Search

Dear colleagues, we are extremely honoured to present to you the Proceedings of the international conference 19th Ružička Days, which was successfully held on September 21–23, 2022, in Vukovar, Croatia, in the hometown town of our famous Croatian scientist and first Nobel laureate, professor Leopold (Lavoslav) Ružička. The main goals of the Conference were to promote excellence, originality and innovation of interdisciplinary scientific research as well as the practical application of the obtained results through collaboration with industry, emphasizing thus the recognizable slogan of the Ružička Days conference: "Today Science – Tomorrow Industry". In addition, the Conference gave the opportunity for meetings, exchanging the ideas, opinions, experiences and cooperation among participants from different working surroundings. It is also important to point out that since 2008, within the international conference Ružička Days, Meeting of Young Chemists has also been held, where high school students, with the help of their mentors, present papers in the field of chemistry, with the aim to direct students and young people to the technical and natural sciences, especially chemistry. In its seventh issue (2011, 2013, 2015, 2017, 2019, 2021) the Proceedings publishes scientific and professional full papers of high quality in the following sections: Chemical analysis and synthesis (10), Chemical and biochemical engineering (3), Food technology and biotechnology (4), Chemistry in agriculture and forestry (1), Environmental protection (9) and 8th Meeting of Young Chemists (3). Full-length papers were subjected to an international review procedure done by eminent experts from the corresponding fields, to whom we express our gratitude, but they were not subjected to linguistic proofreading. On behalf of the Scientific and Organizing Committee of the 19th Ružička Days we cordially thank all the authors, reviewers, participants, lecturers, organizers, especially the international organizers EuCheMS and EHEDG, auspices and sponsors, and all the others who, in any way, supported the Conference and contributed to the preparation of the Proceedings, especially to our highly skilled and committed associates, who have put a lot of effort in the preparation of these Proceedings. At the very end, special thanks to our young, future scientists and their mentors who are faithful participants of the Meeting of Young Chemists of the Ružička Days conference. Enjoy the extremely interesting scientific and professional papers contained in these Proceedings, until the next 20th jubilee of Ružička Days in 2024! We are looking forward to meeting you again in Vukovar! Yours sincerely, Chief Editors Jurislav Babić Vesna Ocelić Bulatović Dajana Kučić Grgić

Crystallization is an important unit operation used in many branches of industry because it is energy efficient way of separating and purifying compounds.[1] It is one of the most important unit operations in the pharmaceutical industry. The industrial control of crystallization is a difficult task because of the stochastic and often unpredictable behavior of the system. With the development of process analytical technology crystallization control has become possible because it has enabled real-time measurement of critical quality attributes. ATR-FTIR spectroscopy has been used to track the concentration during crystallization [2] but the procedure of developing calibration models is time-consuming. The rapid development of the calibration model is done by slowly heating excess solute in a solvent and collecting temperature and spectra. It is assumed that for each temperature the solubility equilibrium has been reached and spectra correspond to the solubility concentration. In this work, the excess amount of glycine was heated from 5 °C to 55 °C during the span of 4.5 hours. The amount of glycine was 5 % higher than the solubility at 55 °C. The particles were grounded with pestle and mortar to achieve a small size of crystals that dissolve quickly. The spectra were recorded every minute together with the temperature. The PLS calibration model relating spectra to saturation temperature was built. The model was tested on an independent data set obtained by repeating the experiment. The results indicate that the model can predict saturation temperature with relatively good accuracy (RMSEP=0.73).

The aim of this work is to investigate the use of in-situ Raman spectroscopy for the monitoring of cooling crystallization process. For this purpose, Raman spectroscopy with a focused laser beam was utilized for the estimation of solution concentration and slurry density. The Raman signal depends on many factors, including the composition of the solid and liquid phase, as well as the size and shape of the crystals. [1] Therefore, the use of Raman spectroscopy for quantitative measurements is challenging and requires advanced chemometric methods. The concentration of an Active Pharmaceutical Ingredient (API), ceritinib was determined experimentally in a solution mixture of 90% acetone and 10% water. The goal was to build a calibration model for the estimation of ceritinib concentration during the crystallization process. Methods used to build the calibration models include Partial Least Square Regression (PLSR) and an Artificial Neural Network (ANN). During model development, various methods for spectra preprocessing were evaluated. In addition, different hyper parameters for the models were tested by ANN. The prediction performance used to evaluate the developed models was the root mean square error of prediction (RMSE). The model created with an artificial neural network proved to perform better than the partial least square regression model for the estimation of solution concentration, but were not able to predict slurry density with good accuracy.

Batch cooling crystallization is a widely employed unit operation in chemical engineering because it can separate and purify the product in a single step. It is used to produce food, chemicals, pharmaceuticals, and many high-value-added products. In the pharmaceutical industry, more than 90 % of active pharmaceutical ingredients (API) require crystallization in at least one step of the production process. To design a process, it is necessary to know the concentration of the slurry during crystallization. Process analytical technology can be used to measure the high concentration of solutions and is an inexpensive and reliable source of analytical information. In this work, a calibration model was developed by dissolving different amounts of the active pharmaceutical ingredient (Mr < 600 g/mol) in the solvent. The concentration range used for data collection was selected to cover the possible concentration range during cooling crystallization. The solutions were linearly cooled while the temperature and in-situ ATR-FTIR spectra were recorded. The recorded spectra were preprocessed using different techniques to determine which combination gives the best results. The region with the largest differences between the spectra was taken to develop the partial least squares (PLS) regression model. In addition, the artificial neural network (ANN) model was developed for comparison with the PLS model. The results indicate that the PLS and ANN models have similar performance and can be used to monitor the concentration of the active pharmaceutical ingredient during crystallization.

Soft sensor models for continuous estimation of 95% distillation point (D95) of a diesel product in a hydrocracking plant of refinery have been developed. Based on continuously measured data collected from the plant and laboratory analysis, data driven models have been developed using different machine learning methods. Data preprocessing included the selection of adequate data for model development, outlier’s detection, as well as data filtering. After selection of influential input variables, neural network and random forest models have been developed. By altering the number of hidden neurons and transfer function, neural network models results have been analyzed. The statistical performance criteria of developed models on validation data set indicate that the neural networks and random forest models estimate D95 of a diesel product reliably. By applying the present case study results in the refinery information system, a more stable operation of the plant, higher product quality and significant savings are expected.

Crystallization in pharmaceutical industry is used as separation and purification process for the production of active pharmaceutical ingredients(APIs), during which multiple laboratory analysis has to be done to ensure the required product quality. Until recently, critical quality attributes (CQA), including concentration, were determined off-line, but improvements in the field of fiber optics and spectroscopic instruments has enabled in-situ measurements in real-time, eliminating manual sampling and time delay influence affecting the analysis’ outcome. In this work, overview and good laboratory practice of spectroscopic techniques for solution concentration measurement is given. Ultraviolet-visible (UV-Vis) and Fourier transform infrared (FTIR) probes work on the principle of attenuated total reflectance (ATR) so that the incident light penetrates slightly into the solution and captures only the liquid phase of the crystallization system. Even though, ATR UV-Vis is cheaper than ATR FTIR, it is limited by the UV absorption capacity of the solute and hence not suitable for compound identification. In-line Raman spectrometers have the ability to simultaneously determine both, the solution and solid concentration using only one probe. Raman data is far more complex than IR and UV-Vis data, and extracting the information on the solution concentration is challenging due to high solid’s signal. The paper presents experiences and practical examples in the application of process analytical technology (PAT) methods for real-time monitoring and control of crystallization processes. The application of PAT requires a thorough knowledge of the crystallization system, the development of calibration models utilizing artificial intelligence and the application of advanced control strategies to obtain a high-quality product.

Large data sets generated and stored by continuous monitoring of process variables in process plants enable the development of data-driven mathematical models of a process. While such models do not provide detailed insight into the process, they do provide a feasible description of the nonlinear complex process dynamics. Data-driven models are often used to estimate important process characteristics in real-time. This paper presents an industrial case study of data-driven models based on multilayer perceptron (MLP) artificial neural networks (ANN) for the isomerization process of light naphtha in refineries. Models were developed for estimating critical quality characteristics of the product. Special attention was paid to the selection and analysis of data sets for different time periods to capture different process conditions. Optimal models were achieved by changing the types of learning function and transfer function, as well as the number of neurons in the hidden layer. Models based on MLP neural networks have shown better generalization capabilities compared to previously developed polynomial linear and nonlinear models for isomerization processes. This makes the developed MLP network models suitable for application to such a process, especially in the area of advanced process control.

Ovo je službeni dokument, samoanaliza Fakulteta kemijskog inženjerstva i tehnologije Sveučilišta u Zagrebu za reakreditaciju u 2022.

In this work, the development of the ANN-based calibration models using in-situ ATR-FTIR spectroscopy for estimating the xylometazoline hydrochloride concentration in the cooling crystallization process is presented.

Accurate measurement of solution concentration is of great importance for crystallization processes since supersaturation is prerequisite for nucleation and crystal growth. The development of process analytical technologies (PAT) enables the continuous measurements of process parameters for the production of high-quality drug products. The current solute concentration among with solubility curve provides information on supersaturation in the solution and thus allows implementation of feedback supersaturation control (SSC). The most commonly used method for in-situ concentration measurements is using ATR-FTIR spectroscopy. The IR absorbance of the dissolved compound is dependent on its concentration in the solution, as well as the temperature of the solvent. Therefore, a calibration is necessary to convert absorbance data into solution concentration [1]. In this work, ATR-FTIR spectra with the corresponding temperature of solutions with different concentrations are used to build two different calibration models: partial least squares regression (PLSR) and principal component regression (PCR) to monitor the concentration of glycine solution in real-time. Cross-validation was used to choose the right number of components to avoid overfitting the model. After visualizing ATR-FTIR spectra, the characteristic absorption bands were marked in the region between 1200 and 1800 cm-1 . The original dataset (W) and the subdivided dataset (C) were preprocessed using several different mathematical treatments (scaling, normalization, standardization and SavitzkyGolay smoothing) and used to build calibration models. Results show that, PLSR has better predictive performance in comparison to PCR with the same number of components used.

Vodeći svjetski sajam procesne industrije ACHEMA 2022 ugostio je više od 2 200 izlagača pristiglih iz više od 50 zemalja. Na sajamskom kompleksu u Frankfurtu od 22. do 26. kolovoza, nakon četiri godine, mogla se vidjeti najsuvremeniju opremu i nova inovativna rješenja kemijske, farmaceutske i prehrambene industrije. Fakultet kemijskog inženjerstva i tehnologije (FKIT) Sveučilišta u Zagrebu i Hrvatsko društvo kemijskih inženjera i tehnologa (HDKI) zainteresiranim studentima omogućili su posjet ACHEMA-i dodjelom financijske potpore.

Owning to the technological advancements in the recent decades, methods and systems relying on fast measurement provided by various PAT are extensively being developed with the purpose of monitoring and control of key process variables in real-time and are an integral part of the Quality by Design (QbD) and Quality by Control (QbC) concepts. The application of these concepts have proven to be beneficial both during the development stage and in the production stage of the API lifecycle in a way that they enable better process understanding, increased process robustness and full control over the Critical Process Parameters (CPP’s) and Critical Quality Attributes (CQA’s). The purpose of this research is to develop a system based on HDR microscopy, ATR-FTIR and Raman spectroscopy in combination with multivariate analytical methods that will enable continuous monitoring and control of crystallization processes. Furthermore, it will aid in finding the optimal crystallization conditions to obtain products of desired CQA’s. HDR microscopy and Raman spectroscopy are incorporated in a single probe used for measurements of chord length distributions, solute concentrations, solid loadings and the detection of polymorphic forms. FTIR is also employed via a separate probe to optionally monitor solute concentrations in real time. The goal is to develop a model free process control system where all of the CPP’s and CQA’s of a given crystallization process can be monitored and controlled in real time with little or no negative impact on the hydrodynamics of the system which is a common issue when multiple PAT tools are applied in monitoring processes at lab scale. The laboratory crystallization setup is comprised of a glass jacketed batch reactor connected to a thermostat unit and equipped with PAT tools. An in-house software platform was developed that enables real time collection of FTIR and Raman spectra together with the corresponding process temperature data. The platform is used to gather data for calibration model development and later on for process monitoring and control. It has proved to be stable and is capable of providing high quality linear and non-linear models used for monitoring and control.

The aim of this work was to investigate different control policies of the batch cooling crystallizer on the granulometric properties of obtained crystals. The study was conducted by computer simulation using population balance modeling in Simulink [1]. The Population balance was solved using the conservation element/solution element (CESE) method and consequently the dynamic evolution of crystal size distribution was obtained. The system under study consisted of 1 liter batch crystallizer with water as solvent and potassium nitrate as solute. The cooling system was not modeled because of the slow nature of the crystallization process and almost instant reaction of thermostat in real life application. The crystallizer was cooled from 32 °C to 20 °C with the starting concentration being saturation concentration at 32 °C. The crystallizer was seeded with 2 g of KNO3 and crystal size distribution of seeds was taken to be normal with mean 200 µm and standard deviation of 50 µm. Two temperature control policies were investigated, namely open loop linear cooling and closed loop (feedback) constant supersaturation control. For open loop linear cooling, different cooling rates were investigated and for the case of closed loop control, different supersaturations were investigated. The linear control policy had shorter operating time, but gave a significantly bigger ratio of nucleated crystals in the final product. Increasing the cooling rate resulted in a higher ratio of nucleated crystals which is considered to be a negative effect. On the other hand, constant supersaturation control had longer operating time, but resulted in a smaller ratio of nucleated crystals in the product. In conclusion, constant supersaturation feedback control is preferred because of improved granulometric properties of the product in comparison with linear open loop cooling policy.

Regulatory agencies, science and technology are forcing rapid change on manufacturing processes. Process Analytical Technology (PAT) is of major significance for the industry. PAT improves manufacturing described in the ‘good manufacturing practices (cGMP) for the 21st century’ initiative. The cGMP initiative stimulates process expertise and understanding. PAT offers companies in chemical, biochemical and food manufacturing the possibility to gain better control of their processes, also to introduce quality into the production process and eventually to move to real-time process control based on critical quality attributes. It brings major benefits in terms of product quality, reduced time to market, and tighter and more responsive supply chains. All industries are challenged to find optimal way to introduce PAT and to select the right tools and infrastructure to match manufacturing needs. A process is well understood when variability in the process is managed by process control that deliver a consistent process output ; and when the key product quality attributes can be accurately and reliably predicted. This lecture will explain how PAT tools can be used for continuous monitoring and process control in the process industry. The method of in-situ spectrophotometric measurements, as well as process microscopy in order to build calibration models will be presented. These models are further used for continuous process monitoring and control thus replace infrequent and occasional laboratory assays that are standardly used for product quality testing. The potential of this technology for the gradual transition from batch to continuous production will also be highlighted.

The batch cooling crystallization is widely used unit operation for separation and purification of a compound in a single step. It is widely applied in pharmaceutical, fine chemicals and food industry among the others. In pharmaceutical industry a large number of products (or intermediates) requires crystallization in at least one step of the process. A number of problems can occur due to excessive formation of fine crystals during the process (long filtering times, clogging of the equipment, cake compression) and because of that crystals generated by nucleation are considered undesirable. In this work we developed a program in Matlab using moment transformation of population balance as the process model. Crystallization kinetics for potassium sulphate was taken from literature because of good quality data and certain kinetics. An optimization problem was formulated to minimize ratio of the third moment of nucleation generated crystals to third moment of seed grown crystal. In addition, constraints were added in order to limit the duration of the process and ensure that appropriate yield is achieved at the end of the process. Optimization was performed first with genetic algorithm in order to find a solution which is near to the optimal solution. Afterwards, a sequential quadratic programming algorithm was applied to further improve the profile obtained by genetic algorithm. The solution obtained by sequential use of algorithms proves to be stable and satisfies conditions set by constraints. Therefore, the developed program can be used for calculation of the optimal temperature profile of batch crystallizer if the kinetics of the crystallizing compound is known.

Pozvano predavanje na skupu NETWORKING @ SIEMENS PHARMA.

Laboratorij za automatiku i mjerenja (LAM) na Fakultetu kemijskog inženjerstva i tehnologije (FKIT) Sveučilišta u Zagrebu čine laboratorijske instalacije sa suvremenim sustavom za vođenje koje obuhvaćaju tipične jediničnih operacija kemijske procesne industrije. Sve to podržano je najnovijom generacijom sustava za vođenje, simulacijskim softverima i cjelovitim softverskim rješenjem za postrojenja. Laboratorij nudi studentima, zaposlenicima i partnerima priliku za usvajanje temeljnih i naprednih znanja o procesnim mjerenjima i vođenju procesa primjenom najmodernije opreme i softvera. Osnovna namjena je educirati studente, doktorande i polaznike stručnih studija i cjeloživotnog obrazovanja o suvremenim mjernim uređajima, metodama vođenja procesa i analize mjernih podataka s postrojenja. U Laboratorij se provodi i europski strukturni projekt sa zadatkom naprednog vođenja procesa kristalizacije (CrystAPC) čiji je cilj, primjenom procesne analitičke tehnologije (PAT) razviti napredne metode vođenja procesa kristalizacije, kao i oformiti tim eksperata koji će provoditi primjenska istraživanja za industriju.

The subject of this research is the development of a calibration model based on in-situ ATR-FTIR measurements using artificial neural network (ANN) for monitoring the concentration of glycine in glycine-water system in a batch crystallizer. Supersaturation, as a driving force of batch cooling crystallization processes, has a strong effect on product properties, i.e. crystal morphology, purity and crystal size distribution.[1] Therefore, the application of process analytical techniques plays a vital role for real-time monitoring and controlling of crystallization processes in order to obtain high-quality products. Attenuated total reflectance - Fourier transform infrared spectroscopy (ATR-FTIR) is used for inline measurement of solute concentration. IR spectrum is characteristic of the vibrational structure of the substance in immediate contact with the ATR immersion probe.[2] To obtain useful information from the spectral data, a calibration model is needed. A calibration model for glycine concentration was developed based on experimental data. Experiments were conducted with different process conditions, including changes of solute concentration and the operating temperature. The developed ANN model has a multilayer perceptron structure. ANN was trained using FTIR spectral data with the corresponding temperature values collected in the laboratory crystallizer. Validation of the model was performed on an independent dataset not used during the neural network training. The application of the developed model for monitoring the concentration in real-time combined with solubility curve and metastable zone width provides information about supersaturation in real-time. That allows the application of a process control method for maintaining the desired degree of supersaturation and thus, obtaining uniform product properties.

U ovom radu modeliran je sustav za provedbu kemijskih reakcija i izmjenu topline u cijevnom reaktoru. U procesu izmjene topline glavna vođena varijabla je temperatura. Temperatura je ključna varijabla za stabilan rad i djelotvornost samog procesa izmjene topline te ostalih vezanih toplinskih i/ili separacijskih procesa. Također je bitna za potrošnju energije i pomoćnih medija te u konačnici za kvalitetu proizvedenih intermedijara i proizvoda. Razvijen je teorijski model procesa izmjene topline za različite režime rada izmjenjivača topline na temelju fundamentalnih jednadžbi, a parametri modela (koeficijenti prijenosa topline) određeni su iz empirijskih formula.[1] Uz to ispitan rad samog modela te su, simulacijom u Simulinku, određene vrijednosti temperatura u reaktoru, prilikom čega je simuliran i utjecaj predhlađenja reaktanata.

Analiza, sinteza i vođenje kemijskih procesa - stručna i znanstvena disciplina koja se bavi primjenom načela kemijskog inženjerstva u proučavanju pretvorbe sirovina u proizvode (analiza), odabirom i dimenzioniranjem opreme te formiranjem procesnih tokova (sinteza) i povezivanjem znanja automatike i kemijskog inženjerstva za postizanje konzistentnosti proizvoda, ekonomičnoga rada i sigurne proizvodnje (vođenje).

U radu je dan osvrt na promociju sveučilišnog udžbenika "Modeliranje u kemijskom inženjerstvu" autora Zorana Gomzija i Želimira Kurtanjeka.

Procesi prijenosa topline često se pojavljuju u industrijskoj praksi. Neki proizvodi u, primjerice, farmaceutskoj ili prehrambenoj industriji, vrlo su osjetljivi na promjene temperature, stoga je potrebno posvetiti veliku pažnju njihovom projektiranju i izvedbi. U ovom radu ispitan je laboratorijski uređaj za analizu dinamičkog vladanja toplinskog procesa. Uređaj je namijenjen demonstraciji razvoja dinamičkih modela procesa na temelju eksperimentalnih podataka, a namjena mu je poučavanje osnovnih metoda vođenja procesa i ugađanja regulatora. U radu je proveden niz eksperimenata kojima su utvrđene statičke i dinamičke karakteristike procesa i regulacijskog kruga. Na temelju provedenih eksperimenata proračunati su i optimirani parametri regulatora. Konačno je ispitano vladanje procesa u regulacijskom krugu i ocijenjen rad regulacijskih krugova.

Kristalizacija je jedan od najvažnijih, no najsloženijih koraka u procesu dobivanja raznih farmaceutskih proizvoda. Na kvalitetu tih proizvoda utječu mnogi čimbenici koje je potrebno konstantno nadzirati. Jedan od čimbenika je i raspodjela veličina čestica kristala koja je veoma bitna za procese koji slijede nakon kristalizacije (filtracija i sušenje), primjenska svojstva te vijek trajanja proizvoda. Budući da postoji težnja za unapređenjem vođenja cijelog procesa, u sustav se nastoje uvesti uređaji i metode praćenja promjena svih čimbenika pa tako i raspodjele veličine čestica. Raspodjelu veličina čestica u stvarnom vremenu može se pratiti sondom za mjerenje refleksije usredotočene zrake (engl. Focused Beam Reflectance Measurement, FBRM). Rezultat mjerenja FBRM sonde je linijska raspodjela veličina čestica (LRVČ) koju je, za pravilno tumačenje, potrebno korelirati s volumnom raspodjelom veličina kristalnih čestica (VRVČ). Za određivanje korelacije mogu se upotrijebiti teorijski ili empirijski modeli. U istraživanju se primjenjuje empirijski pristup modeliranju. Korištene su statistička metode glavnih komponenti (engl. Principal Component Analysis, PCA) i parcijalne regresije najmanjih kvadrata (engl. Partial Least Squares Regression, PLSR).

Fosamprenavir‐kalcij koristi se u liječenju HIV‐ a, a djeluje kao prolijek amprenavira Metaboličkim djelovanjem u ljudskom tijelu iz fosamprenavira nastaje amprenavir koji služi kao inhibitor proteaze humanog virusa imunodeficijencije (HIV) s antivirusnim djelovanjem. Lijek se primjenjuje u tom obliku jer se tako sporije oslobađa u ljudskom tijelu te je njegovo djelovanje produljeno. Molekula fosamprenavir‐ kalcija može postojati u različitim kristalnim i amorfnim oblicima ovisno o načinu pripreme što je opisano u više patenata. U ovom istraživanju koristio se kristalni oblik 1. Odabir otapala za provedbu rekristalizacije proveden je pomoću uređaja Crystal16. Cilj je bio odrediti otapalo koje pokazuje zadovoljavajuću ovisnost koncentracije otopljene tvari o temperaturi sustava. Voda, izopropanol, aceton, acetonitril, etanol gotovo ne otapaju ispitivanu djelatnu tvar (manje od 1 %) niti pri povišenim temperaturama, dok se u metanolu topljivost povećava s povišenjem temperature (pri T = 35 °C Xs > 150 mg ml–1). Za metanol i mješavinu metanola i vode (19:1) određene su širine metastabilne zone. Dodatak vode znatno povećava širinu metastabilne zone od 15 °C, koliko iznosi za čisti metanol, na 25 °C. Kristalizacija hlađenjem provedena je za sustav fosamprenavir‐kalcij – metanol, dok je kristalizacija precipitacijom provedena za sustav fosamprenavir‐kalcij – metanol uz dodatak vode i etanola kao antiotapala. Rezultati različitih načina provođenja kristalizacije prikazani su kroz usporedbu XRD‐snimaka, mikroskopskih slika kristala i raspodjele veličina dobivenih kristala. Istraživanje je provedeno kao preliminarno ispitivanje sustava koji će se primijeniti u daljnjem radu, koji se odnosi na mogućnost on‐ line praćenja raspodjele veličine kristala u procesu kristalizacije. Krajnji cilj istraživanja je razvoj napredne metode za vođenje procesa kristalizacije koja bi osiguravala željenu raspodjelu veličina kristala na kraju procesa.

The subject of the research is development of a calibration model for a real-time measurement of the crystal size distribution (CSD) in laboratory crystallizer. CSD is crucial for product quality, i.e. process efficiency, as well as energy and utility consumption in subsequent batch production processes. In addition, the continuous measurement of CSD would have an important application in the transition from batch to continuous production which is trend in pharmaceutical industry. Focused Beam Reflectance Measurement (FBRM) is a precise and sensitive technology for measuring particle size, number and shape changes. The measured data can be used for optimization and understanding of the process, and potentially for the crystallization process control. Output data of the FBRM instrument is Chord Length Distribution (CLD), which is dependent of the actual CSD, but the two are not identical. Calibration model for the conversion of CLD to CSD was developed using experimental data. The developed model is based on multilayer perceptron neural network. Neural network model is trained on the FBRM data collected experimentally in the laboratory crystallizer and analytically determined CSD by the laser diffraction. Validation of the model was performed on an independent data set not used for the neural network training. The real-time application of developed model would allow monitoring of the CSD during the course of process. The main advantage of developed method is in the fact it does not require theoretical knowledge of the FBRM operation, and allows monitoring of CSD for irregularly shaped particles. Real-time CSD information is prerequisite for the development of real-time CSD control methods, which would consequently ensure production of the crystalline product with consistent quality.