Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Cristina Roxana Popa has a total of 20 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Deep Learning-Based Reconstruction and Representation Learning of Open-Hole Well Logs Using Machine Learning

    by Dan-Romulus Jacota, Cristina-Roxana Popa, Daniela Doina Neagu

    Well-log measurements provide essential information for lithological identification, petrophysical characterization, and reservoir interpretation but incomplete log intervals may occur because of technical, operational, or economic constraints. This study investigates machine-learning-based reconstruction of missing open-hole well-log curves using a depth-indexed dataset from a single well in the Bărbuncești oil and gas field, Romania. Seven predictors (DEPTH, AC, GR, RFOC, RILD, RILM, and SP) were used to estimate ten target properties (CALI, CGR, NPHI, NPOR, PEF, POTA, RHOB, SGR, THOR, and URAN) using Random Forest, Extra Trees, and Artificial Neural Networks. Approximately 2350 valid observations were used for model development and evaluation. Robustness was first assessed using 20 repeated 70/15/15 training/validation/testing partitions and an additional 80/10/10 sensitivity analysis. Under repeated 70/15/15 random partitioning, mean R2 values ranged from 0.809 for URAN to 0.935 for RHOB, with relatively small standard deviations, while the 80/10/10 configuration produced only limited changes in performance. However, a nested 10-fold depth-blocked cross-validation procedure, in which model-family selection was performed exclusively on training/validation data and each outer-test block remained completely untouched during model development, yielded negative mean outer-test R2 values for all ten target properties when DEPTH was included as a predictor. This marked discrepancy demonstrates that random hold-out validation can substantially overestimate predictive performance in densely sampled, depth-indexed, single-well data and that strong random-split performance should not be interpreted as evidence of generalization to unseen contiguous depth intervals. The reconstructed missing intervals are therefore presented as model-based estimates showing qualitative geological and petrophysical consistency rather than as quantitatively validated reconstructions, because measured target values are unavailable within those intervals. These findings highlight the importance of depth-structured validation for assessing machine-learning-based well-log reconstruction and the need for independent validation using additional wells and geological settings.

  • Open Access

    Machine-Learning Algorithm and Decline-Curve Analysis Comparison in Forecasting Gas Production

    by Dan-Romulus Jacota, Cristina Roxana Popa, Maria Tănase, Cristina Veres

    This study utilizes machine-learning algorithms to reinterpret existing datasets originally plotted using Decline-Curve Analysis (DCA), aiming to enhance predictive accuracy without requiring new field-data acquisition. Historical production records were compiled: monthly oil/gas rates, bottom-hole pressures, and cumulative productions, which were fitted to Arps equations via least-squares optimization, and key decline parameters, such as initial rate, nominal decline rate, and hyperbolic exponent, served as input data. Four machine-learning models were trained and validated: Artificial Neural Networks (ANN), Support Vector Machines (SVM), and Linear Regression (LR), using 80/20 train–test splits and 5-fold cross-validation. Models were evaluated using Mean Squared Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (R2). The ANN emerged as the best-performing method, achieving near-unity predictive accuracy (R2 ≈ 1) on the independent test set, with low error values (MSE = 0.0012 Ncm2/month2, RMSE = 0.035 Ncm/month, MAE = 0.028 Ncm/month) for oil production rates. Similar levels of accuracy were obtained for gas rates and pressures. These results reflect the strong and highly regular relationships present in the dataset analyzed rather than an exact zero-error fit. The multi-layer architecture of the ANN effectively captured the nonlinear interactions between Arps parameters and transient flow regimes, outperforming the empirical and physics-constrained approaches. Linear regression yielded strong results (R2 = 0.98, RMSE = 0.15 Ncm/month) but faltered in high-decline scenarios, failing to model exponential tails accurately. SVM exhibited the highest deviations (RMSE = 0.42 Ncm/month, R2 = 0.89), attributable to kernel sensitivity in sparse, noisy decline data. RF provided intermediate performance (R2 = 0.97). This ANN-driven approach redefines decline analysis by automating parameter tuning and uncertainty quantification, reducing forecasting errors by 85% versus classical Arps methods.


  • AN APLICATION FOR THE SELECTION AND SIZING OF CONTROL VALVE FOR CONTROL LOOP

    by Cristina Popa

    The paper presents the research of the author in developing an application for sizing and selection of control valves for typical process control applications. The proper sizing, selection, installation, and maintenance of control valve are all significant to obtain precise control performance. The paper has three parts. The first part describes the steps necessary to size and selection control valve based on the specific standards. The second part is destined to present the application developed by the author using Python programming language. In the last part, the author has presented and analyzed the numerical results of selection control valve using data from a real process.


  • HETEROGENOUS DISTRIBUTION OF INITIAL WATER SATURATION USING ARTIFICIAL NEURAL NETWORKS

    by Dan-Romulus Jacotă, Cristina Popa

    Artificial neuronal networks (ANNs) are rapidly becoming a very useful tool in the petroleum industry for predicting different evolution types for different parameters and use the human nervous system principles in creating the required prediction algorithm. Main objective of paper is to use a feedforward neural network to estimate the distribution of initial water saturation, as a small part of reservoir characterization in the presence of heterogeneities. It is very known that ANNs are complicated structures, take a long time in programing, are computer-time consuming and often require specialized aid in using them. Therefore, it will be an asset to know if reservoir heterogeneities can be pointed out with ANNs, or other prediction methods are indicated for these cases.

  • Academic Education Involvement in Refinery Advanced Process Control

    by Cristian Pătrăşcioiu, Marian Popescu, Nicolae Paraschiv, Cristina R. Popa, Nicoleta Nicolae

  • Open Access

    Estimation of the Kinetic Parameters of a Catalytic Cracking Model Case Study

    by Cristina Popa, Nicoleta Nicolae, Cristian Patrascioiu

    Cristina Popa, Nicoleta Nicolae, Cristian Patrascioiu - Estimation of the Kinetic Parameters of a Catalytic Cracking Model Case Study


  • Estimation of the Kinetic Parameters of a Catalytic Cracking Model Case Study

    by Cristina Roxana Popa

    This paper presents the research regarding determination of kinetic model parameters from a catalytic cracking process. Starting from the Weekman kinetic model, the authors proposed a simplified version of this model and, based on experimental data form a catalytic cracking plant, they have numerical determined the coefficients of the new kinetic model. For this purpose, there were defined two objective functions; the first function is based on errors generated by estimation of the riser outlet temperature and the second function associated to the errors generated by the estimation of the gasoline yield. The minimization of the two objective functions has been solve by using Optimization Toolbox from MATLAB programming language. The results showed that the objective function that depends on gasoline yield allows more accurate estimation of the kinetic parameters from this model. Keywords: kinetic model, optimization, catalytic cracking

  • Five-Lump Kinetic Model for the Catalytic Cracking Process\

    by Raluca Dragomir, Paul Rosca, Cristina Popa

    The main objectives of the present paper are to adaptation the five-kinetic model of the catalytic cracking process and simulation the riser to predicts the FCC products yields when one of the major input variable of the process is change. The simulation and adaptation are based on the industrial data from Romanian refinery. The adaptation is realize using a computational method from Optimization Toolbox from Matlab programming language. The new model can be used for optimization and control of FCC riser.

  • Improving FCC plant performance with model reference adaptive control based on neural network

    by Popa Cristina, Popa Alexandru

    The current paper presents the research results obtained by the authors regarding the automation of the reactor - regenerator group associated with the catalytic cracking process, using a model reference adaptive control system based on neural network. The first part of the paper describes the conventional control structure of the reactor-regenerator group. The second part of paper presents neural network model reference adaptive control structure proposed for the reactor-regenerator group. The proposed NN-model reference adaptive controller can significantly improve the reactor-regenerator group behavior and force the system to follow the reference model and minimize the error between the model and plant output. The results of the simulation have shown comparable or even superior performance against to the specific use of conventional techniques.