Josip Budimir Sacher
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Josip Budimir Sacher has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Nenad Bolf
Department of Measurements and Process Control, Faculty of Chemical Engineering and Technology, University of Zagreb, 10000 Zagreb, Croatia
6 shared publications - Ivan Vrban
Krka, 8501 Novo Mesto, Slovenia
2 shared publications Matea Gavran
Sveučilište u Zagrebu, Fakultet kemijskog inženjerstva i tehnologije
2 shared publications- Željka Ujević Andrijić
Faculty of Chemical Engineering, University of Zagreb, 10000 Zagreb, Croatia
2 shared publications - Damir Šahnić
PLIVA Croatia Ltd. (member of TEVA Group), Prilaz baruna Filipovića 25, 10000 Zagreb, Croatia
1 shared publication Ivan Vrban
Pliva Hrvatska d.o.o.
1 shared publication
Latest Publications
- Open Access
Robustness-Aware Genetic Algorithm for Batch Crystallization with an LSTM Digital Twin
Batch crystallization processes are prone to batch-to-batch inconsistencies arising from operational uncertainties and equipment-induced noise. This study presents a Robustness-Aware Genetic Algorithm (RAGA) integrated with a Long Short-Term Memory (LSTM) digital twin for the design of robust crystallization procedures. The RAGA employs a hierarchical fitness function that strictly enforces a target median crystal size D50 as the primary constraint while maximizing process yield as a secondary objective. Robustness is incorporated directly into the optimization by requiring candidate trajectories to satisfy the D50 specification across five independent stochastic realizations with perturbed operating conditions. A candidate is promoted in the evolutionary search only if all five evaluations produce a predicted D50 within ±2 µm of the target. The framework was applied to seeded cooling crystallization of creatine monohydrate across three target crystal sizes of 115, 125, and 135 µm. Robustness of optimal crystallization procedures was independently verified through 100-run Monte Carlo simulations under ±10% parameter perturbations with success defined as D50 within ±5 µm of target. Experimental validation at laboratory scale confirmed that optimized procedures translate to practice, with two of three target sizes achieved within the ±5 µm specification and the third deviating due to the combined effect of LSTM prediction uncertainty and thermal lag. Despite having no embedded mechanistic knowledge, the optimizer successfully converged on physically coherent crystallization strategies. Its variations in seed loading, batch time, and cooling trajectory parameters remained entirely consistent with established principles of supersaturation management. The results demonstrate that embedding robustness directly within the evolutionary optimization loop enables consistent crystal size control using data-driven models.
Open AccessRobustness-Aware Genetic Algorithm for Batch Crystallization with an LSTM Digital Twin
Batch crystallization processes are prone to batch-to-batch inconsistencies arising from operational uncertainties and equipment-induced noise. This study presents a Robustness-Aware Genetic Algorithm (RAGA) integrated with a Long Short-Term Memory (LSTM) digital twin for the design of robust crystallization procedures. The RAGA employs a hierarchical fitness function that strictly enforces a target median crystal size D50 as the primary constraint while maximizing process yield as a secondary objective. Robustness is incorporated directly into the optimization by requiring candidate trajectories to satisfy the D50 specification across five independent stochastic realizations with perturbed operating conditions. A candidate is promoted in the evolutionary search only if all five evaluations produce a predicted D50 within ±2 µm of the target. The framework was applied to seeded cooling crystallization of creatine monohydrate across three target crystal sizes of 115, 125, and 135 µm. Robustness of optimal crystallization procedures was independently verified through 100-run Monte Carlo simulations under ±10% parameter perturbations with success defined as D50 within ±5 µm of target. Experimental validation at laboratory scale confirmed that optimized procedures translate to practice, with two of three target sizes achieved within the ±5 µm specification and the third deviating due to the combined effect of LSTM prediction uncertainty and thermal lag. Despite having no embedded mechanistic knowledge, the optimizer successfully converged on physically coherent crystallization strategies. Its variations in seed loading, batch time, and cooling trajectory parameters remained entirely consistent with established principles of supersaturation management. The results demonstrate that embedding robustness directly within the evolutionary optimization loop enables consistent crystal size control using data-driven models.
- Open Access
Direct Nucleation Control with External Heating Loop: Process System Engineering, Simulation and Experimental Investigation
The aim of this work was to explore and test the concept of a novel direct nucleation control method with external dissolution of fine crystals (E-DNC). It was postulated that the heating cycles of the internal DNC method could be transferred from the crystallizer jacket to an electrically heated recirculation tube, thereby using less energy and requiring a shorter time compared to internal DNC. The conceptual model was explored and developed by reviewing previous research on the topic and addressing known drawbacks. Engineering experience and a heuristic approach led to the conclusion that five commonly used controllers would suffice for a straightforward implementation of the method. A simplified simulation was developed to compare the time and energy requirements for the internal and external DNC methods, and it was concluded that external DNC uses 37% less energy and 19% less time compared to internal DNC. The laboratory system was then constructed by modifying the internal DNC apparatus with inexpensive and commonly used components. A linear cooling experiment was performed to establish the baseline for comparison with E-DNC experiments and to set the expected count range. The E-DNC experiments were then conducted with the aim of obtaining larger crystal sizes, and it was shown that the process could be designed in only three experiments. Adjusting the heating rate and count limit led to a significant increase in median crystal size (22.5%) compared to linear cooling.
Open AccessDirect Nucleation Control with External Heating Loop: Process System Engineering, Simulation and Experimental Investigation
The aim of this work was to explore and test the concept of a novel direct nucleation control method with external dissolution of fine crystals (E-DNC). It was postulated that the heating cycles of the internal DNC method could be transferred from the crystallizer jacket to an electrically heated recirculation tube, thereby using less energy and requiring a shorter time compared to internal DNC. The conceptual model was explored and developed by reviewing previous research on the topic and addressing known drawbacks. Engineering experience and a heuristic approach led to the conclusion that five commonly used controllers would suffice for a straightforward implementation of the method. A simplified simulation was developed to compare the time and energy requirements for the internal and external DNC methods, and it was concluded that external DNC uses 37% less energy and 19% less time compared to internal DNC. The laboratory system was then constructed by modifying the internal DNC apparatus with inexpensive and commonly used components. A linear cooling experiment was performed to establish the baseline for comparison with E-DNC experiments and to set the expected count range. The E-DNC experiments were then conducted with the aim of obtaining larger crystal sizes, and it was shown that the process could be designed in only three experiments. Adjusting the heating rate and count limit led to a significant increase in median crystal size (22.5%) compared to linear cooling.
- Open Access
Data-Driven Prediction of Crystal Size Metrics Using LSTM Networks and In Situ Microscopy in Seeded Cooling Crystallization
This work presents a data-driven modeling framework for predicting image-derived crystal size metrics in seeded cooling crystallization using Long Short-Term Memory (LSTM) neural networks. The model leverages in situ microscopy data to predict square-weighted D10, D50, D90, and particle counts based solely on seed loading and temperature profiles, without requiring real-time supersaturation measurements. To enhance predictive power, engineered process descriptors—including temperature derivatives and integrals—were incorporated as dynamic features. Experimental validation was performed using creatine monohydrate crystallization from aqueous solution, with LSTM models trained on a diverse dataset encompassing variable seed loadings and cooling profiles. The feature-engineered LSTM model consistently outperformed its non-engineered counterpart, particularly under nonlinear cooling conditions where crystallization dynamics were the most complex. This approach offers a practical alternative to mechanistic models and spectroscopic process analytical technology (PAT) tools by enabling accurate prediction of chord length distribution (CLD) metrics from routinely collected data. The framework is easily transferable to other crystallization systems and provides a low-complexity, high-accuracy tool for accelerating lab-scale crystallization development.
Open AccessData-Driven Prediction of Crystal Size Metrics Using LSTM Networks and In Situ Microscopy in Seeded Cooling Crystallization
This work presents a data-driven modeling framework for predicting image-derived crystal size metrics in seeded cooling crystallization using Long Short-Term Memory (LSTM) neural networks. The model leverages in situ microscopy data to predict square-weighted D10, D50, D90, and particle counts based solely on seed loading and temperature profiles, without requiring real-time supersaturation measurements. To enhance predictive power, engineered process descriptors—including temperature derivatives and integrals—were incorporated as dynamic features. Experimental validation was performed using creatine monohydrate crystallization from aqueous solution, with LSTM models trained on a diverse dataset encompassing variable seed loadings and cooling profiles. The feature-engineered LSTM model consistently outperformed its non-engineered counterpart, particularly under nonlinear cooling conditions where crystallization dynamics were the most complex. This approach offers a practical alternative to mechanistic models and spectroscopic process analytical technology (PAT) tools by enabling accurate prediction of chord length distribution (CLD) metrics from routinely collected data. The framework is easily transferable to other crystallization systems and provides a low-complexity, high-accuracy tool for accelerating lab-scale crystallization development.
- Open Access
Batch Cooling Crystallization of a Model System Using Direct Nucleation Control and High-Performance In Situ Microscopy
The aim of this study was to investigate the use of automated high performance in situ microscopy (HPM) for monitoring and direct nucleation control (DNC) during cooling crystallization. Compared to other techniques, HPM enables the detection of small crystals in the range of 1 to 10 μm. Therefore, a novel DNC-controlled variable was investigated to determine the potential improvement of the method. The laboratory system and process control software were developed in-house. A well-studied crystallization model system, the seeded batch cooling crystallization of α-glycine from water, was investigated under normal conditions and temperatures below 60 °C. It was postulated that length-weighted edge-to-edge counts in the range of 1 to 10 μm would be most sensitive to the onset of secondary nucleation and are therefore, used as a control variable. Linear cooling experiments were conducted to determine the initial setpoint for the DNC experiments. Three DNC experiments were then performed with different setpoints and an upper and lower counts limit. It was found that the DNC method can be destabilized with a low setpoint and narrow counts limits. In addition, the new controlled variable is highly sensitive to the formation of bubbles at the microscope window and requires careful evaluation. To address the advantages of the DNC method, an additional linear cooling experiment of the same duration was performed, and it was found that the DNC method resulted in a larger average crystal size. Overall, it can be concluded that the HPM method is suitable for DNC control and could be improved by modifying the image processing algorithm.
Open AccessBatch Cooling Crystallization of a Model System Using Direct Nucleation Control and High-Performance In Situ Microscopy
The aim of this study was to investigate the use of automated high performance in situ microscopy (HPM) for monitoring and direct nucleation control (DNC) during cooling crystallization. Compared to other techniques, HPM enables the detection of small crystals in the range of 1 to 10 μm. Therefore, a novel DNC-controlled variable was investigated to determine the potential improvement of the method. The laboratory system and process control software were developed in-house. A well-studied crystallization model system, the seeded batch cooling crystallization of α-glycine from water, was investigated under normal conditions and temperatures below 60 °C. It was postulated that length-weighted edge-to-edge counts in the range of 1 to 10 μm would be most sensitive to the onset of secondary nucleation and are therefore, used as a control variable. Linear cooling experiments were conducted to determine the initial setpoint for the DNC experiments. Three DNC experiments were then performed with different setpoints and an upper and lower counts limit. It was found that the DNC method can be destabilized with a low setpoint and narrow counts limits. In addition, the new controlled variable is highly sensitive to the formation of bubbles at the microscope window and requires careful evaluation. To address the advantages of the DNC method, an additional linear cooling experiment of the same duration was performed, and it was found that the DNC method resulted in a larger average crystal size. Overall, it can be concluded that the HPM method is suitable for DNC control and could be improved by modifying the image processing algorithm.
- Open Access
Development of a Calibration Model for Real-Time Solute Concentration Monitoring during Crystallization of Ceritinib Using Raman Spectroscopy and In-Line Process Microscopy
Raman spectroscopy is a useful tool for polymorphic form-monitoring during the crystallization process. However, its application to solute concentration estimation in two-phase systems like crystallization is rare, as the Raman signal is influenced by various changing factors in the crystallization process. The development of a robust calibration model that covers all variations is complex and represents a major challenge for the implementation of Raman spectroscopy for in-line monitoring and control of the solution crystallization process. This paper describes the development of a Raman-based calibration model for estimating the solute concentration of the active pharmaceutical ingredient ceritinib. Several different calibration approaches were tested, which included both temperature and spectra of clear solutions and slurries/suspensions. It was found that the concentration of the ceritinib solution could not be accurately predicted when suspended crystals were present. To overcome this challenge, the approach was enhanced by including additional variables related to crystal size and solid concentration obtained via in-line process microscopy (chord-length distribution percentiles D10, D50 and D90) and turbidity. Partial least squares regression (PLSR) and artificial neural network (ANN) models were developed and compared based on root mean square error (RMSE). ANN models estimated the solute concentration with high accuracy, with the prediction error not exceeding 1% of the nominal solute concentration.
Open AccessDevelopment of a Calibration Model for Real-Time Solute Concentration Monitoring during Crystallization of Ceritinib Using Raman Spectroscopy and In-Line Process Microscopy
Raman spectroscopy is a useful tool for polymorphic form-monitoring during the crystallization process. However, its application to solute concentration estimation in two-phase systems like crystallization is rare, as the Raman signal is influenced by various changing factors in the crystallization process. The development of a robust calibration model that covers all variations is complex and represents a major challenge for the implementation of Raman spectroscopy for in-line monitoring and control of the solution crystallization process. This paper describes the development of a Raman-based calibration model for estimating the solute concentration of the active pharmaceutical ingredient ceritinib. Several different calibration approaches were tested, which included both temperature and spectra of clear solutions and slurries/suspensions. It was found that the concentration of the ceritinib solution could not be accurately predicted when suspended crystals were present. To overcome this challenge, the approach was enhanced by including additional variables related to crystal size and solid concentration obtained via in-line process microscopy (chord-length distribution percentiles D10, D50 and D90) and turbidity. Partial least squares regression (PLSR) and artificial neural network (ANN) models were developed and compared based on root mean square error (RMSE). ANN models estimated the solute concentration with high accuracy, with the prediction error not exceeding 1% of the nominal solute concentration.
- Open Access
- 19 November 2023
- Published by Croatian Society of Chemical Engineers/HDKI in Kemija u industriji
- Vol. 72 (11-12) , 639-650
Primjena neuronskih mreža za procjenu koncentracije otopine ksilometazolin hidroklorida u n-butanolu primjenom ATR-FTIR spektroskopije in situ
Procesna analitička tehnologija (PAT) sve se češće primjenjuje u procesu kristalizacije za kontinuirano praćenje neke od ključnih procesnih veličina i značajki kvalitete proizvoda. Vrlo bitne procesne varijable za kristalizaciju hlađenjem su temperatura i koncentracija otopine. Kontinuirano mjerenje koncentracije omogućeno je naprednim in situ spektroskopskim instrumentima. U takve metode spada i prigušena ukupna reflektancija Fourier transformacijske infracrvene spektroskopije (engl. Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy, ATR-FTIR), primijenjena u ovom istraživanju. U ovom radu razvijeni su kemometrijski modeli za kontinuiranu procjenu koncentracije otopine u stvarnom vremenu metodama strojnog učenja. Cilj ovog rada bio je razviti i ispitati modele parcijalne regresije metodom najmanjih kvadrata (engl. Partial Least Squares Regression, PLSR) i modele neuronskih mreža (engl. Neural Networks, NN) za modeliranje ovisnosti koncentracije djelatne tvari, ksilometazolin hidroklorid, u n-butanolu, o temperaturi i spektralnim podatcima dobivenim mjerenjima ATR-FTIR spektrometrom. U radu je provedena predobrada prikupljenih podataka tehnikom MSC (engl. Multiplicative Scatter Correction) te je provedeno skaliranje podataka prema normalizacijama Min-Max i Z-score. Analiziran je broj neurona u prvom i drugom skrivenom sloju, broj skrivenih slojeva, vrsta primijenjenog algoritma učenja (ADAM, NADAM, RMSprop) kao i utjecaj vrste prijenosne funkcije (ReLU, sigmoid, tanh) na kvalitetu razvijenih neuronskih mreža. Razvijeni modeli na sva četiri skupa podataka dali su iznimno dobre rezultate s obzirom na iznose koeficijenta determinacije i srednje kvadratne pogreške. Modelom neuronske mreže dobiveni su koeficijenti determinacije u rasponu vrijednosti od 0,9979 do 0,9989, dok je srednja kvadratna pogreška od 0,0020 do 0,0011. PLSR modelom dobiveni su koeficijenti determinacije od 0,9990 do 0,9995, odnosno srednja kvadratna pogreška od 0,0009 do 0,0005. Dobiveni rezultati pokazali su da predobrada podataka i dodavanje drugoga skrivenog sloja neuronske mreže u ovoj studiji nisu imali velik utjecaj na konačne rezultate. Takav način praćenja i kontrole procesa rezultira efikasnijom proizvodnjom uz smanjenu vjerojatnost pogreške, što u konačnici farmaceutskoj industriji omogućuje brže plasiranje proizvoda na tržište.
- 19 November 2023
- Published by Croatian Society of Chemical Engineers/HDKI in Kemija u industriji
- Vol. 72 (11-12) , 639-650
Open AccessPrimjena neuronskih mreža za procjenu koncentracije otopine ksilometazolin hidroklorida u n-butanolu primjenom ATR-FTIR spektroskopije in situ
Procesna analitička tehnologija (PAT) sve se češće primjenjuje u procesu kristalizacije za kontinuirano praćenje neke od ključnih procesnih veličina i značajki kvalitete proizvoda. Vrlo bitne procesne varijable za kristalizaciju hlađenjem su temperatura i koncentracija otopine. Kontinuirano mjerenje koncentracije omogućeno je naprednim in situ spektroskopskim instrumentima. U takve metode spada i prigušena ukupna reflektancija Fourier transformacijske infracrvene spektroskopije (engl. Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy, ATR-FTIR), primijenjena u ovom istraživanju. U ovom radu razvijeni su kemometrijski modeli za kontinuiranu procjenu koncentracije otopine u stvarnom vremenu metodama strojnog učenja. Cilj ovog rada bio je razviti i ispitati modele parcijalne regresije metodom najmanjih kvadrata (engl. Partial Least Squares Regression, PLSR) i modele neuronskih mreža (engl. Neural Networks, NN) za modeliranje ovisnosti koncentracije djelatne tvari, ksilometazolin hidroklorid, u n-butanolu, o temperaturi i spektralnim podatcima dobivenim mjerenjima ATR-FTIR spektrometrom. U radu je provedena predobrada prikupljenih podataka tehnikom MSC (engl. Multiplicative Scatter Correction) te je provedeno skaliranje podataka prema normalizacijama Min-Max i Z-score. Analiziran je broj neurona u prvom i drugom skrivenom sloju, broj skrivenih slojeva, vrsta primijenjenog algoritma učenja (ADAM, NADAM, RMSprop) kao i utjecaj vrste prijenosne funkcije (ReLU, sigmoid, tanh) na kvalitetu razvijenih neuronskih mreža. Razvijeni modeli na sva četiri skupa podataka dali su iznimno dobre rezultate s obzirom na iznose koeficijenta determinacije i srednje kvadratne pogreške. Modelom neuronske mreže dobiveni su koeficijenti determinacije u rasponu vrijednosti od 0,9979 do 0,9989, dok je srednja kvadratna pogreška od 0,0020 do 0,0011. PLSR modelom dobiveni su koeficijenti determinacije od 0,9990 do 0,9995, odnosno srednja kvadratna pogreška od 0,0009 do 0,0005. Dobiveni rezultati pokazali su da predobrada podataka i dodavanje drugoga skrivenog sloja neuronske mreže u ovoj studiji nisu imali velik utjecaj na konačne rezultate. Takav način praćenja i kontrole procesa rezultira efikasnijom proizvodnjom uz smanjenu vjerojatnost pogreške, što u konačnici farmaceutskoj industriji omogućuje brže plasiranje proizvoda na tržište.
- Open Access
- 13 July 2023
- Published by Croatian Society of Chemical Engineers/HDKI in Kemija u industriji
- Vol. 72 (7-8) , 443-453
Proračun optimalnog temperaturnog profila hlađenja šaržnog kristalizatora
- 13 July 2023
- Published by Croatian Society of Chemical Engineers/HDKI in Kemija u industriji
- Vol. 72 (7-8) , 443-453
Open AccessProračun optimalnog temperaturnog profila hlađenja šaržnog kristalizatora