Method and system for monitoring biological products and method for predicting microbial load in biological products

Abstract

The present invention relates to a method and system for monitoring biological products, and also to a method for generating a prediction of microbial load in biological products, comprising at least one synchronous electronic sensor (1) equipped with a data transmission device (11) in data communication with a gateway (2) via wireless networks and designed to be positioned near to or in direct contact with a biological product (7); or at least one asynchronous sensor (1') equipped with a removable storage unit (11'); a gateway (2) connected via the internet (3) to the data management module (4), wherein said module (4) comprises a trained artificial neural network and a data storage unit (42) both operated in a virtual environment from a cloud server (5) or on a local physical server (5') and connected to the user interface unit (6).

Description

[0001] METHOD AND SYSTEM FOR MONITORING BIOLOGICAL PRODUCTS AND METHOD FOR PREDICTING MICROBIAL LOAD IN BIOLOGICAL PRODUCTS

[0002] Technical sector

[0003]

[0001] The present invention belongs to the technological sector of predictive microbiology, more specifically it refers to a method and system for monitoring biological products and also a method for predicting microbial load in biological products.

[0004]

[0002] The method involves entering information about (the biological product, the microorganism to be monitored, the intrinsic and extrinsic properties to be monitored) into a computer program, which may be installed on a local network or in the cloud. The system uses temperature sensors positioned in the vicinity of or in direct contact with the biological product to be monitored, which transmit data in real time to a server, either local or in the cloud, via a wireless network. This information is processed in a virtual environment that houses an automated calculation application and, through mathematical models, generates predictions for the microbial load over time.

[0005]

[0003] The results of these predictions for microbial load kinetics are presented to the user through a graphical interface, accessible via browser or mobile application, allowing for decision-making regarding the temperature, packaging, and storage of the biological product. The predictions provide data on shelf life, packaging conditions, and quality of biological products, as well as the kinetics of microbial populations applicable to both heat treatment and fermentation processes, as well as quality control of transported material, ensuring safety and efficiency in different application scenarios.

[0006] State of the art

[0007]

[0004] Controlling the growth and inactivation of microorganisms in products of biological origin represents a critical challenge to ensure the safety and quality of these products. Products of biological origin, or biological products, are defined as those derived from living organisms. This includes a wide range of items, such as raw or processed foods, products for animal use, enzymatic preparations, and biological medicines, stored under refrigerated or non-refrigerated conditions.

[0008]

[0005] Microorganisms that undesirably contaminate biological products can be classified as pathogenic or spoilage microorganisms. Controlling pathogenic microorganisms is a public health concern, as they are directly linked to the transmission of diseases. Although spoilage microorganisms do not cause disease, they contribute to the degradation of biological products, resulting in the loss of quality characteristics such as texture, color, and flavor. Discarding products contaminated by microorganisms leads to financial losses for producers and retailers and contributes to the waste of natural resources and the emission of greenhouse gases resulting from the organic decomposition of these products.

[0009]

[0006] On the other hand, some microorganisms can be intentionally added to biological materials to cause desirable modifications. They are used, for example, in the production of fermented foods, or for the synthesis of bioproducts derived from microbial cultivation. Monitoring the evolution of the microbial population throughout these processes allows for its control and optimization, increasing productivity and ensuring product quality. In these matrices, effective control of the microbial population is essential to promote safe, economical, and sustainable practices.

[0010]

[0007] Microbial kinetics encompasses the study of the growth, metabolism, and death rates of microorganisms, playing an essential role in the quality of biological products. Mathematical models can be used to predict microbial kinetics, a field of study within predictive microbiology, through the analysis of the following factors:

[0011] - intrinsic properties (related to the characteristics of the product, such as pH and water activity, i.e., the amount of free water for microbial growth and the development of degradation reactions);

[0012] - extrinsic properties (related to the environment, such as temperature and relative humidity); and

[0013] - implicit properties (related to competition with the microbiota naturally present in the product).

[0014]

[0008] Among extrinsic properties, temperature is especially relevant to microbial kinetics. Temperatures below the optimal growth range of a microorganism can retard or inhibit its growth. This is why refrigeration is a common measure to prevent the proliferation of microorganisms. Temperatures within the microorganism's optimal growth range, generally close to room temperature, accelerate microbial growth, increasing the microbial load and product spoilage. Finally, very high temperatures, above the optimal growth range, are used to eliminate microorganisms and extend the shelf life of biological products. In addition, temperature affects the production of metabolites (such as toxins or decomposition products) by microorganisms.

[0015]

[0009] Based on this, it becomes evident that temperature control plays a crucial role in preserving the quality and safety of biological products. In fact, in industrial thermal processes, the use of high temperatures is responsible for reducing the microbial load of bioproducts to acceptable levels. Furthermore, products are normally subjected to the cold chain, which corresponds to their handling and storage at controlled temperatures until final consumption. These practices ensure the safety of the products, delay or prevent their deterioration, extend shelf life, and reduce economic losses associated with the disposal of deteriorated items. In the present invention, shelf life is defined as the useful lifespan during which a product can be preserved under certain environmental conditions undergoing minor changes that are considered acceptable by the manufacturer, the consumer, and current legislation.

[0016]

[0010] It is worth noting that, in addition to temperature, other extrinsic properties, such as relative air humidity, and intrinsic properties, such as product pH, can fluctuate along the supply chain, directly influencing microbial kinetics. Monitoring and controlling these additional factors can be important for ensuring the quality and safety of biological products.

[0017]

[0011] Traditional predictive microbiology models often fail to adequately capture the non-linear interactions between intrinsic and extrinsic factors involved in microbial kinetics. In this sense, artificial neural network models based on machine learning emerge as a promising alternative. Artificial Intelligence is a field of computer science dedicated to creating computational routines capable of performing tasks that normally require human intelligence. This includes skills such as learning, reasoning, perception, pattern recognition, decision-making, and natural language understanding. One of the most prominent areas of Artificial Intelligence is machine learning, which are algorithms that allow computational systems to improve via training data without explicit programming.Machine learning is able to recognize and model non-linear patterns that influence microbial kinetics in a highly effective way by identifying complex interactions that escape traditional models.

[0018]

[0012] Patent documents CN108254513B and CN109816150A describe the use of neural networks to predict the shelf life of biological products, based on the measurement of specific variables and their correlation with shelf life. Document CN108254513B applies this concept to predicting the shelf life of fish, using gas measurements to correlate with microbial counts. Document CN109816150A refers to predicting the shelf life of fresh grapes post-harvest, using criteria such as product appearance, temperature and storage conditions, and microbial counts to assess shelf life. Both documents propose systems and processes that use neural networks to predict the quality of biological products along a temperature chain during the monitoring period.

[0019]

[0013] Document KR101864252B1, in turn, describes a system that uses artificial intelligence to predict the contamination index of food by pathogenic microorganisms, based on the composition of the products and the presence of specific microorganisms under a constant ambient temperature.

[0020]

[0014] Although the systems and methods described in these documents present relative practicality in modeling the prediction of product quality over time, they are limited to monitoring specific types of biological products. Each patent deals with a single product, and even document KR101864252B1, which proposes monitoring food on a broader scale, remains restricted to controlling bacteria that cause food poisoning, without considering the monitoring of microorganisms beneficial to the product or spoilage microorganisms.

[0021]

[0015] Patent document US2022 / 147755A1 describes a system that uses image-based neural networks to estimate microbial susceptibility and the minimum inhibitory concentration (MIC) of antimicrobials for different microorganisms. Similarly, patent EP41250641 uses neural networks to predict the susceptibility of microorganisms to antimicrobials through image processing. However, the need for image capture and processing makes the prediction process considerably more expensive and complex. The installation of cameras and image processing hinder the implementation of this technology in biological product production and distribution environments, reducing the feasibility of its application in these scenarios.

[0022]

[0016] Patent application BR 10 2020 010345 describes a system for detecting microbiological loads in various biological products, using readers and sensors based on changes in the electromagnetic properties of biosensors via RFID tags. Each RFID sensor functions as a chemically treated biosensor to improve the selective analytical response. Upon detecting microorganisms, the data is transmitted to a computer program with artificial intelligence that performs real-time analyses using mathematical calculations and algorithms.

[0023]

[0017] Although the system is designed to monitor various types of products, it relies on chemically treated sensors and a specific handheld device for its operation. This device requires sample preparation on a plate containing a receiver slide, which makes the diagnostic process more complex and less practical. Furthermore, the system does not operate autonomously with simple sensors, such as temperature sensors, which limits its functionality and practicality in continuous monitoring environments.

[0024]

[0018] Document US2024 / 093262 describes a monitoring system that uses sensors to assess in real time the microbial population, its health, and metabolic activity in aquatic and industrial environments. This biosensor measures water quality in natural and artificial systems, detecting variations in microbiological conditions. Integrated with artificial intelligence, it performs continuous parameterization and data analysis. However, the system is restricted to aqueous environments, using selectively permeable membranes to collect microbial gas signals, which are sent to sensors that measure the concentration of gaseous compounds, without predicting future changes in the microbial population.

[0025]

[0019] Documents US2024 / 209306 and US2021 / 262047 describe bioreactor monitoring and optimization systems with different approaches. The first uses integrated sensors to measure parameters such as temperature, pH, dissolved oxygen, and nutrient rates, sending data to machine learning software that adjusts cultivation conditions in real time. The second combines a kinetic model of cell growth with a metabolic model, monitoring metabolite consumption and production. Both are limited to controlled environments, such as bioreactors, due to the reliance on specific sensors and the need for direct contact with the product being monitored, restricting their application to closed systems.

[0026]

[0020] Patent application US11748766B2 describes a monitoring system and method that provides a coding for product quality along a cold chain. The patent proposes the use of temperature sensor data, among other information, in artificial intelligence models to assess product shelf life, generating qualitative classifications such as A (high remaining shelf life), B (intermediate shelf life), and C (product nearing perishability). However, the system has limitations regarding real-time microbial load prediction, as the quality estimate is only performed after the product's journey, associating the data collected by the sensors with expert and consumer opinions obtained after receiving the product.

[0027] New features and purpose of the invention

[0028]

[0021] The present invention aims to provide a method and system for monitoring biological products based on temperature or other product and environmental properties, and also a method for predicting microbial load in biological products that effectively resolves the limitations of the prior art mentioned above.

[0029]

[0022] The system used to monitor biological products comprises at least one synchronous electronic sensor, such as those that detect temperature and continuously monitor the product's storage conditions. The sensor is equipped with a wireless data transmission device that sends the data to a gateway, which in turn communicates via the internet with the data management module. The data management module executes the steps of the method for predicting the microbial load of the monitored product in a cloud-based virtual environment and comprises a previously trained machine learning-based artificial neural network, a data storage device, and a user interface module. Optionally, the electronic sensors can be equipped with a removable storage unit for collecting the monitored data. Furthermore, the virtual environment can be run from a local physical server.

[0030]

[0023] The biological product monitoring method begins with the input of mandatory data (type of microorganism and biological product) and optional data on intrinsic factors (pH and water activity, for example) and extrinsic factors (for example, relative humidity). This data will be entered by the user directly in the visual interface. The next step involves the real-time collection and transmission of the monitored variables by at least one synchronous electronic sensor equipped with a data transmission device positioned near the biological product or optionally in direct contact with it. The collected data is sent wirelessly to the gateway, which forwards the received data to the data management module in the cloud-based virtual environment. The data management module uses an artificial neural network algorithm based on machine learning.In this way, the data management module processes and stores the information in its own database, and simultaneously transmits it to the user interface module, which can be accessed through a web browser.

[0031]

[0024] The steps involved in the method for predicting microbial load in biological products are performed in the data management module and include receiving data obtained by the sensors and forwarding it to the artificial neural network algorithm. The artificial neural network algorithm is based on machine learning and was previously trained using information available in a database with kinetic curves of microbial load containing information annotated by experts in the field.The algorithm coefficients are estimated during pre-training, and the model takes as input the product type, microorganism type, and intrinsic and extrinsic properties (provided in the user interface module) in association with data obtained from electronic sensors; and as output, it provides the results of the calculations performed, providing a quantitative estimate of the microbial load being monitored, thus generating a prediction for the increase, reduction, or inactivation of the microbial load in the monitored product. Therefore, this method is capable of generating estimates of the growth or inactivation of the microbiological load present in the monitored product in real time, contributing to better decision-making.

[0032]

[0025] Optionally, the method for predicting microbial load in biological products can generate automatic alerts when the microbial load approaches or exceeds predefined limits, such as shelf life, notifying users in real time to ensure the maintenance of product quality and safety. This information will then be sent to the user interface module.

[0033] Advantages and technical effects of the invention

[0034]

[0026] The method and system for monitoring biological products and the method for predicting microbial load in biological products, objects of the present invention, result in the following advantages and achieve the following technical effects over prior art systems:

[0035] - Versatility: The method and system were designed to produce predictive reports on microbial kinetics, allowing its application to a wide variety of biological products, such as food, pharmaceuticals, and fermented products. Therefore, a wide variety of products of biological origin can be monitored using the same method and system.

[0036] - Real-Time Monitoring: Unlike existing solutions, the method and system described allow for continuous monitoring of product conditions, generating reports predicting microbial kinetics at predetermined intervals. This speeds up decision-making and enables rapid adjustments to preserve product integrity;

[0037] - Proactive Management: The system facilitates the efficient control of critical safety processes, such as the elimination of pathogens in heat treatments, ensuring compliance with health regulations;

[0038] - Logistics Optimization: Real-time monitoring improves logistics agility, allowing for quick adjustments in production, distribution, and storage processes;

[0039] Accurate Predictions: The system generates reports that provide quantitative estimates of the microorganisms present in a given biological product, ensuring greater accuracy in quality and safety monitoring.

[0040] - Accessibility and Ease of Operation: The method does not require the use of sophisticated sensors or cameras that demand complex and expensive installations, as in existing systems. This makes implementation more economical and feasible for different industries;

[0041] - Broad Applicability: The system is applicable not only to perishable products during food preparation or transportation, but also to industrial heat treatment processes, fermentation processes, and logistical processes such as storage and transportation.

[0042] List of attached drawings

[0043]

[0027] In order that the present invention may be fully understood and put into practice by any technician in this technological sector, it is described in a clear, precise and sufficient manner, based on the attached drawings listed below:

[0044] Figure 1 - Diagram of the biological product monitoring system using a synchronous sensor;

[0045] Figure 2 - Diagram of the biological product monitoring system using an asynchronous sensor;

[0046] Figure 3 - Flowchart of the method for monitoring biological products using a synchronous sensor;

[0047] Figure 4 - Flowchart of the method for monitoring biological products using an asynchronous sensor; Figure 5 - Flowchart of the method for predicting microbial load in biological products.

[0048] Detailed description of the invention

[0049]

[0028] For context, it is known that at high temperatures (typically above 50°C, depending on the microorganism) microbial inactivation occurs, resulting in a reduction in the number of microorganisms. Conversely, when temperatures are close to room temperature (on average between 20 and 45°C), microorganisms proliferate, leading to an increase in the microbial load. The microbial reproduction rate peaks within an ideal temperature range for growth, which varies according to the microorganism species in question, and is reduced at temperatures that deviate from this range, whether higher or lower.The invention proposes a method and system for monitoring biological products, and a method for predicting the rate of reduction, inactivation, or reproduction of a microbial population based on the temperature measured by a sensor, allowing the monitoring and adjustment of microbiological behavior according to the observed thermal conditions.

[0050]

[0029] The selection of the sensor for a given application must take into account the characteristics of the product, the range of environmental conditions to which the product will be subjected, as well as the parameters that the electronic sensor must measure, whether they are temperature, relative humidity, or other parameters. For example, if the relative humidity of the environment is relevant to the evolution of microbial kinetics in the product in question, the sensor must be able to measure it at the levels typically observed.

[0051]

[0030] Different electronic sensors, such as thermocouples, thermistors, and thermo-hygrometers, can be used to monitor the desired properties. The described system can operate with any sensing device capable of accurately measuring these properties and transmitting or storing the collected data. For illustrative purposes, the use of a synchronous electronic sensor (1), equipped with a data transmission device (11), will be described.

[0031] Figure 1 illustrates the synchronous system employing the biological product monitoring method. The system comprises at least one synchronous electronic sensor (1) equipped with a data transmission device (11) with communication established via wireless network with a gateway (2). Said gateway (2) maintains real-time communication via the internet (3) with the data management module (4) which is operated in a virtual environment from a cloud server (5).The aforementioned data management module (4) is responsible for receiving and processing the data obtained by the synchronous electronic sensor (1) as well as for executing the method for predicting the microbial load in the biological product being monitored.

[0052]

[0032] The data management module (4) comprises a previously trained artificial neural network and a data storage device (42) to store the data collected by the synchronous sensor (1) as well as the output data containing predictions about the microbial load. Furthermore, the data management module (4) sends in real time via the internet (3) the generated prediction about the microbial load to the user interface module (6) which can be accessed through a web browser or mobile device application. Said synchronous electronic sensor (1) is designed to be positioned in the vicinity of or directly in contact with the biological product (7) to be monitored.

[0053]

[0033] By way of example, we can consider the shelf-life monitoring of a biological product (7) when it is packaged. In this case, the area close to the biological product (7) can be considered the ideal location for positioning the synchronous electronic sensor, close to the packaging. On the other hand, in situations where the biological product (7) is inside a bioreactor, such as in fermentation processes, the sensor can be positioned inside the bioreactor itself, coming into direct contact with the biological product (7) being monitored.

[0054]

[0034] Figure 2 illustrates the asynchronous monitoring system where optionally at least one asynchronous electronic sensor (1') can be used, equipped with a removable storage unit (11'). In this case, the data collected by the asynchronous electronic sensor (T) are stored in the removable storage unit (11') and manually transferred to the data management module (4) at predetermined times. Optionally, the data management module (4) can be operated from a local physical server (5').

[0055]

[0035] Likewise, the data management module (4) comprises a previously trained artificial neural network and a data storage device (42) to store the data collected by the asynchronous sensor (T) as well as the output data containing predictions about the microbial load. Furthermore, the data management module (4) has a connection to the user interface module (6) which, for example, can be established via wireless networks or alternatively via local physical cabling. The said asynchronous electronic sensor (T) is designed to be positioned in the vicinity of, or directly in contact with, the biological product (7) to be monitored.

[0056]

[0036] Figure 3 details the method for monitoring biological products that employs the use of at least one synchronous temperature sensor (1) equipped with a data transmission device (11) comprising the following steps: a) The user fills in the user interface unit (6) of the data management module (4) the mandatory input data (type of microorganism and biological product); optional input data of intrinsic factors (pH and water activity, for example) and extrinsic factors (for example, relative humidity); b) The user positions the synchronous electronic sensor (1) in the vicinity of the biological product (7), or optionally in direct contact with the biological product (7); c) The synchronous electronic sensor (1) continuously collects temperature data in the vicinity of the biological product (7);d) The data transmission device (11) continuously transmits the data collected by the synchronous electronic sensor (1) via wireless network to the gateway (2); e) The gateway (2) transmits, via internet (3) to the data management module (4), the data received from the data transmission device (11) of the synchronous electronic sensor (1); f) The data management module (4) executes the steps of the method for predicting the microbial load of the biological product (7); g) The data management module (4) sends in real time the results obtained in the method for predicting the microbial load to the data storage device (42) and user interface unit (6).

[0057]

[0037] Figure 4 illustrates the monitoring method where, optionally, at least one asynchronous temperature sensor (T) equipped with a removable storage unit (11') and a local physical server (5') can be used. In this case, the method comprises the following steps: a') The user fills in the mandatory input data (type of microorganism and biological product) in the user interface unit (6) of the data management module (4); optional input data of intrinsic factors (pH and water activity, for example) and extrinsic factors (for example, relative humidity); b') The user positions the asynchronous electronic sensor (T) on the biological product (7), or optionally in direct contact with the biological product (7); c') The asynchronous electronic sensor (T) continuously collects temperature data in the vicinity of the biological product (7);d') The removable storage unit (11') of the asynchronous electronic sensor (T) continuously receives and stores the collected data; e') The user removes the removable storage unit (1T) from the asynchronous electronic sensor (T) containing the collected data; f) The user connects the removable storage unit (11') to the local physical server (5') for data transmission to the data management module (4); g') The data management module (4) executes the steps of the method for predicting microbial load in the biological product (7); h') The data management module (4) sends the results obtained in the method for predicting microbial load to the data storage device (42) and user interface unit (6).

[0058]

[0038] Artificial neural networks are algorithms inspired by the functioning of the human brain that comprise several layers of interconnected artificial neurons. These networks are designed to recognize patterns and perform complex tasks, such as, in the case of the present invention, calculating predictions based on variables collected by synchronous electronic sensors (1). These variables are applied in specific formulas of predictive microbiology, allowing the prediction of the evolution of the microbial load. For this purpose, it is possible to use, for example, the MLP (Multilayer Perceptron) artificial neural network.

[0059]

[0039] Furthermore, specialized databases in the areas of biology and predictive microbiology, such as COMBASE (www.combase.ee), are widely available on the internet. These databases contain information generated by experts on the behavior of microorganisms under different environmental conditions, and are frequently updated with scientific articles and empirical data. Such databases can be used to train the artificial neural network applied in the present invention, allowing the artificial neural network to process the variables and perform the calculations necessary to accurately predict the behavior of microbial loads in monitored biological products in an automated way and with the provision of real-time information, assisting in risk management and ensuring the safety and quality of biological products.

[0060]

[0040] To exemplify the different information generated by the model, consider the effects of temperature on microbial behavior. At temperatures high relative to the optimum growth temperature, microorganisms become inactivated, resulting in a reduction of the microbial load. The mathematical model predicts this inactivation rate based on temperature readings from the sensors, where higher temperatures correspond to higher inactivation rates. These predictions are essential for optimizing and monitoring the effectiveness of thermal processes and ensuring product safety. On the other hand, at temperatures close to ambient, microbial reproduction increases, with the microbial load growing. The model also predicts this behavior, considering that the reproduction rate reaches its maximum at an optimum growth temperature, which varies according to the species analyzed.

[0061]

[0041] Although artificial neural network training is widely used in various fields, the following steps highlight the particularities of this process for the present invention, applied to the development of the model for the kinetic modeling of microbial load: a) Identification of target microorganisms: Specific microorganisms related to the product are selected, such as pathogens or spoilage organisms. For example, Salmonella for egg-based products or Clostridium botulinum for vacuum-packed products; b) Data collection: Data are gathered from various sources, such as experiments, scientific literature and databases, covering variables such as temperature, humidity and product properties (pH, water activity, etc.).c) Data preprocessing: The data is standardized, outliers are removed, and incomplete data is excluded to ensure the uniformity and integrity of the dataset; d) Dataset division: The dataset is randomly and balancedly divided into three parts: training, validation, and testing, to ensure the generalization of the model; e) Model training: Supervised learning algorithms are used to train the model, based on input data such as temperature and pH, to predict microbial counts over time; f) Model validation: The model is validated with separate data (validation dataset), adjusting hyperparameters through cross-validation, to ensure that the model is robust and avoids overfitting; g) Model testing: Performance is evaluated with independent test data (test dataset), using metrics such as mean squared error (MSE) and coefficient of determination (R²).2 ), to ensure its accuracy; h) Final adjustments and integration: if necessary, adjustments are made to the model. Once approved, the model is integrated into the data management system, generating microbial load predictions based on sensor parameters.

[0062]

[0042] This systematic and rigorous process allows the development of robust and accurate models to predict the evolution of microbial load in biological products, offering a reliable tool to optimize industrial processes and ensure microbiological safety.

[0063]

[0043] Figure 5 details the steps of the method for predicting microbial load in the biological product performed by the data management module (4) which comprises the following steps:

[0064] I. The data management module (4) receives data from the synchronous electronic sensor (1) or asynchronous electronic sensor (T);

[0065] II. The data management module (4) feeds the artificial neural network algorithm with data on microorganisms, on the product and on intrinsic and extrinsic properties entered by the user in the user interface unit (6);

[0066] III. The data management module (4) feeds the artificial neural network algorithm with the temperature data from the synchronous electronic sensor (1);

[0067] IV. Based on the data entered in steps II and III, the model generates quantitative values ​​for the microbiological load as results;

[0068] V. The data management module (4) sends the generated results in real time to the data storage device (42);

[0069] VI. The data management module (4) provides real-time quantitative predictions of the microbial load in the user interface unit (6).

[0070]

[0044] Optionally, the data management module (4) can generate automatic alerts implemented to notify the user when the microbial load approaches or exceeds predefined limits.

[0071]

[0045] The present invention can be applied to the monitoring of food during transport and storage, for example in a dairy industry that stores pasteurized milk in refrigerated chambers before distribution. The shelf life of milk is strongly influenced by the presence of spoilage microorganisms, making continuous monitoring of these agents essential to guarantee product quality.

[0072]

[0046] The process begins with the user entering into the data management module (4) via the user interface unit (6) the type of microorganism (in this case, spoilage microorganisms) and the type of product (pasteurized milk), as well as relevant intrinsic and extrinsic parameters. Synchronous electronic sensors (1) are installed in the milk packaging to monitor the temperature in real time. This data is transmitted by the data transmission device (11) to a local gateway (2), which forwards it via the internet (3) to the data management module (4).

[0073]

[0047] In the data management module (4), an artificial neural network algorithm uses the variables collected by the synchronous electronic sensor (1) and inserts them into the previously trained model. With this, the necessary calculations are performed to estimate microbial growth.

[0074]

[0048] Temperature and microbial load data are stored in the data storage unit (42) and made available to the team via the user interface unit (6), accessible via web browser or mobile application. The data management module (4) can generate automatic alerts if temperature or microorganism levels reach critical limits before the estimated expiration date. For example, if a synchronous electronic sensor (1) detects a refrigeration failure, the module will send an alert indicating that the microbial load is rapidly increasing, allowing corrective actions to be taken, such as transferring the batch to a functional chamber or accelerating its distribution.

[0075]

[0049] If transport is carried out in a truck equipped with a gateway (2), the data management module (4) can continuously monitor the microbial load throughout the entire logistics chain. In situations of temperature abuse, such as when doors remain open for long periods, the module will send alerts about possible compromises in product quality.

[0076]

[0050] The object of the present invention is also applicable to products derived from fermentation processes, such as bread, beer, and yogurt. Microbial kinetics are crucial to ensure the quality and uniformity of the final product. For example, in the beer production process using the yeast Saccharomyces cerevisiae.

[0077]

[0051] The process begins with the user entering into the data management module (4) via the user interface unit (6) information about the microorganism (Saccharomyces cerevisiae) and the type of product (beer), as well as intrinsic and extrinsic parameters relevant to predicting the microbial load. Synchronous electronic sensors (1) are installed in the fermentation tanks to measure temperature and gas concentration, for example.

[0078]

[0052] These data are transmitted to the data management module (4), where the artificial neural network algorithm processes the variables collected by the sensor (1) and inserts them into the previously trained mathematical model. This allows generating an estimate of microbial growth and predicting the evolution of the microbial population over time.

[0079]

[0053] The data management module (4) can calculate, based on quantitative data, the ideal point to stop the fermentation process, ensuring that the product reaches the desired sensory characteristics, such as flavor, texture and microbiological safety.

[0080]

[0054] Additionally, the module generates automatic alerts in case of variations in fermentation conditions that could compromise product quality. For example, if the temperature deviates from the ideal range for the growth of fermenting microorganisms, the module will send automatic notifications, enabling quick adjustments and preventing production losses.

Claims

MODIFIED CLAIMS Received by the International Secretariat on March 27, 2026 (27.03.2026) 1. METHOD FOR MONITORING BIOLOGICAL PRODUCTS characterized by comprising the following steps: a) Filling in the user interface unit (6) of the data management module (4) the input data on the biological product (7) and type of microorganism to be monitored, including filling in input data on the intrinsic factors of the biological product (7); b) Positioning the synchronous electronic sensor (1) in the vicinity of or directly in contact with the biological product (7), the said sensor being configured to obtain continuous temporal data of at least one extrinsic and / or intrinsic factor; c) The synchronous electronic sensor (1) continuously collects temperature data and / or other extrinsic and intrinsic factors optionally monitored in the vicinity of the biological product (7);d) The data transmission device (11) continuously transmits the data collected by the synchronous electronic sensor (1) via wireless network to the gateway (2); e) The gateway (2) transmits, via internet (3) to the data management module (4), the data received from the data transmission device (11) of the synchronous electronic sensor (1); f) The data management module (4) executes the steps of the method for predicting the microbial load of the biological product (7); g) The data management module (4) sends in real time the results obtained in the method for predicting the microbial load to the data storage device (42) and user interface unit (6); h) Processing of the set of extrinsic data and intrinsic data by means of an artificial neural network previously trained based on kinetic curves of microbial load, for the biological product (1) and for the type of microorganism to be monitored;i) Generate a quantitative estimate as output from the artificial neural network; MODIFIED SHEET (ARTICLE 19) Continuous monitoring of microbial load over time from a prediction model based on previously trained artificial neural networks that considers the interrelation between intrinsic, extrinsic, and implicit data, in addition to product and microorganism information to be monitored; j) Iteratively update the microbial load estimate obtained from the training of neural networks based on the temporal evolution of extrinsic and / or intrinsic data measured by the sensors; k) Determine, from the estimate, microbial growth and / or inactivation regimes under variable non-stationary environmental conditions.

2. MONITORING METHOD, according to claim 1, characterized in that step a) comprises filling in input data on intrinsic factors and recording data on intrinsic and extrinsic factors of the biological product (7) to be monitored.

3. METHOD FOR MONITORING BIOLOGICAL PRODUCTS characterized by employing the use of at least one asynchronous sensor (1') equipped with a removable storage unit (1T) and a local physical server (5') comprising the following steps: a') Filling in the user interface unit (6) of the data management module (4) the input data on the biological product (7) and type of microorganism to be monitored, including filling in input data on the intrinsic factors of the biological product (7); b') Positioning the asynchronous electronic sensor (T) in the vicinity of or directly in contact with the biological product (7), the said sensor being configured to obtain continuous temporal data of at least one extrinsic and / or intrinsic factor; c') The asynchronous electronic sensor (T) continuously collects temperature data and / or other extrinsic and intrinsic factors optionally monitored in the vicinity of the biological product (7);d') The removable storage unit (1 T) of the asynchronous electronic sensor (1') continuously receives and stores the collected data; MODIFIED SHEET (ARTICLE 19) e') Remove the removable storage unit (11') from the asynchronous electronic sensor (T) containing the collected data; f') Connect the removable storage unit (11') to the local physical server (5') for data transmission to the data management module (4); g') Execute the steps of the microbial load prediction method on the biological product (7) through the data management module (4); h') The data management module (4) sends the results obtained in the method to generate microbial load prediction to the data storage device (42) and user interface unit (6); i') Processing of the set of extrinsic and intrinsic data by means of an artificial neural network previously trained based on microbial load kinetic curves, for the biological product (1) and for the type of microorganism to be monitored;j') Generate, as output from the artificial neural network, a continuous quantitative estimate of the microbial load over time from a previously trained artificial neural network-based prediction model that considers the interrelation between intrinsic, extrinsic, and implicit data, in addition to information on the product and the microorganism to be monitored; k') Iteratively update the microbial load estimate obtained from the training of the neural networks as a function of the temporal evolution of extrinsic and / or intrinsic data measured by the sensors; I') To determine, based on estimation, microbial growth and / or inactivation regimes under variable non-stationary environmental conditions.

4. MONITORING METHOD, according to claim 3, characterized in that step a') comprises filling in input data on intrinsic factors and recording data on intrinsic and extrinsic factors of the biological product (7) to be monitored.

5. BIOLOGICAL PRODUCT MONITORING SYSTEM that MODIFIED SHEET (ARTICLE 19) employs the method for monitoring biological products as defined in claim 1, characterized by - comprising at least one synchronous electronic sensor (1) equipped with a data transmission device (11) with data communication with a gateway (2) established via wireless networks; - being the gateway (2) connected via internet (3) with the data management module (4); - the data management module (4) comprises a previously trained artificial neural network and a data storage device (42); - the data management module (4) connected to the user interface unit (6) via wireless networks.

6. SYSTEM, according to claim 5, characterized by the synchronous electronic sensor (1) being designed to be positioned in the vicinity of the biological product (7) to be monitored.

7. SYSTEM, according to claim 5, characterized by the synchronous electronic sensor (1) being designed to be positioned directly in contact with the biological product (7) to be monitored.

8. SYSTEM, according to claim 5, characterized in that the data management module (4) operates in a virtual environment from a cloud server (5).

9. SYSTEM, according to claim 5, characterized in that the data management module (4) operates in a virtual environment from a local physical server (5').

10. SYSTEM, according to claim 5, characterized in that the neural network training involves the steps of data collection and pre-processing of microbial growth curves, supervised network training, testing and performance evaluation, being developed for each product or process and microorganism to be monitored.

11. BIOLOGICAL PRODUCT MONITORING SYSTEM employing the biological product monitoring method as defined. MODIFIED SHEET (ARTICLE 19) in claim 3, characterized by - comprising at least one asynchronous electronic sensor (T) equipped with a removable storage unit (11'); - the data management module (4) comprises a previously trained artificial neural network and a data storage device (42); - the data management module (4) connected to the user interface unit (6).

12. SYSTEM, according to claim 11, characterized by the asynchronous electronic sensor (T) being designed to be positioned directly in contact with the biological product (7) to be monitored.

13. SYSTEM, according to claim 11, characterized by the asynchronous electronic sensor (T) being designed to be positioned in the vicinity of the biological product (7) to be monitored.

14. SYSTEM, according to claim 11, characterized by being the data management module (4) operated in a virtual environment from a local physical server (5').

15. METHOD FOR PREDICTING MICROBIAL LOAD IN BIOLOGICAL PRODUCTS implemented in the data management module (4) and employing the biological product monitoring system as defined in claims 5 and 11, characterized by comprising the following steps: I. The data management module (4) receives the data generated by the synchronous electronic sensor (1) or asynchronous electronic sensor (T); II. The data management module (4) feeds the artificial neural network algorithm with data on microorganisms, on the product and on intrinsic and extrinsic properties entered by the user in the user interface unit (6); III. The data management module (4) feeds the artificial neural network algorithm with temperature data from the synchronous electronic sensor (1) or asynchronous electronic sensor (T); IV. Based on the data entered in steps II and III, the model generates the MODIFIED SHEET (ARTICLE 19) quantitative results of the microbiological load; V. The data management module (4) sends the generated results in real time to the data storage device (42); VI. The data management module (4) provides real-time quantitative predictions of the microbial load in the user interface unit (6).

16. METHOD FOR PREDICTING MICROBIAL LOAD, according to claim 15, characterized in that the data management module (4) in step VI generates automatic notification alerts to the user when the microbial load approaches or exceeds predefined limits. MODIFIED SHEET (ARTICLE 19) [0001] DECLARATION IN ACCORDANCE WITH ARTICLE 19(1) [0002]INTERNATIONAL DEPOSIT: PCTBR2025050487 [0003] INTERNATIONAL DEPOSIT DATE: 10 / 29 / 2025 [0004]In accordance with article 19.1, a new set of claims is presented where the scope of the present invention is better defined. Claims 1, 2, 3, 4 and 16 have been amended by transferring material from claims 3 and 6 and paragraphs [024, 031 and 052] of the Descriptive Report, claim 10 is new and finds support in paragraphs [024, 047 and 052] of the Descriptive Report. The modifications to the claims were made in order to make the differences between the invention and the prior art clearer. [0005]Milton Lucídio Leão Barcellos Eng. Luiz Alberto Rosenstengel OAB / RS 43707 - API / BR 838 CREA / RS 57.679 - API / BR 813 [0006] Lawyer and Industrial Property Agent Mechanical Engineer and Industrial Property Agent

Images