Intelligent bacterium detection equipment and method for milk production

Abstract

The invention relates to the technical field of dairy product quality detection, in particular to an intelligent bacterium detection device and method for milk production, the intelligent bacterium detection device comprises a detection box, a centrifugal module is fixedly mounted outside the detection box, an edge calculation unit is fixedly mounted on the outer wall of the detection box above the centrifugal module, and a random forest algorithm is built in the edge calculation unit; a storage cavity is formed in the upper half part in the centrifugal module, a detection cavity is formed below the storage cavity, and a microfluidic console is arranged at the bottom of the inner wall of the detection cavity. The defects in the prior art are overcome, traditional 24-48 hours of laboratory detection is shortened to be completed within 5 minutes through centrifugal layering, multi-sensor synchronous collection and edge calculation, the continuous detection requirement of a production line is met, a detection result is directly linked with pasteurization equipment, technological parameters are adjusted in real time, the risk of batch pollution is avoided, and the production efficiency is improved. Escherichia coli, staphylococcus aureus and other common pathogenic bacteria can be distinguished, and the deterioration risk can be early warned 12 hours in advance.

Description

Technical Field

[0001] The present invention relates to the technical field of dairy product quality detection, and specifically to an intelligent bacterial detection device and method for milk production. Background Art

[0002] Traditional bacterial detection methods in the dairy industry (such as plate counting method, PCR method) rely on laboratory operations, which are time-consuming (24 - 48 hours) and cannot meet the real-time monitoring requirements of the production line; single sensors (such as electrochemical sensors) are easily interfered by milk components (fat, protein), resulting in a high false positive rate (>30%); existing equipment lacks automatic pretreatment and self-cleaning functions, and is prone to detection result drift due to residue contamination.

[0003] To solve the problems existing in the above technologies, an intelligent bacterial detection device and method for milk production are proposed. Summary of the Invention

[0004] In view of the deficiencies of the prior art, the present invention provides an intelligent bacterial detection device and method for milk production, which overcomes the deficiencies of the prior art and solves the problems mentioned in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solution: An intelligent bacterial detection device for milk production, including a detection box, an external centrifugation module is fixedly installed on the detection box, an edge computing unit is fixedly installed on the outer wall of the detection box above the centrifugation module, and a random forest algorithm is built in the edge computing unit. An upper half part inside the centrifugation module forms a storage cavity, a detection cavity is formed below the storage cavity, a microfluidic platform is arranged at the bottom of the inner wall of the detection cavity, a multi-sensor array is arranged on the inner wall of the detection cavity, and a self-cleaning module is configured inside the detection cavity.

[0006] As a preferred technical solution of the present invention, an automatic sampling pump is fixedly installed on the top of the microfluidic platform, a sample storage tank is fixedly installed on the inner wall of the detection box on one side of the automatic sampling pump, the input end of the automatic sampling pump is connected to the sample storage tank through a pipeline, a sample delivery port is opened on the sample storage tank, and the sample delivery port passes through the detection box and extends to the outside.

[0007] As a preferred technical solution of the present invention, a spiral channel is configured on the top of the microfluidic platform, a filter membrane is arranged inside the spiral channel, and the pore diameter of the filter membrane is 0.45 μm. The bottom end of the spiral channel is connected to the output end of the automatic sampling pump through a pipeline, the top end of the spiral channel is connected to a recovery pipe, and a recovery tank is fixedly installed at one end of the recovery pipe away from the spiral channel.

[0008] As a preferred technical solution of the present invention, the multi-sensor array includes an impedance sensor, an optical sensor, and a temperature sensor;

[0009] The impedance sensor adopts an interdigital electrode structure with an electrode spacing of 20 μm and is surface-modified with a nano-gold-graphene composite coating. When the sample flows through the interdigital electrodes of the impedance sensor, an alternating electric field (frequency scanned from 1 kHz to 10 MHz) is applied to the electrodes, and the impedance change of the solution is measured in real time;

[0010] The optical sensor integrates an LED light source (wavelength 405 nm) and a photodiode. The LED light source irradiates the sample, and the photodiode receives the bacterial scattered light signal. The two achieve synchronous data acquisition through timestamp alignment;

[0011] On one side of the temperature sensor, a micro-heating sheet is fixedly installed on the upper surface of the microfluidic platform. The temperature sensor monitors the temperature of the microfluidic platform in real time. If the temperature fluctuation exceeds ±0.5 °C, through the micro-heating sheet on the top of the microfluidic platform, a constant temperature (25 ± 0.2 °C) is maintained to eliminate the influence of temperature on impedance measurement.

[0012] As a preferred technical solution of the present invention, the self-cleaning module includes an ultrasonic cleaning module and a hydrogen peroxide disinfection module;

[0013] The ultrasonic cleaning module includes an ultrasonic generator configured and installed in the middle of the microfluidic platform, a piezoelectric ceramic transducer configured and installed at the bottom of the microfluidic platform, and an aluminum alloy vibration rod provided on the outer wall of the microfluidic platform;

[0014] The hydrogen peroxide disinfection module includes a hydrogen peroxide storage tank, a disinfection tube installed at the output port of the hydrogen peroxide storage tank, a first micro solenoid valve configured and installed on the disinfection tube, a discharge tube fixedly installed at the bottom end of the spiral channel, a second micro solenoid valve configured and installed on the discharge tube, and a waste liquid collection tank fixedly installed at the tail of the discharge tube. A piezoelectric atomization sheet and a micro air pump are configured and installed inside the hydrogen peroxide storage tank. The piezoelectric atomization sheet atomizes liquid H2O2 into 1 - 5 μm particles, and the micro air pump provides an air flow to drive the atomized particles to diffuse along the hydrogen peroxide storage tank into the spiral channel.

[0015] As a preferred technical solution of the present invention, a nitrogen gas tank is fixedly installed on the outer wall of the detection box. The output end of the nitrogen gas tank is fixedly installed with an air drying tube, and one end of the air drying tube away from the nitrogen gas tank is connected to the spiral channel. A third micro solenoid valve is configured and installed on the air drying tube. The sterile nitrogen gas inside the nitrogen gas tank can enter the spiral channel through opening the third micro solenoid valve to blow the residual liquid in the channel. At the same time, after the spiral channel is disinfected with hydrogen peroxide droplets, a platinum catalyst coating is integrated at the outlet of the discharge tube, which can decompose hydrogen peroxide into water and oxygen to avoid residual corrosion of the channel.

[0016] A method for intelligent detection of bacteria used in milk production includes:

[0017] S1. The operator puts the milk sample into the centrifugation module, and the centrifugation module starts automatically (3000 rpm, 5 minutes) to separate the fat layer and the whey.

[0018] Inject the skimmed whey in the middle layer of the separated sample into the sample input port. The sample input port enters the spiral channel through the automatic sampling pump, and the flow rate is precisely controlled by the stepping motor of the pump (0.5 mL / min) to ensure that the sample flows evenly through the sensor detection area. A filter membrane can be set inside the spiral channel to remove residual particles and avoid blocking the sensor surface. At the same time, the extended path of the spiral channel (total length 5 cm) allows sufficient time for bacteria to contact the sensor.

[0019] S2. The self-cleaning module performs pre-cleaning before sample injection: ultrasonic vibration (40 kHz) to remove residues on the electrode surface, and hydrogen peroxide spray sterilization (concentration 3%).

[0020] S3. The impedance sensor and the optical sensor work alternately according to the preset timing sequence:

[0021] From 0 to 10 seconds: The impedance sensor scans the full frequency band and records the impedance amplitude / phase data.

[0022] From 10 to 15 seconds: The optical sensor emits pulsed light (sampling rate 100 Hz) to capture the scattered light intensity fluctuations.

[0023] The temperature sensor monitors throughout the process, and the data is used as a correction parameter for the impedance value.

[0024] S4. The original data eliminates high-frequency noise through the wavelet denoising algorithm.

[0025] The feature extraction engine calculates:

[0026] Impedance feature: Extract the rate of change of the phase angle (Δθ / Δt) at 1 MHz frequency, which reflects the bacterial metabolic activity.

[0027] Optical feature: Calculate the variance (σ 2 ) of the light intensity fluctuations to distinguish differences in bacterial density.

[0028] Temperature-impedance coupling coefficient: Dynamically correct the impedance baseline value according to the temperature change.

[0029] S5. Input the above features into the pre-trained random forest model.

[0030] The model output results include:

[0031] Total bacterial count (TBC): The regression prediction value (unit: CFU / mL).

[0032] Probability of pathogenic bacteria: Classified output (such as 85% probability of Escherichia coli, 12% probability of Staphylococcus aureus).

[0033] If the probability of detecting pathogenic bacteria > preset threshold (e.g., Salmonella > 70%), immediately trigger a red alert.

[0034] S6. Send control instructions to the pasteurization equipment through the Modbus protocol:

[0035] If TBC > 10 5 CFU / mL, adjust the sterilization temperature from 72°C to 75°C and extend the duration by 10 seconds;

[0036] Synchronize the test results to the MES system, generate a quality report and store it in the blockchain.

[0037] Compared with the prior art, the beneficial effects of the present invention are:

[0038] 1. Through centrifugal stratification + multi-sensor synchronous acquisition + edge computing, the traditional laboratory test that takes 24 - 48 hours is shortened to be completed within 5 minutes, meeting the continuous detection requirements of the production line. The test results directly link to the pasteurization equipment to immediately adjust the process parameters and avoid the risk of batch contamination.

[0039] 2. Through the joint analysis of impedance, optical, and temperature data, the misjudgment rate caused by fat / protein interference is reduced from > 30% of the traditional single sensor to < 5%. The closed-loop control of the temperature sensor and the micro heating sheet eliminates the influence of ambient temperature fluctuations on impedance measurement.

[0040] 3. It can distinguish common pathogenic bacteria such as Escherichia coli and Staphylococcus aureus, and give an early warning of the deterioration risk 12 hours in advance.

[0041] 4. The residual hydrogen peroxide is catalytically decomposed into water and oxygen, without chemical pollution. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a three-dimensional schematic diagram of the present invention; Figure 1 ;

[0043] Figure 2 is a three-dimensional schematic diagram of the present invention; Figure 2 ;

[0044] Figure 3 is a partial cross-sectional schematic diagram of the detection box of the present invention;

[0045] Figure 4 is a front cross-sectional schematic diagram of the present invention;

[0046] Figure 5 is a side cross-sectional schematic diagram of the present invention.

[0047] In the figure: 1. Detection box; 2. Centrifugation module; 3. Edge computing unit; 4. Microfluidic platform; 5. Automatic sampling pump; 6. Sample storage tank; 7. Sample input port; 8. Ultrasonic generator; 9. Piezoelectric ceramic transducer; 10. Spiral channel; 11. Recovery pipe; 12. Recovery tank; 13. Impedance sensor; 14. Optical sensor; 15. Hydrogen peroxide storage tank; 16. Disinfection pipe; 17. First micro solenoid valve; 18. Discharge pipe; 19. Second micro solenoid valve; 20. Waste liquid collection tank; 21. Micro heating sheet; 22. Temperature sensor; 23. Nitrogen tank; 24. Air drying pipe; 25. Third micro solenoid valve. Detailed implementation manners

[0048] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0049] Please refer to Figures 1-5 , a bacterial intelligent detection device for milk production, including a detection box 1, a centrifugation module 2 is fixedly installed outside the detection box 1, an edge computing unit 3 is fixedly installed on the outer wall of the detection box 1 above the centrifugation module 2, and the edge computing unit 3 incorporates a random forest algorithm. An upper half part inside the centrifugation module 2 forms a storage cavity, and a detection cavity is formed below the storage cavity. A microfluidic platform 4 is arranged at the bottom of the inner wall of the detection cavity, a multi-sensor array is arranged on the inner wall of the detection cavity, and a self-cleaning module is configured inside the detection cavity.

[0050] Specifically, an automatic sampling pump 5 is fixedly installed on the top of the microfluidic platform 4, a sample storage tank 6 is fixedly installed on the inner wall of the detection box 1 on one side of the automatic sampling pump 5, the input end of the automatic sampling pump 5 is connected to the sample storage tank 6 through a pipeline, a sample input port 7 is opened on the sample storage tank 6, and the sample input port 7 extends through the detection box 1 to the outside.

[0051] Specifically, a spiral channel 10 is configured on the top of the microfluidic platform 4, a filter membrane is arranged inside the spiral channel 10, and the pore diameter of the filter membrane is 0.45 μm. The bottom end of the spiral channel 10 is connected to the output end of the automatic sampling pump 5 through a pipeline, and the top end of the spiral channel 10 is connected to a recovery pipe 11. A recovery tank 12 is fixedly installed at one end of the recovery pipe 11 away from the spiral channel 10.

[0052] Specifically, the multi-sensor array includes an impedance sensor 13, an optical sensor 14, and a temperature sensor 22;

[0053] The impedance sensor 13 adopts an interdigital electrode structure with an electrode spacing of 20 μm. The surface is modified with a nano-gold-graphene composite coating. When the sample flows through the interdigital electrodes of the impedance sensor 13, an alternating electric field (frequency scanned from 1 kHz to 10 MHz) is applied to the electrodes, and the impedance change of the solution is measured in real time;

[0054] The optical sensor 14 integrates an LED light source wavelength (405 nm) and a photodiode. The LED light source irradiates the sample, and the photodiode receives the bacterial scattered light signal. The two achieve synchronous data acquisition through timestamp alignment;

[0055] On one side of the temperature sensor 22, a micro-heating sheet 21 is fixedly installed on the upper surface of the microfluidic platform 4. The temperature sensor 22 monitors the temperature of the microfluidic platform 4 in real time. If the temperature fluctuation exceeds ±0.5 °C, through the micro-heating sheet 21 on the top of the microfluidic platform 4, a constant temperature (25 ± 0.2 °C) is maintained to eliminate the influence of temperature on impedance measurement.

[0056] Specifically, the self-cleaning module consists of an ultrasonic cleaning module and a hydrogen peroxide disinfection module;

[0057] The ultrasonic cleaning module includes an ultrasonic generator 8 configured and installed in the middle of the microfluidic platform 4, a piezoelectric ceramic transducer 9 configured and installed at the bottom of the microfluidic platform 4, and an aluminum alloy vibrating rod provided on the outer wall of the microfluidic platform 4;

[0058] The hydrogen peroxide disinfection module includes a hydrogen peroxide storage tank 15, a disinfection pipe 16 installed at the output port of the hydrogen peroxide storage tank 15, a first micro solenoid valve 17 configured and installed on the disinfection pipe 16, a discharge pipe 18 fixedly installed at the bottom end of the spiral channel 10, a second micro solenoid valve 19 configured and installed on the discharge pipe 18, and a waste liquid collection tank 20 fixedly installed at the tail of the discharge pipe 18. Inside the hydrogen peroxide storage tank 15, a piezoelectric atomization sheet and a micro air pump are configured and installed. The piezoelectric atomization sheet atomizes liquid H2O2 into 1 - 5 μm particles, and the micro air pump provides an air flow to drive the atomized particles to diffuse along the hydrogen peroxide storage tank 15 into the spiral channel 10.

[0059] Specifically, a nitrogen gas tank 23 is fixedly installed on the outer wall of the detection box 1. The output end of the nitrogen gas tank 23 is fixedly installed with an air drying pipe 24. One end of the air drying pipe 24 away from the nitrogen gas tank 23 is connected to the spiral channel 10. A third micro solenoid valve 25 is configured and installed on the air drying pipe 24. The sterile nitrogen gas inside the nitrogen gas tank 23 can enter the spiral channel 10 through opening the third micro solenoid valve 25 to purge the residual liquid in the channel. At the same time, after the spiral channel 10 is disinfected by hydrogen peroxide droplets, a platinum catalyst coating is integrated at the outlet of the discharge pipe 18, which can decompose hydrogen peroxide into water and oxygen to avoid residual corrosion of the channel.

[0060] A method for intelligent detection of bacteria used in milk production includes:

[0061] S1. The operator puts the milk sample into the centrifugation module 2, and the centrifugation module 2 starts automatically (3000 rpm, 5 minutes) to separate the fat layer from the whey.

[0062] Inject the skimmed whey in the middle layer of the separated sample into the sample input port 7. The sample input port 7 enters the spiral channel 10 through the automatic sampling pump 5, and the flow rate is precisely controlled by the stepping motor of the pump (0.5 mL / min) to ensure that the sample flows evenly through the sensor detection area. A filter membrane can be set inside the spiral channel 10 to remove residual particles and avoid blocking the sensor surface. At the same time, the extended path (total length 5 cm) of the spiral channel 10 allows enough time for bacteria to contact the sensor.

[0063] S2. The self-cleaning module performs pre-cleaning before sample injection: ultrasonic vibration (40 kHz) to remove residues on the electrode surface, and hydrogen peroxide spray sterilization (concentration 3%).

[0064] S3. The impedance sensor 13 and the optical sensor 14 work alternately according to the preset time sequence:

[0065] From 0 to 10 seconds: The impedance sensor 13 performs a full-frequency band scan and records impedance amplitude / phase data.

[0066] From 10 to 15 seconds: The optical sensor 14 emits pulsed light (sampling rate 100 Hz) and captures the scattered light intensity fluctuations.

[0067] The temperature sensor 22 monitors throughout the process, and the data is used as a correction parameter for the impedance value.

[0068] S4. The original data eliminates high-frequency noise through the wavelet denoising algorithm.

[0069] The feature extraction engine calculates:

[0070] Impedance feature: Extract the rate of change of the phase angle (Δθ / Δt) at a frequency of 1 MHz, which reflects the bacterial metabolic activity.

[0071] Optical feature: Calculate the variance (σ 2 ) of the light intensity fluctuations to distinguish differences in bacterial density.

[0072] Temperature-impedance coupling coefficient: Dynamically correct the impedance baseline value according to the temperature change.

[0073] S5. Input the above features into the pre-trained random forest model.

[0074] The model output results include:

[0075] Total bacterial count (TBC): The regression prediction value (unit: CFU / mL);

[0076] Pathogenic bacteria probability: Classified output (e.g., E. coli probability 85%, Staphylococcus aureus probability 12%);

[0077] If the detected pathogenic bacteria probability > preset threshold (e.g., Salmonella > 70%), immediately trigger a red alarm.

[0078] S6. Send control instructions to the pasteurization equipment through the Modbus protocol:

[0079] If TBC > 10 5 CFU / mL, adjust the sterilization temperature from 72°C to 75°C and extend the duration by 10 seconds;

[0080] Synchronize the test results to the MES system, generate a quality report and store it in the blockchain.

[0081] Finally, it should be noted that in the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "vertical", "upper", "lower", "horizontal", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation to the present invention.

[0082] In the description of the present invention, it should also be noted that unless otherwise clearly specified and defined, the terms "set", "installed", "connected", "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0083] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A bacterial intelligent detection device for milk production, comprising a detection box (1), characterized in that: An external centrifugation module (2) is fixedly installed on the detection box (1). An edge computing unit (3) is fixedly installed on the outer wall of the detection box (1) above the centrifugation module (2), and the edge computing unit (3) incorporates a random forest algorithm. In the upper half of the centrifugation module (2), a storage cavity is formed, and a detection cavity is formed below the storage cavity. At the bottom of the inner wall of the detection cavity, a microfluidic platform (4) is provided. A multi-sensor array is arranged on the inner wall of the detection cavity, and a self-cleaning module is configured inside the detection cavity.

2. The bacterial intelligent detection device for milk production according to claim 1, characterized in that: An automatic sampling pump (5) is fixedly installed on the top of the microfluidic platform (4). A sample storage tank (6) is fixedly installed on the inner wall of the detection box (1) on one side of the automatic sampling pump (5). The input end of the automatic sampling pump (5) is connected to the sample storage tank (6) through a pipeline. A sample delivery port (7) is provided on the sample storage tank (6), and the sample delivery port (7) extends through the detection box (1) to the outside.

3. The bacterial intelligent detection device for milk production according to claim 1, wherein: A spiral channel (10) is configured on the top of the microfluidic platform (4). A filter membrane is arranged inside the spiral channel (10), and the pore diameter of the filter membrane is 0.45 μm. The bottom end of the spiral channel (10) is connected to the output end of the automatic sampling pump (5) through a pipeline. The top end of the spiral channel (10) is connected to a recovery pipe (11), and a recovery tank (12) is fixedly installed at one end of the recovery pipe (11) away from the spiral channel (10).

4. The bacterial intelligent detection device for milk production according to claim 1, characterized in that: The multi-sensor array includes an impedance sensor (13), an optical sensor (14), and a temperature sensor (22); The impedance sensor (13) adopts an interdigital electrode structure with an electrode spacing of 20 μm and is surface-modified with a nano-gold / graphene composite coating. When the sample flows through the interdigital electrodes of the impedance sensor (13), an alternating electric field (frequency scanned from 1 kHz to 10 MHz) is applied to the electrodes to measure the impedance change of the solution in real time; The optical sensor (14) integrates an LED light source (wavelength 405 nm) and a photodiode. The LED light source irradiates the sample, and the photodiode receives the bacterial scattered light signal. The two achieve synchronous data acquisition through timestamp alignment; A micro heating sheet (21) is fixedly installed on the upper surface of the microfluidic platform (4) on one side of the temperature sensor (22). The temperature sensor (22) monitors the temperature of the microfluidic platform (4) in real time. If the temperature fluctuation exceeds ±0.5 °C, the micro heating sheet (21) on the top of the microfluidic platform (4) is used to maintain a constant temperature (25 ± 0.2 °C) to eliminate the influence of temperature on impedance measurement.

5. The bacterial intelligent detection device for milk production according to claim 1, characterized in that: The self-cleaning module consists of an ultrasonic cleaning module and a hydrogen peroxide disinfection module; The ultrasonic cleaning module includes an ultrasonic generator (8) configured and installed in the middle of the microfluidic platform (4), a piezoelectric ceramic transducer (9) configured and installed at the bottom of the microfluidic platform (4), and an aluminum alloy vibrating rod arranged on the outer wall of the microfluidic platform (4); The hydrogen peroxide disinfection module includes a hydrogen peroxide storage tank (15), a disinfection pipe (16) installed at the output port of the hydrogen peroxide storage tank (15), a first micro solenoid valve (17) configured and installed on the disinfection pipe (16), a discharge pipe (18) fixedly installed at the bottom end of the spiral channel (10), a second micro solenoid valve (19) configured and installed on the discharge pipe (18), and a waste liquid collection tank (20) fixedly installed at the tail of the discharge pipe (18). A piezoelectric atomization sheet and a micro air pump are configured and installed inside the hydrogen peroxide storage tank (15). The piezoelectric atomization sheet atomizes liquid H2O2 into particles with a size of 1 - 5 μm, and the micro air pump provides an air flow to drive the atomized particles to diffuse along the hydrogen peroxide storage tank (15) into the spiral channel (10).

6. The intelligent bacterial detection device for milk production according to claim 1, wherein: A nitrogen gas tank (23) is fixedly installed on the outer wall of the detection box (1). The output end of the nitrogen gas tank (23) is fixedly installed with an air drying pipe (24). One end of the air drying pipe (24) away from the nitrogen gas tank (23) is connected to the spiral channel (10), and a third micro solenoid valve (25) is configured and installed on the air drying pipe (24).

7. A bacterial intelligent detection method for milk production according to claim 1, characterized in that Comprising: S1. The operator puts the milk sample into the centrifugation module (2), and the centrifugation module (2) automatically starts (3000 rpm, 5 minutes) to separate the fat layer and the whey; Inject the skimmed whey in the middle layer of the separated sample into the sample input port (7). The sample input port (7) enters the spiral channel (10) through the automatic sampling pump (5). The flow rate is precisely controlled by the stepping motor of the pump (0.5 mL / min) to ensure that the sample evenly flows through the sensor detection area. A filter membrane can be set inside the spiral channel (10) to remove residual particles and avoid clogging the surface of the sensor. At the same time, the extended path of the spiral channel (10) (total length 5 cm) allows sufficient time for bacteria to contact the sensor. S2. The self - cleaning module performs pre - cleaning before sample injection: Ultrasonic vibration (40 kHz) to remove residues on the electrode surface, and hydrogen peroxide spray sterilization (concentration 3%). S3. The impedance sensor (13) and the optical sensor (14) work alternately according to a preset time sequence: From 0 - 10 seconds: The impedance sensor (13) performs a full - band scan and records impedance amplitude / phase data; From 10 - 15 seconds: The optical sensor (14) emits pulsed light (sampling rate 100 Hz) and captures the fluctuations in the scattered light intensity; The temperature sensor (22) monitors the whole process, and the data is used as a correction parameter for the impedance value. S4. The original data eliminates high - frequency noise through the wavelet denoising algorithm; The feature extraction engine calculates: Impedance feature: Extract the rate of change of the phase angle (Δθ / Δt) at a frequency of 1 MHz, which reflects the bacterial metabolic activity; Optical feature: Calculate the variance (σ 2 ) of the light intensity fluctuation to distinguish the difference in bacterial density; Temperature - impedance coupling coefficient: Dynamically correct the impedance baseline value according to the temperature change. S5. Input the above features into a pre - trained random forest model. The model output results include: Total bacterial count (TBC): The regression prediction value (unit: CFU / mL); Probability of pathogenic bacteria: Classification output (such as the probability of Escherichia coli is 85%, and the probability of Staphylococcus aureus is 12%); If the detected probability of pathogenic bacteria > the preset threshold (such as Salmonella > 70%), immediately trigger a red alarm. S6. Send control instructions to the pasteurization equipment via the Modbus protocol: If TBC > 10 5 CFU / mL, adjust the sterilization temperature from 72 °C to 75 °C and extend the duration by 10 seconds; Synchronize the detection results to the MES system, generate a quality report and store it in the blockchain.

Images