire genomes. In this study, we focus on the identification of DNA 4mC sites in the mouse genome. Although there are already some computational methods that can predict DNA 4mC sites in the mouse genome, there is still significant room for improvement in accurately predicting them due to their inability to fully capture the multifaceted characteristics of DNA sequences. To address this issue, we propose a new deep learning predictor called Mus4mCP... [more]

Sustainable and green waste management has become increasingly crucial due to the rising volume of waste driven by urbanization and population growth. Deep learning models based on image recognition offer potential for advanced waste classification and recycling methods. However, traditional image recognition approaches usually rely on single-label images, neglecting the complexity of real-world waste occurrences. Moreover, there is a scarcity of recognition efforts directed at actual municipal waste data, with most studies confined to laboratory settings. Therefore, we introduce an efficient Query2Label (Q2L) framework, powered by the Vision Transformer (ViT-B/16) as its backbone and complemented by an innovative asymmetric loss function, designed to effectively handle the complexity of multi-label waste image classification. Our experiments on the newly developed municipal waste dataset “Garbage In, Garbage Out”, which includes 25,000 street-level images, each potentially containing... [more]

In additive manufacturing, such as Selective Laser Melting (SLM), identifying fabrication defects poses a significant challenge. Existing identification algorithms often struggle to meet the precision requirements for defect detection. To accurately identify small-scale defects in SLM, this paper proposes a deep learning model based on the original YOLOv5 network architecture for enhanced defect identification. Specifically, we integrate a small target identification layer into the network to improve the recognition of minute anomalies like keyholes. Additionally, a similarity attention module (SimAM) is introduced to enhance the model’s sensitivity to channel and spatial features, facilitating the identification of dense target regions. Furthermore, the SPD-Conv module is employed to reduce information loss within the network and enhance the model’s identification rate. During the testing phase, a set of sample photos is randomly selected to evaluate the efficacy of the proposed model... [more]

The eEF2K, a member of the α-kinase family, plays a crucial role in cellular differentiation and the stability of the nervous system. The development of eEF2K inhibitors has proven to be significantly important in the treatment of diseases such as cancer and Alzheimer’s. With the advancement of big data in pharmaceuticals and the evolution of molecular generation technologies, leveraging artificial intelligence to expedite research on eEF2K inhibitors shows great potential. Based on the recently published structure of eEF2K and known inhibitor molecular structures, a generative model was used to create 1094 candidate inhibitor molecules. Analysis indicates that the model-generated molecules can comprehend the principles of molecular docking. Moreover, some of these molecules can modify the original molecular frameworks. A molecular screening strategy was devised, leading to the identification of five promising eEF2K inhibitor lead compounds. These five compound molecules demonstrated e... [more]

Since the “13th Five-Year Plan”, the exploration of large-scale structural oil and gas reservoirs in the Bohai oilfield has become more complex, and the exploration of igneous oil and gas reservoirs has become the focus of current attention. At present, igneous rock reservoir fluid identification methods are mainly based on the evaluation method of logging single parameter construction, which is primarily a qualitative identification due to lithology, physical property, and engineering factors. Accurate acquisition of interference logging data, and multi-parameter coupling and recording coupling methods are few, lacking systematic and comprehensive evaluation and analysis of logging data. Since conventional logging data in the study area have difficulty accurately and quickly identifying reservoir fluid properties, a systematic analysis was conducted of three factors: lithology, physical properties, and engineering, as well as a variety of logging parameters (gas measurement, three-dim... [more]

As a distinct type of reservoir, tuffaceous tight reservoirs have attracted much attention. However, previous studies on tuffaceous tight reservoirs formed in the burial diagenetic stage are few, particularly regarding the genesis of micropores, which restricts the in-depth exploration of tuffaceous tight oil. According to thin section observation, scanning electron microscopy (SEM) identification, X-ray diffraction (XRD) experiments, elemental analyses, porosity and permeability tests, and pore structure analyses, the micropore characteristics of the Carboniferous tuffaceous tight reservoir formed by burial dissolution in the Santanghu Basin, NW China, are studied. In addition, the cause of the tuff micropore formation and its geological significance are also researched in this paper. The results are as follows: (1) The tuffaceous tight reservoir formed by burial dissolution mainly consists of quartz, feldspar, dolomite, and clay minerals. The reservoir space mainly consists of interg... [more]

Recently, the “sweet spot” of a fractured reservoir, controlled by a strike-slip fault, has been found and become the favorable target for economic exploitation of deep (>4500 m) tight gas reservoirs in the Sichuan Basin, Southwestern China. However, hidden faults of small vertical displacements (<20 m) are generally difficult to identify using low signal−noise rate seismic data for deep subsurfaces. In this study, we propose a seismic processing method to improve imaging of the hidden strike-slip fault in the central Sichuan Basin. On the basis of the multidirectional and multiscale decomposition and reconstruction processes, seismic information on the strike-slip fault can be automatically enhanced to improve images of it. Through seismic processing, the seismic resolution increased to a large extent enhancing the fault information and presenting a distinct fault plane rather than an ambiguous deflection of the seismic wave, as well as a clearer image of the sectional seismic attr... [more]

Overflow is one of the complicated working conditions that often occur in the drilling process. If it is not discovered and controlled in time, it will cause gas invasion, kick, and blowout, which will bring inestimable accidents and hazards. Therefore, overflow identification and early warning has become a hot spot and a difficult problem in drilling engineering. In the face of the limitations and lag of traditional overflow identification methods, the poor application effect, and the weak mechanisms of existing models and methods, a method of series fusion of feature data obtained from physical models as well as sliding window and random forest machine learning algorithm models is proposed. The overflow identification and early warning model of managed pressure drilling based on a series fusion data-driven model is established. The research results show that the series fusion data-driven model in this paper is superior to the overflow identification effect of other feature data and a... [more]

The massive integration of distributed renewable energy sources and nonlinear power electronic equipment has given rise to power quality issues such as waveform distortion, voltage instability, and increased harmonic components. Nowadays, the pollution of power quality is becoming increasingly severe, posing a potential threat to the security of the power grid and the stable operation of electrical equipment. Due to the presence of significant noise interference in the collected signals, existing methods still face issues such as low accuracy in disturbance identification and high computational complexity. To address these problems, this paper proposes a hybrid data-driven strategy that can significantly improve the accuracy and speed of identification. Firstly, the wavelet packet transform (WPT) method is employed to denoise the power disturbance signals. Subsequently, the local mean decomposition (LMD) algorithm is used to adaptively decompose the nonlinear and complex time series in... [more]

Stuck pipe refers to the accidental phenomenon whereby drilling tools are stuck in a well during the drilling process and cannot move freely due to various reasons. As a result, the stuck pipe can consume a lot of manpower and material resources. With the development of artificial intelligence, the intelligent prediction and identification of stuck pipe risk has gradually advanced. However, there are usually only a few stuck samples, so the intelligent model is not sufficient to excavate the stuck feature law, and then the model overfitting phenomenon occurs. Regarding the above issue, this paper proposed a limited incident dataset method based on data augmentation. Firstly, in terms of data processing, by applying percentage scaling and random dithering to the original data and combining it with GAN to generate new data, the training dataset was effectively extended, solving the problem of insufficient sample size. Then, in the selection and training of the intelligent model, an LSTM... [more]

Modeling the non-linear dynamics of a system from measurement data accurately is an open challenge. Over the past few years, various tools such as SINDy and DySMHO have emerged as approaches to distill dynamics from data. However, challenges persist in accurately capturing dynamics of a system especially when the physical knowledge about the system is unknown. A promising solution is to use a hybrid paradigm, that combines mechanistic and black-box models to leverage their respective strengths. In this study, we combine a hybrid modeling paradigm with sparse regression, to develop and identify models simultaneously. Two methods are explored, considering varying complexities, data quality, and availability and by comparing different case studies. In the first approach, we integrate SINDy-discovered models with neural ODE structures, to model unknown physics. In the second approach, we employ Multifidelity Surrogate Models (MFSMs) to construct composite models comprised of SINDy-discover... [more]

Machine learning presents opportunities to improve the scale-specific accuracy of mechanistic models in a data-driven manner. Here we demonstrate the use of a machine learning technique called Sparse Identification of Nonlinear Dynamics (SINDy) to improve a simple mechanistic model of algal growth. Time-series measurements of the microalga Chlorella Vulgaris were generated under controlled photobioreactor conditions at the University of Technology Sydney. A simple mechanistic growth model based on intensity of light and temperature was integrated over time and compared to the time-series data. While the mechanistic model broadly captured the overall growth trend, discrepancies remained between the model and data due to the model's simplicity and non-ideal behavior of real-world measurement. SINDy was applied to model the residual error by identifying an error derivative correction term. Addition of this SINDy-informed error dynamics term shows improvement to model accuracy while maint... [more]

The determination of the design space (DS) in a pharmaceutical process is a crucial aspect of the quality-by-design (QbD) initiative which promotes quality built into the desired product. This is achieved through a deep understanding of how the critical quality attributes (CQAs) and process parameters (CPPs) interact that have been demonstrated to provide quality assurance. For computational inexpensive models, the original process model can be directly deployed to identify the design space. One such crucial process is the Tablet Press (TP), which directly compresses the powder blend into individual units of the final product or adds dry or wet granulation to meet specific formulation needs. In this work, we identify the design space of input variables in a TP such that there is a (probabilistic) guarantee that the tablets meet the quality constraints under a set of operating conditions. A reduced-order model of TP is assigned for this purpose where the effects of lubricants and glidan... [more]

This paper presents an algorithm for developing sparse surrogate models that satisfy mass/energy conservation even when the training data are noisy and violate the conservation laws. In the first step, we employ the Bayesian Identification of Dynamic Sparse Algebraic Model (BIDSAM) algorithm proposed in our previous work to obtain a set of hierarchically ranked sparse models which approximate system behaviors with linear combinations of a set of well-defined basis functions. Although the model building algorithm was shown to be robust to noisy data, conservation laws may not be satisfied by the surrogate models. In this work we propose an algorithm that augments a data reconciliation step with the BIDSAM model for satisfaction of conservation laws. This method relies only on known boundary conditions and hence is generic for any chemical system. Two case studies are considered-one focused on mass conservation and another on energy conservation. Results show that models with minimum bia... [more]

To address challenges in the detection of wind turbine blade damage images, characterized by complex backgrounds and multiscale feature distribution, we propose a method based on an enhanced YOLOV8 model. Our approach focuses on three key aspects: First, we enhance the extraction of small target features by integrating the CBAM attention mechanism into the backbone network. Second, the feature fusion process is refined using the Weighted Bidirectional Feature Pyramid Network (BiFPN) to replace the path aggregation network (PANet). This modification prioritizes small target features within the deep features and facilitates the fusion of multiscale features. Lastly, we improve the loss function from CIoU to EIoU, enhancing sensitivity to small targets and the perturbation resistance of bounding boxes, thereby reducing the gap between computed predictions and real values. Experimental results demonstrate that compared with the YOLOV8 model, the CBAM-BiFPN-YOLOV8 model exhibits improvement... [more]