Online visible near-infrared fermentation liquid food production process monitoring system and method

Abstract

The invention discloses an online visible near-infrared fermentation liquid food production process monitoring system and method, and the monitoring system comprises a bioreactor 8 which is used for generating to-be-detected fermentation liquor, the fermentation liquor is input into a detection bin through a liquid pipeline assembly, the fermentation liquor entering the detection bin is detected through a visible near-infrared detection assembly, and the detected fermentation liquor is sent to the detection bin; and the detected infrared spectrum is sent to the upper computer 7, the upper computer 7 detects the infrared spectrum, and the fermentation liquor after detection flows back to the bioreactor 8. The monitoring system provided by the invention can realize non-contact detection of the fermentation liquid, reduces the possibility of being polluted in the fermentation process, has low requirements on users, and can be used without professional knowledge of the monitoring system.

Description

Technical Field

[0001] The invention belongs to the technical field of food processing, and in particular relates to an online visible-near infrared fermented liquid food production process monitoring system and method. Background Art

[0002] Fermented liquid foods occupy a large proportion in people's daily lives and are widely popular due to their unique health benefits. The quality of fermented liquid foods directly affects their nutritional value and the health of consumers, so it is necessary to improve the quality of fermented liquid foods to meet people's demand for high-quality food.

[0003] During the fermentation production process, the internal components of fermented liquid foods change in a complex manner, which makes it difficult to optimize and control the fermentation process and improve product quality. The existing fermentation detection system has many deficiencies in real-time monitoring of key indicators of fermented liquid foods, mainly including: traditional detection methods usually rely on laboratory analysis, and cannot achieve real-time monitoring of the fermentation process, resulting in difficulty in timely grasping the reaction situation; the data acquisition and analysis of the existing system often have time delays and cannot quickly respond to changes in the fermentation process; changes in environmental factors such as temperature and pressure during the fermentation process may have a significant impact on the detection results, and the existing system is not adaptable enough to this. The premise of optimizing and controlling the fermentation process is to monitor the key indicators of the internal components during the fermentation process, so it is urgent to develop a monitoring system that can be applied to the fermentation process of fermented liquid foods to meet the high-quality production needs of fermented liquid foods.

[0004] In order to realize real-time online monitoring of key indicators in the production process of fermented liquid foods, and then optimize the control of the production process to improve product quality, an online visible-near infrared fermented liquid food production process monitoring system and method are proposed. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes an online visible-near-infrared fermented liquid food production process monitoring system and method, which can realize real-time online monitoring of key indicators in the production process of fermented liquid food, and then realize optimized control of the production process to improve product quality.

[0006] The present invention provides an online visible-near-infrared fermented liquid food production process monitoring system, comprising: a bioreactor, used to produce a fermentation liquid to be tested, the fermentation liquid is input into a detection chamber through a liquid pipeline component, the fermentation liquid entering the detection chamber is detected by a visible-near-infrared detection component, and the detected infrared spectrum is sent to a host computer, the host computer detects the infrared spectrum, and the fermentation liquid after detection flows back to the bioreactor.

[0007] Optionally, the bioreactor is a bioreactor with a stirring motor.

[0008] Optionally, the liquid pipeline assembly includes a latex tube, a filter, and a constant flow pump connected in sequence, the constant flow pump is connected to the liquid inlet of the detection chamber, the liquid outlet of the detection chamber is connected to one end of a reflux pipe, and the other end of the reflux pipe is inserted into the bioreactor.

[0009] Optionally, the constant current pump is also connected to a lower computer, and the lower computer is used to control the speed, direction, start and stop of the constant current pump, and the lower computer is connected to the upper computer.

[0010] Optionally, the visible near-infrared detection component includes a visible near-infrared light generating system, a first optical fiber, a second optical fiber and a spectrometer. The visible near-infrared light generating system is connected to the optical path entrance of the detection chamber through the optical fiber, and the optical path exit of the detection chamber is connected to the spectrometer through the second optical fiber. The spectrometer is connected to the host computer, and the visible near-infrared light generating system is also connected to the slave computer.

[0011] Optionally, a blocking component is provided inside the detection chamber along the light path to prevent the fermentation liquid from entering the light path.

[0012] The present invention also provides an online visible-near infrared fermented liquid food production process monitoring method, comprising:

[0013] Acquire the spectral data to be monitored;

[0014] The spectral data to be monitored is input into a calibration model to obtain a monitoring result; wherein the calibration model is calibrated using a multivariate calibration method, and the calibration model is trained using a training set, and the training set is spectral data collected using a visible near-infrared spectrometer.

[0015] Optionally, before the spectral data collected by the visible near infrared spectrometer, the following is also included:

[0016] According to the turbidity of the liquid to be tested, the integration time, number of scans and smoothing points of the initial signal of the visible near-infrared spectrometer are adjusted to ensure that the detection range of the visible near-infrared spectrometer can cover the entire fermentation process. If it cannot cover the entire fermentation process, the integration time is adjusted to ensure that the entire fermentation process is covered.

[0017] Optionally, the spectral data to be monitored is input into a calibration model, and obtaining the monitoring result includes: performing dimensionality reduction processing on the spectral data to be monitored, calibrating the spectral data after dimensionality reduction, and obtaining the monitoring result.

[0018] Compared with the prior art, the present invention has the following advantages and technical effects:

[0019] An online visible-near-infrared fermented liquid food production process monitoring system proposed in the present invention can realize automatic online monitoring of key indicator parameters in the fermentation process, provide key technical parameters for the fermentation process, and provide a technical basis for the high-quality production of fermented liquid foods; the monitoring system proposed in the present invention can realize contactless detection of fermentation liquid, reducing the possibility of contamination of the fermentation process, and has low requirements for users, and does not require professional knowledge of the monitoring system to use it.

[0020] The invention proposes an online visible-near-infrared fermented liquid food production process monitoring method, which is helpful for the use of the monitoring system; liquid circulation reflux is achieved through a liquid pipeline component; a visible-near-infrared detection component is used to realize online acquisition of the spectrum of the flowing liquid, and then a calibration model is established using a multivariate calibration method; the established calibration model is encapsulated as a file in a dynamic link library format, which is convenient for the host computer to call, and also ensures the encapsulation of the model file, and can be used for online monitoring of the fermentation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0022] Figure 1 This is a structural diagram of an online visible near-infrared fermentation liquid food production process monitoring system according to an embodiment of the present invention; wherein, 1, a stirring motor, 2, a latex tube, 3, a filter, 4, a constant current pump, 5, a signal line, 6, a lower computer, 7, a host computer, 8, a bioreactor, 9, a halogen tungsten lamp, 10, an optical fiber, 11, a four-way valve, 12, a spectrometer, 13, a data line;

[0023] Figure 2 is a modeling flow chart of the monitoring method of an embodiment of the present invention, wherein: Figure 2 (a) is the original spectrum, Figure 2 (b) is the original spectrum preprocessing result. Figure 2 (c) PLS model diagram for using the selected feature variable subset. Figure 2 (d) Diagram of the PLS model built using a subset of high-frequency variables. Figure 2 (e) is the diagram for determining the best model for each indicator. DETAILED DESCRIPTION

[0024] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0025] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0026] The present invention provides an online visible near infrared fermentation liquid food production process monitoring system. Figure 1 As shown, it specifically includes: a bioreactor 8, which produces a fermentation liquid to be tested, and the fermentation liquid is input into a detection chamber through a liquid pipeline component. The fermentation liquid entering the detection chamber is detected by a visible near-infrared detection component, and the detected infrared spectrum is sent to a host computer 7. The host computer 7 detects the infrared spectrum, and the fermentation liquid after the detection flows back to the bioreactor 8.

[0027] Specifically, the liquid pipeline assembly includes a latex tube 2, a filter 3, a constant flow pump 4 and a four-way valve 11;

[0028] The visible near infrared detection component includes a halogen tungsten lamp 9, an optical fiber 10, a four-way valve 11, a spectrometer 12 and a data line 13;

[0029] The lower computer 6 is a single chip microcomputer in this embodiment, and the upper computer 7 is a notebook computer in this embodiment;

[0030] The latex tube 2 has 4 sections, which are connected to the bioreactor 8, the filter 3, and the four-way valve 11 through a pagoda-type joint; the latex tube 2 is embedded in the constant flow pump head to achieve liquid circulation; the specification of the latex tube 2 can be 25#, but there is no limit.

[0031] The filter 3 is cylindrical, and the inner cavity is provided with a filter screen; the filter screen is 100 mesh and is made of 304 stainless steel.

[0032] Furthermore, the bioreactor 8 is a bioreactor with a stirring motor 1 .

[0033] Specifically, the bioreactor 8 is a 5L glass small fermentation tank, which can be stirred by the stirring motor 1 and also has a constant temperature function.

[0034] Furthermore, the liquid pipeline assembly includes a latex tube 2, a filter 3, and a constant flow pump 4 connected in sequence, the constant flow pump 4 is connected to the liquid inlet of the detection chamber, the liquid outlet of the detection chamber is connected to one end of a reflux pipe, and the other end of the reflux pipe is inserted into the bioreactor 8;

[0035] The constant flow pump 4 is also connected to a lower computer 6 , which is used to control the speed, direction, start and stop of the constant flow pump 4 . The lower computer 6 is connected to an upper computer 7 .

[0036] Specifically, the constant flow pump 4 has two modes: manual and automatic. A standard DB15 external control interface is provided on the outside, which can be connected to the controller. In the automatic mode, the speed, direction, and start and stop can be automatically set.

[0037] In this embodiment, the detection chamber adopts a four-way valve 11 device, which is made of brass, resistant to acid and alkali corrosion, and has a liquid channel (at both ends of A and B) and an optical channel (at both ends of C and D); the liquid channel and the optical channel cross each other in space, with latex tubes connected to both ends of the liquid path A and B, and optical fibers connected to both ends of the optical path C and D; at the intersection of the liquid path and the optical path, a glass transparent lens is installed at the port inside the optical path that is connected to the liquid path, which prevents liquid from entering the optical path and does not affect the transmission of light through the liquid, thereby facilitating the collection of the transmission spectrum of the liquid flowing through.

[0038] Furthermore, the visible near-infrared detection component includes a visible near-infrared light generating system, a first optical fiber 10, a second optical fiber and a spectrometer 12. The visible near-infrared light generating system is connected to the optical path entrance of the detection chamber through the optical fiber 10, and the optical path exit of the detection chamber is connected to the spectrometer 12 through the second optical fiber. The spectrometer 12 is connected to the upper computer 7, and the visible near-infrared light generating system is also connected to the lower computer 6.

[0039] Specifically, in this embodiment, the visible near-infrared light generating system adopts a halogen tungsten lamp 9 .

[0040] Furthermore, along the light path direction, a blocking component is provided inside the detection chamber to prevent the fermentation liquid from entering the light path.

[0041] Specifically, the monitoring system is composed of a bioreactor 8, a liquid pipeline assembly, a visible near infrared detection assembly, a lower computer control unit and an upper computer control unit;

[0042] The bioreactor is a 5L glass small fermentation tank, which can be stirred by a stirring motor 1 and also has a constant temperature function;

[0043] The liquid pipeline assembly includes a latex tube 2, a filter 3, a constant flow pump 4 and a four-way valve 11;

[0044] The two ends corresponding to the upper and lower parts of the four-way valve 4 are marked as ends A and B; the two ends corresponding to the left and right parts are marked as ends C and D;

[0045] The latex tube 2 is connected to the bioreactor 8, the filter 3, the constant flow pump 4 and the four-way valve 11 in sequence to form a closed-loop liquid circuit, wherein the latex tube 2 is connected to the A and B ends of the four-way valve 11;

[0046] Furthermore, the latex tube has 4 sections in total, which are connected to the bioreactor, the filter and the four-way valve through pagoda-type connectors.

[0047] Furthermore, the latex tube is embedded in the pump head of the constant flow pump to achieve liquid circulation.

[0048] Furthermore, the specification of the latex tube can be 25#, but is not limited.

[0049] Furthermore, the filter is cylindrical, and the inner cavity is provided with a filter screen; the filter screen is 100 mesh and is made of 304 stainless steel.

[0050] Furthermore, the constant flow pump has two modes: manual and automatic. A standard DB15 external control interface is provided on the outside for connection with the controller. In the automatic mode, the speed, direction and start and stop can be automatically set.

[0051] Furthermore, the four-way valve is made of brass, resistant to acid and alkali corrosion, and has a liquid channel (at both ends of A and B) and an optical channel (at both ends of C and D); the liquid channel and the optical channel cross each other in space, the liquid channels A and B are connected to latex tubes, and the optical channels C and D are connected to optical fibers; at the intersection of the liquid and optical paths, a glass light-transmitting lens is installed at the port inside the optical path that communicates with the liquid path, which prevents liquid from entering the optical path and does not affect the light passing through the liquid, so as to facilitate the collection of the transmission spectrum of the liquid flowing through;

[0052] The visible near infrared detection component includes a halogen tungsten lamp 9, an optical fiber 10, a four-way valve 11, a spectrometer 12 and a data line 13;

[0053] The optical fiber 10 is connected to the halogen tungsten lamp 9, the C and D ends of the four-way valve 11 and the spectrometer 12 in sequence, and the light source 9 is electrically connected to the lower computer 6;

[0054] The lower computer 6 and the spectrometer 12 are electrically connected to the upper computer 7) via a data line 13.

[0055] Furthermore, the halogen tungsten lamp can emit light in the visible and near-infrared bands to meet the use of the selected visible and near-infrared instruments, with a rated voltage of 3.3V and a rated current of 0.43A.

[0056] Furthermore, the optical fiber has two sections, which are single-core quartz optical fibers and can be bent appropriately. The interface has an optical fiber coupler for convenient connection with a halogen tungsten lamp, a four-way valve and a spectrometer, and complies with the SMA905 standard.

[0057] Furthermore, the spectrometer is a miniature visible near-infrared spectrometer with an operating wavelength range of 340 to 1100 nm, a total of 2048 variable points, and supports multiple communication methods such as USB and serial port.

[0058] Furthermore, the data cables are respectively a USB to serial port data cable and a Micro USB data cable, which can respectively realize the communication between the host computer and the slave computer, and the host computer and the visible near-infrared spectrometer.

[0059] Furthermore, the lower computer control unit is an STM32 single-chip microcomputer of the ARM system, which can control the pump through the serial port, control the halogen tungsten lamp through pin operation with the help of relays, and communicate with the upper computer through the serial port.

[0060] Furthermore, the host control unit is a computer, which realizes functions such as spectrum acquisition, spectrum display, online monitoring and result display in combination with a control program.

[0061] Furthermore, the host computer software is a form application written on the Visual Studio 2019 platform, which has a visual human-computer interaction interface, can realize independent setting of parameters, and has serial port and USB communication functions; the host computer software can realize the import of model files to monitor key indicators of the production process; the model files are saved in the format of dynamic link libraries; the host computer software can realize automatic periodic detection by setting the number of detections and detection time intervals;

[0062] Furthermore, the lower computer software is an application program written based on the Keil uVision5 platform, which controls the execution components by receiving instructions sent by the upper computer.

[0063] The present invention also provides an online visible-near infrared fermented liquid food production process monitoring method, comprising:

[0064] Acquire the spectral data to be monitored;

[0065] The spectral data to be monitored is input into the calibration model to obtain the monitoring results; wherein, the calibration model is calibrated using a multivariate calibration method, and the calibration model is trained using a training set, which is spectral data collected using a visible near-infrared spectrometer.

[0066] Furthermore, before the spectral data collected by the visible near infrared spectrometer, the method further includes:

[0067] According to the turbidity of the liquid to be tested, adjust the integration time, number of scans and smoothing points of the initial signal of the visible near-infrared spectrometer to ensure that the detection range of the visible near-infrared spectrometer can cover the entire fermentation process. If it cannot cover the entire fermentation process, adjust the integration time to ensure that the entire fermentation process is covered.

[0068] Furthermore, the spectral data to be monitored is input into the calibration model, and obtaining the monitoring result includes: performing dimensionality reduction processing on the spectral data to be monitored, calibrating the spectral data after dimensionality reduction, and obtaining the monitoring result.

[0069] Specifically, the monitoring methods include:

[0070] Step 1, conducting a fermentation experiment; wherein the fermentation experiment is liquid fermentation;

[0071] Step 2, adjusting the detection parameters of the spectrometer; wherein, adjusting the detection parameters of the spectrometer includes: flexibly adjusting the integration time, the number of scans and the number of smoothing points of the initial signal of the spectrometer according to the turbidity of the liquid to be tested, to ensure that the detection range of the spectrometer can cover the entire fermentation process. If it cannot cover the entire fermentation process, increase the integration time in the software according to the needs to achieve two-stage detection or multi-stage detection, so that an independent calibration model needs to be established for each stage, and adjacent detection stages have partial overlap to ensure that the entire fermentation process is covered;

[0072] Step 3, spectrum acquisition; wherein, spectrum acquisition: after setting the spectrometer parameters, the spectrum data of the sample can be saved in the host computer program;

[0073] Step 4, physical and chemical value determination; wherein, the physical and chemical value determination can be measured according to traditional standard methods (for example, pH value determination, total acidity determination, and sugar content determination, etc.);

[0074] Step 5, spectral preprocessing: by preprocessing the spectrum, the data quality and feature extraction of the original spectrum are improved, the effect of spectral analysis is enhanced, and the accuracy and reliability of modeling are improved. Among them, the spectral preprocessing methods include regularization, centering, multivariate scattering correction, standard normal transformation, wavelet transformation, first-order derivative and second-order derivative, etc.;

[0075] Step 6, spectral data dimension reduction processing; wherein, spectral data dimension reduction processing includes feature compression and feature variable selection; the feature compression method is generally principal component analysis; the feature variable selection method is intelligent optimization algorithm;

[0076] Step 7, calibration model establishment and determination; wherein, calibration model establishment and determination, using the new variable subset after dimensionality reduction processing to establish a mathematical model through linear or nonlinear multivariate calibration method, and determining the best model by comparing model performance; multivariate calibration methods include partial least squares (PLS), multiple linear regression (MLR), principal component regression (PCR), artificial neural network (ANN), support vector machine (SVM) and other methods; model performance evaluation indicators include correlation coefficient Rc and root mean square error RMSEP of the prediction set, etc.;

[0077] Step 8, calibration model verification; wherein, calibration model verification, a new batch of fermentation experiments are designed to verify the prediction performance of the model;

[0078] Step 9, importing the calibration model; wherein, when importing the calibration model, the model file can be named in a unified format and placed in a designated folder for software to call; the calibration model is encapsulated as a file in a dynamic link library format;

[0079] Step 10, setting automatic monitoring parameters; wherein, setting automatic monitoring parameters mainly refers to setting spectrum acquisition parameters and detection mode parameters in the host computer program.

[0080] Steps 5 to 8 were all completed in Matlab 2014b;

[0081] like Figure 1 As shown, this embodiment provides a fermented liquid food monitoring system based on online visible near-infrared spectroscopy technology, which is used for online monitoring of key growth indicators such as bacterial concentration, soluble sugar and total acid in the fermentation process of kombucha. The system includes a bioreactor 8 with a stirring motor 1, a latex tube 2, a filter 3, an adjustable constant flow pump 4, a control signal line 5, a single-chip microcomputer 6, a laptop computer 7, a halogen tungsten lamp 9, an optical fiber 10, a four-way valve 11, a visible near-infrared spectrometer 12 and a data transmission line 13;

[0082] The bioreactor 8 is equipped with a constant temperature system and a pH meter to facilitate the fermentation experiment; the bioreactor 8 is connected to the filter 3, the constant flow pump 4 and the four-way valve 11 through a latex tube to achieve the circulation of the fermentation liquid; Figure 1 As shown, the circulation reflux direction is clockwise;

[0083] The latex tube 2 meets the use specifications of the constant flow pump 4, and a filter 3 is connected between the bioreactor 8 and the adjustable constant flow pump 4 to filter out suspended particles and flocculent substances, so as to ensure the collection of a stable spectrum;

[0084] The adjustable constant current pump 4 needs a separate external power supply, and the latex tube can be pressed into the pump head to provide circulation power for the fermentation liquid, and the pump liquid is separated; the constant current pump 4 communicates with the lower computer 6 through the RS485 serial port, and receives the instructions sent by the single chip microcomputer, including speed, direction and start and stop;

[0085] The control signal line 5 is a general single-strand copper signal line;

[0086] The lower computer 6 is an STM32 development board, which controls the lighting and extinguishing of the halogen tungsten lamp 9 through a relay; the single-chip computer 6 realizes communication with the upper computer laptop computer 7 through a serial port;

[0087] Laptop computer 7 is a general-purpose laptop computer. Based on this hardware, a form application program is compiled that integrates the functions of communication setting, spectrum measurement, spectrum spectrum display, automatic periodic detection, online monitoring of key growth indicators, and online monitoring result display;

[0088] The halogen tungsten lamp 9 needs to be powered by an external power supply, the required rated voltage is 3.3V, and the rated current is 0.43A; and a rectangular lampshade is designed for the halogen tungsten lamp 9; the lampshade is connected to the optical fiber through an optical fiber coupler and complies with the SMA905 standard;

[0089] Two optical fibers 10 are required in total, which can pass light in the visible near-infrared band to meet the use of the spectrometer; through the optical fiber coupler, one end of the optical fiber 1 is connected to the halogen tungsten lamp 9 and the other end is connected to the C end of the four-way valve 11, and one end of the optical fiber 2 is connected to the D end of the four-way valve 11 and the other end is connected to the visible near-infrared spectrometer 12; all of the above meet the SMA905 standard;

[0090] The liquid path of the four-way valve 11 crosses with the optical path, the two ends of the liquid path are connected to the latex tube 3, and the two ends of the optical path are connected to the optical fiber 10;

[0091] The visible near-infrared spectrometer 12 is a miniature visible near-infrared spectrometer with an operating wavelength range of 340 to 1100 nm, a total of 2048 variable points, and supports multiple communication methods such as USB and serial port;

[0092] The data cable is a Micro USB data cable.

[0093] This embodiment provides a fermented liquid food monitoring method based on online visible near-infrared spectroscopy technology, which is used to online monitor key growth indicators such as bacterial concentration, soluble sugar and total acid in the fermentation process of kombucha. Figure 2 As shown, the main steps are as follows:

[0094] Step 1: Set up the monitoring system and carry out a batch of fermentation experiments;

[0095] Step 2: During the fermentation process, according to the turbidity of the liquid to be tested, flexibly adjust the integration time, scanning times and smoothing points of the initial signal of the spectrometer to ensure that the detection range of the spectrometer can cover the entire fermentation process;

[0096] Step 3: After setting the spectrometer parameters, the spectral data of the sample can be saved in the host computer program;

[0097] Step 4: Start multiple batches of fermentation experiments and collect samples; collect visible and near-infrared spectra during the fermentation process. The original spectra are as follows: Figure 2 (a) and the standard values ​​of bacterial concentration, soluble sugar and total acid were determined by national standard method;

[0098] Step 5: Use the first-order derivative to preprocess the original spectrum to eliminate the baseline drift during the detection process. The processing result is as follows: Figure 2 (b)

[0099] Step 6: Establish a mathematical model. The modeling process is as follows:

[0100] The samples were divided into a calibration set and a prediction set in a ratio of 7:3. Table 1 shows the results of preprocessing. The model performance of the total acid index decreased after the first-order derivative, so the original spectrum was selected for modeling.

[0101] Table 1

[0102]

[0103] The Sparrow Search Algorithm (SSA), Whale Optimization Algorithm (WOA) and African Vulture Optimization Algorithm (AVOA) were used in combination with PLS to extract the characteristic variables of the original spectrum, and the PLS model was established using the selected characteristic variable subset, such as Figure 2 (c) As shown in Table 2, the three indicators were optimized using SSA, WOA and AVOA methods respectively. The original number of variables was 2048, and the model structure was well optimized.

[0104] Table 2

[0105]

[0106]

[0107] Each optimization method was run 20 times and compared with the model established using the full variable spectrum. A variable subset with better results was selected, and a high-frequency variable subset was screened out based on these variable subsets. The high-frequency variable subset was then used to establish the PLS model, such as Figure 2 (d) As shown; after comparison, the best model for each indicator is finally determined, as shown in Figure 2 (e) As shown in Table 3, the modeling results of the selected high-frequency variables are compared with the variable results selected by the optimization method. The model performance has been improved and the number of variables has been reduced;

[0108] Table 3

[0109]

[0110] Step 7: Model import; put the model file into the specified folder, and the model file is uniformly named "Model";

[0111] Step 8: Set detection parameters, including spectrum acquisition parameters and detection mode parameters;

[0112] Step 9: Example verification; start a new batch of fermentation experiments, use the developed online visible near-infrared monitoring system to obtain fermentation process parameters, and synchronously collect samples at the detection time point to determine the standard physical and chemical values.

[0113] Step 10: Result evaluation; The maximum errors between the predicted values ​​and the actual measured values ​​of the three indicators of bacterial concentration, soluble sugar and total acid were 0.0566 (Abs), 1.0028 g / L and 0.3567 g / L, respectively. The differences are not large and can meet the needs of online monitoring.

[0114] The present invention has a wide application range and good application effect, and can be widely used in the production process of fermented liquid foods to monitor key indicators in the production process online and provide technical parameters for optimized control of production.

[0115] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An online visible-near infrared fermented liquid food production process monitoring system, characterized in that: include: The bioreactor (8) is used to produce a fermentation liquid to be tested. The fermentation liquid is input into the detection chamber through a liquid pipeline component. The fermentation liquid entering the detection chamber is detected by a visible near-infrared detection component, and the detected infrared spectrum is sent to a host computer (7). The host computer (7) detects the infrared spectrum, and the fermentation liquid after the detection is returned to the bioreactor (8).

2. The online visible-near infrared fermented liquid food production process monitoring system according to claim 1 is characterized in that: The bioreactor (8) is a bioreactor with a stirring motor (1).

3. The online visible-near infrared fermented liquid food production process monitoring system according to claim 1 is characterized in that: The liquid pipeline assembly includes a latex tube (2), a filter (3), and a constant flow pump (4) which are connected in sequence. The constant flow pump (4) is connected to the liquid inlet of the detection chamber. The liquid outlet of the detection chamber is connected to one end of a reflux pipe, and the other end of the reflux pipe is inserted into the bioreactor.

4. The online visible-near infrared fermented liquid food production process monitoring system according to claim 3 is characterized in that: The constant flow pump (4) is also connected to a lower computer (6), and the lower computer (6) is used to control the speed, direction, start and stop of the constant flow pump (4). The lower computer (6) is connected to the upper computer (7).

5. The online visible-near infrared fermented liquid food production process monitoring system according to claim 1, characterized in that: The visible near-infrared detection component comprises a visible near-infrared light generating system, a first optical fiber (10), a second optical fiber and a spectrometer (12); the visible near-infrared light generating system is connected to the optical path entrance of the detection chamber via the optical fiber (10); the optical path exit of the detection chamber is connected to the spectrometer (12) via the second optical fiber; the spectrometer (12) is connected to the host computer (7); and the visible near-infrared light generating system is also connected to the slave computer (6).

6. The online visible-near infrared fermented liquid food production process monitoring system according to claim 1, characterized in that: Along the light path direction, a blocking component is arranged inside the detection chamber to prevent the fermentation liquid from entering the light path.

7. An online visible-near infrared fermented liquid food production process monitoring method, characterized in that: include: Acquire the spectral data to be monitored; The spectral data to be monitored is input into a calibration model to obtain a monitoring result; wherein the calibration model is calibrated using a multivariate calibration method, and the calibration model is trained using a training set, and the training set is spectral data collected using a visible near-infrared spectrometer.

8. The method for monitoring the production process of an online visible-near infrared fermented liquid food according to claim 7, characterized in that: Before the spectral data collected by the visible near infrared spectrometer also includes: According to the turbidity of the liquid to be tested, the integration time, number of scans and smoothing points of the initial signal of the visible near-infrared spectrometer are adjusted to ensure that the detection range of the visible near-infrared spectrometer can cover the entire fermentation process. If it cannot cover the entire fermentation process, the integration time is adjusted to ensure that the entire fermentation process is covered.

9. The method for monitoring the production process of an online visible-near infrared fermented liquid food according to claim 7, characterized in that: Inputting the spectral data to be monitored into the calibration model to obtain the monitoring results includes: The spectral data to be monitored are subjected to feature compression and feature variable selection processing, and the processed spectral data are calibrated to obtain the monitoring result.

Images