Detection instrument for fully automatically detecting concentration of glucose in fermentation liquor

Abstract

The invention discloses a full-automatic detection instrument for detecting the concentration of glucose in fermentation broth, and relates to the field of glucose concentration detection.The full-automatic detection instrument comprises a fermentation tank set, a filtering sampler is arranged on one side of the fermentation tank set, a feeding pump is arranged on one side of the filtering sampler, a sample pool is arranged on one side of the feeding pump, and a mechanical arm sampler is arranged on one side of the sample pool; a standard sample pool and a micro-injection pump are respectively arranged on two sides of the mechanical arm sampler; a glucose electrochemical detection pool is arranged on one side of the micro-injection pump, a cleaning pool is connected to one side of the glucose electrochemical detection pool through a cleaning pump, a collection pool is connected to the bottom end of the glucose electrochemical detection pool through a waste liquid pump, and a concentration detection circuit is arranged in the glucose electrochemical detection pool; and one end of the glucose electrochemical detection pool is connected with a master controller. The full-automatic on-line detection instrument provided by the invention solves the problems of high glucose content and concentration in fermentation liquor, high difficulty in on-line real-time detection in the fermentation process, need of manual guarding and the like.

Description

Technical Field

[0001] The invention relates to glucose concentration detection, in particular to a fully automatic detection instrument for glucose concentration in fermentation liquid. Background Art

[0002] Biomanufacturing is an important foundation and guarantee for the development of bioeconomy today, and the bio-fermentation industry occupies a dominant position in biomanufacturing. However, in the current existing technologies, the bio-fermentation industry is still dominated by "extensive" production, and the understanding of the fermentation process is not sufficient. The fermentation process is still in a "black box" state. In order to improve the understanding of the fermentation process, the glucose concentration in the fermentation liquid can be detected to reflect the growth of cells and microorganisms, fermentation efficiency and product quality in the current fermentation process.

[0003] At present, the main instruments for detecting glucose concentration in biological laboratories include: liquid chromatography instruments, near-infrared spectrometers and biosensor analysis instruments. Liquid chromatography instruments are the most accurate in measurement, but their high cost and complex operating logic make them difficult to popularize. Infrared spectrometers can realize contactless online monitoring, but they are not sensitive enough for low-concentration glucose solutions and are expensive. However, the detection range of biosensor analysis instruments in the prior art is narrow, and the concentration in the fermentation broth is often several times higher than the detection range. Therefore, artificial dilution is required before measurement before detection, thereby introducing errors caused by artificial dilution, resulting in inaccurate final test results. In addition, the above instruments have a common disadvantage: they require manual supervision. If you want to obtain the status of the fermentation broth in real time, you must manually take samples and process them for detection at night, which greatly consumes human resources.

[0004] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0005] In view of the problems in the related art, the present invention proposes a fully automatic instrument for detecting the glucose concentration in the fermentation liquid to overcome the above technical problems existing in the existing related art.

[0006] To this end, the specific technical solution adopted by the present invention is as follows:

[0007] A fully automatic instrument for detecting glucose concentration in fermentation liquid comprises a fermentation tank group, a filter sampler is arranged on one side of the fermentation tank group, a feed pump is arranged on one side of the filter sampler, a sample pool is arranged on one side of the feed pump, a mechanical arm sampler is arranged on one side of the sample pool, and standard sample pools and micro-injection pumps are arranged on both sides of the mechanical arm sampler; a glucose electrochemical detection cell is arranged on one side of the micro-injection pump, a cleaning cell is connected to one side of the glucose electrochemical detection cell via a cleaning pump, a collection cell is connected to the bottom end of the glucose electrochemical detection cell via a waste liquid pump, a concentration detection circuit is arranged inside the glucose electrochemical detection cell, and a general controller is connected to one end of the glucose electrochemical detection cell.

[0008] Preferably, the master controller is used to combine the preset algorithm with the historical concentration detection data to generate a concentration calibration curve and a driving instruction, use the concentration calibration curve and the driving instruction to perform the fermentation liquid glucose concentration detection work, and verify the accuracy of the concentration detection according to the detection results;

[0009] The main controller includes a detection concentration curve calibration unit, a drive instruction setting generation unit, a concentration detection drive execution unit and a concentration result detection verification unit;

[0010] A detection concentration curve calibration unit is used to analyze the relationship between concentration, fermentation liquid volume and concentration based on historical concentration detection data and curve relationship calibration algorithm, and establish a detection concentration calibration curve according to the relationship results;

[0011] A drive instruction setting generation unit, used to determine the detection time and sampling point based on the historical concentration detection data and the time segmentation sequence algorithm, and generate a detection drive instruction according to the detection time and the sampling point;

[0012] The concentration detection drive execution unit is used to start the robot arm sampler and the glucose electrochemical detection cell according to the drive instruction to perform the fermentation liquid glucose concentration detection work, and obtain the fermentation liquid glucose concentration in combination with the concentration calibration curve and the concentration detection circuit;

[0013] The concentration result detection and verification unit is used to analyze the concentration error between the glucose concentration of the fermentation liquid and the concentration prediction result. Based on the error result, a verification model is built to analyze the accuracy of the detection concentration calibration curve and the driving instructions.

[0014] Preferably, the detection concentration curve calibration unit comprises:

[0015] A current information corresponding acquisition module, used to obtain the corresponding current value in the glucose electrochemical detection cell when there is no glucose solution and when a preset concentration of glucose solution is injected based on historical concentration detection data;

[0016] A relationship curve analysis module is used to describe the intensity change between the current value and the volume of the glucose solution based on the dilution effect relationship, and analyze the relationship curve between the concentration conductivity logarithm and the current according to the intensity change and the fitting algorithm;

[0017] An objective function establishment module is used to derive a concentration relationship between the logarithm of the concentration conductivity and the concentration of the glucose solution according to the relationship curve, and use the concentration relationship as a target function for calibrating the detection concentration curve;

[0018] The concentration curve calibration module is used to find the optimal solution of the detection concentration curve calibration objective function using the curve relationship calibration algorithm, and establish the glucose concentration detection calibration curve based on the optimal solution.

[0019] Preferably, using a curve relationship calibration algorithm to find an optimal solution for the detection concentration curve calibration objective function, and establishing a glucose concentration detection calibration curve based on the optimal solution includes:

[0020] After converting the detection concentration curve calibration objective function into a relative error least square function based on the curve relationship calibration algorithm, the initial parameters are set to perform Taylor series expansion operations to calculate the Jacobian matrix;

[0021] The relative error is analyzed based on the Jacobian matrix, and the iterative increment of the calibration objective function of the expanded detection concentration curve is obtained according to the relative error;

[0022] The relative error sum of squares is obtained based on the iterative increment and error analysis algorithm, and the accuracy of the iterative increment is judged by the relative error sum of squares. If the accuracy result does not meet the requirements, the iterative increment is obtained again;

[0023] If the accuracy result meets the requirements, the iterative increment is substituted into the relative error least squares function as the optimal solution of the objective function, and a glucose concentration detection calibration curve is established based on it.

[0024] Preferably, the expression of the error analysis algorithm is:

[0025]

[0026] In the formula, S represents the relative error sum of squares, v represents the total number of iterations, and r u represents the residual error of the u-th iteration, f(y u ,p) represents the relative error least square function corresponding to the calibration objective function of the detection concentration curve at the uth iteration, and p represents the iteration increment.

[0027] Preferably, the drive instruction setting generation unit includes:

[0028] A time series data acquisition module is used to acquire the corresponding liquid volume, number of liquid fermentation tanks, sampling time interval and sampling time required in the glucose concentration detection process as time series data based on historical concentration detection data;

[0029] The time series processing module is used to normalize the time series data, initialize the segmented sampling time set, and add the initial point and the end point of the sampling time required to the segmented sampling time set;

[0030] A segmented sequence representation module is used to sort the time series data in order according to the sampling time interval, and obtain the segmented sequence representation result of the time series data based on the sorting result and the liquid sampling amount;

[0031] The sampling interval time module is used to analyze the fitting error between the segmented sequence and the time series data through the time segmented sequence algorithm, output the segmented sampling time set based on the fitting error, and determine the sampling drive instruction according to the segmented sampling point set.

[0032] Preferably, the time series data are sorted in order according to the sampling time interval, and a segmented sequence representation result of the time series data is obtained based on the sorting result and the liquid extraction amount, including:

[0033] The time series data are sorted in order according to the sampling time interval, and a cumulative segmentation rule based on the liquid volume threshold is generated based on the sorting result and the liquid volume;

[0034] The cumulative liquid volume is calculated step by step based on the number of liquid fermentation tanks and the cumulative segmentation rule, and the time point when the liquid volume threshold is reached is recorded. The time point, the initial point and the end point of the sampling time are combined and recorded as a segmented time set;

[0035] Based on the liquid volume threshold, the average liquid volume and time interval change trends within the segmented time set are analyzed, and the segmented sequence representation results are output in combination with the liquid volume threshold and the segmented time set.

[0036] Preferably, analyzing the fitting error between the segmented sequence and the time series data by a time segmented sequence algorithm, outputting a segmented sampling time set based on the fitting error, and determining a sampling drive instruction according to the segmented sampling point set includes:

[0037] Based on the time segmentation sequence algorithm, the linear regression value of the segmented sequence is analyzed, and the linear regression value is subjected to linear interpolation to obtain the segmented time series data, and the fitting error between the time series data and the segmented time series data is analyzed;

[0038] Based on the fitting error results, the segmented sequence with the smallest fitting error is obtained for priority segmentation, and the maximum and minimum values ​​of the original curve and the fitting curve of the segmented sequence corresponding to the priority segmentation are analyzed;

[0039] Segment the segment sequence corresponding to the priority segment according to the size relationship between the maximum value and the minimum value to form a time series segment, and output a new sampling time set based on the time series segmentation result;

[0040] The new sampling time is combined with the extracted sampling points to generate a driving instruction including a time instruction, a liquid extraction volume instruction set and a target tank number for each sampling point.

[0041] Preferably, the concentration result detection and verification unit includes:

[0042] A concentration mean error calculation module is used to analyze the maximum and minimum current values ​​in the measurement process and the prediction process based on the fermentation liquid glucose concentration detection result and the concentration prediction result, and calculate the concentration mean error according to the maximum and minimum current values;

[0043] The accuracy verification model building module is used to analyze the linear relationship between the predicted concentration and the actual concentration based on the concentration mean error results, and to establish the accuracy verification model based on the linear relationship results;

[0044] The accuracy analysis feedback module is used to check the adaptability of the calibration curve to the concentration prediction by using the accuracy verification model, and to analyze the accuracy of the driving instruction to capture the concentration change based on the accuracy verification model.

[0045] Preferably, the calculation formula for the concentration mean error is:

[0046]

[0047] In the formula, F represents the mean concentration error, K M-max Indicates the maximum current during the measurement process, K M-min Indicates the minimum current value during the measurement process, K D-max Indicates the maximum current during the prediction process, K D-min represents the minimum current value during the prediction process, x M Indicates the glucose concentration test result of the fermentation broth, x D represents the concentration prediction result, and i represents the current interval value.

[0048] The beneficial effects of the present invention are:

[0049] 1. The fully automatic online detection instrument proposed in the present invention is designed with a weak signal acquisition system and a concentration detection circuit, a robotic arm sampler and other functions, so that the instrument can extract and filter the fermentation liquid from the fermentation tank at a regular time, and automatically detect the concentration of the fermentation liquid. This greatly solves the problems in the biological fermentation industry such as high glucose content in the fermentation liquid, the difficulty in online real-time detection of the fermentation process, and the inability to measure the fermentation liquid in time at night, which requires manual supervision, etc.

[0050] 2. The present invention can perform detection without human supervision, which changes the problem that traditional offline instruments must take samples manually and use injection needles to inject solutions manually. It realizes automatic sampling and automatic detection, and there is no human participation in the sampling process, and no external artificial interference factors will be introduced. There is no need for manual injection to avoid the problem of large manual injection errors and inaccurate noise concentration detection results.

[0051] 3. The present invention integrates the functions of concentration calibration curve generation, drive instruction setting, concentration detection execution and detection result verification by setting a master controller. It can automatically calibrate the correlation curve between concentration and fermentation liquid volume by analyzing the relationship between historical concentration data and curves, and automatically set the sampling point and detection time in combination with the time segmentation algorithm and historical data, avoiding the tedious manual planning of the operation path.

[0052] 4. The present invention can make full use of historical concentration detection data, dynamically adjust the curve model, ensure that the calibration results match the actual process changes, accurately plan the detection time and sampling points, ensure the balanced distribution of detection points, and reduce the risk of missing key concentration changes. Finally, by analyzing the concentration error, generating feedback signals, optimizing the calibration curve and driving instructions, the accuracy and efficiency of fermentation broth glucose concentration detection are significantly improved through intelligence, automation and dynamic optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 is a schematic diagram of a fully automatic instrument for detecting glucose concentration in fermentation liquid according to an embodiment of the present invention;

[0055] Figure 2 This is a principle block diagram of a general controller in a fully automatic instrument for detecting glucose concentration in fermentation liquid according to an embodiment of the present invention;

[0056] Figure 3 The present invention is a circuit diagram of a concentration detection circuit in a fully automatic instrument for detecting the glucose concentration in a fermentation liquid according to an embodiment of the present invention.

[0057] In the figure:

[0058] 1. Fermentation tank group; 2. Filter sampler; 3. Feed pump; 4. Sample pool; 5. Robotic arm sampler; 6. Standard sample pool; 7. Microinjection pump; 8. Glucose electrochemical detection pool; 9. Cleaning pump; 10. Cleaning pool; 11. Waste liquid pump; 12. Collection pool; 13. Concentration detection circuit; 14. Main controller; 1401. Detection concentration curve calibration unit; 1402. Drive instruction setting generation unit; 1403. Concentration detection drive execution unit; 1404. Concentration result detection verification unit. DETAILED DESCRIPTION

[0059] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0060] According to an embodiment of the present invention, a fully automatic instrument for detecting the glucose concentration in a fermentation liquid is provided.

[0061] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1-Figure 3 As shown, the fully automatic detection instrument for glucose concentration in fermentation liquid according to the embodiment of the present invention comprises a fermentation tank group 1, a filter sampler 2 is arranged on one side of the fermentation tank group 1, a feed pump 3 is arranged on one side of the filter sampler 2, a sample pool 4 is arranged on one side of the feed pump 3, a robotic arm sampler 5 is arranged on one side of the sample pool 4, and a standard sample pool 6 and a micro-injection pump 7 are arranged on both sides of the robotic arm sampler 5;

[0062] A glucose electrochemical detection cell 8 is arranged on one side of the microinjection pump 7, one side of the glucose electrochemical detection cell 8 is connected to a cleaning cell 10 via a cleaning pump 9, the bottom end of the glucose electrochemical detection cell 8 is connected to a collecting cell 12 via a waste liquid pump 11, a concentration detection circuit 13 is arranged inside the glucose electrochemical detection cell 8, and one end of the glucose electrochemical detection cell 8 is connected to a main controller 14.

[0063] In this embodiment, the master controller 14 is used to combine the preset algorithm with the historical concentration detection data to generate a concentration calibration curve and a driving instruction, use the concentration calibration curve and the driving instruction to perform the fermentation liquid glucose concentration detection work, and verify the accuracy of the concentration detection according to the detection results;

[0064] The overall controller 14 includes a detection concentration curve calibration unit 1401, a drive instruction setting generation unit 1402, a concentration detection drive execution unit 1403 and a concentration result detection verification unit 1404;

[0065] The detection concentration curve calibration unit 1401 is used to analyze the relationship between the concentration, the fermentation liquid volume and the concentration based on the historical concentration detection data and the curve relationship calibration algorithm, and establish the detection concentration calibration curve according to the relationship result.

[0066] Specifically, in this embodiment, the detection concentration curve calibration unit 1401 includes:

[0067] A current information corresponding acquisition module, used to acquire the corresponding current value in the glucose electrochemical detection cell 8 when there is no glucose solution and when a preset concentration of glucose solution is injected based on historical concentration detection data;

[0068] A relationship curve analysis module is used to describe the intensity change between the current value and the volume of the glucose solution based on the dilution effect relationship, and analyze the relationship curve between the concentration conductivity logarithm and the current according to the intensity change and the fitting algorithm;

[0069] In this embodiment, when describing the intensity change between the current value and the volume of the glucose solution based on the dilution effect relationship, and analyzing the relationship curve between the logarithm of the concentration conductivity and the current according to the intensity change and the fitting algorithm, the dilution effect relationship can be described according to the volume V of the glucose solution, the current value I, the conductivity σ and the concentration C, that is, the nonlinear relationship between the solution volume V and the current value I: I = f (V, C), assuming that the current value and the conductivity and concentration show a certain correlation (for example, a linear or power relationship), a preliminary modeling is performed: σ = k1*C, I = k2*V, k1 and k2 represent the proportional constants to be fitted, and the influence of the volume change ΔV on the current intensity change ΔI during the dilution process is analyzed: ΔI = I(V+ΔV-I(V), and the glucose solution is described based on the influence relationship. The influence trend of liquid volume on current was plotted, and the relationship curve of liquid volume and current intensity was drawn. At the same time, the relationship between the logarithm of concentration conductivity and current was analyzed through logarithmic transformation: ln(σ)=ln(k1)+ln(C), ln(I)=ln(k2)+ln(σ)+ln(V); it was preliminarily assumed that the relationship between the logarithm of concentration conductivity and current was a linear relationship: ln(I)=aln(C)+b, a and b both represented parameters to be fitted, and the least squares method was used to fit a straight line to the data points (ln(C), ln(I)). Based on the fitting results, a function expression and curve of the relationship between the logarithm of concentration conductivity and current were generated, and the generated curve of the relationship between the logarithm of concentration conductivity and current was applied to the concentration detection of glucose solution, and the unknown current value was predicted to reversely infer the solution concentration.

[0070] The objective function establishment module is used to derive a concentration relationship between the concentration conductivity logarithm and the glucose solution concentration according to the relationship curve, and use the concentration relationship as the objective function for calibrating the detection concentration curve.

[0071] In this embodiment, when the concentration relationship between the concentration conductivity logarithm and the glucose solution concentration is derived from the relationship curve, and the concentration relationship is used as the target function for calibrating the detection concentration curve, the concentration C expression can be reversely derived based on the fitted function form ln(σ)=f(C) (where e represents the base):

[0072] If ln(σ)=aC+b, then:

[0073]

[0074] If ln(σ)=aln(C)+b, then:

[0075]

[0076] If the polynomial relationship ln(σ)=aC 2 +bC+c, c represents the parameter to be fitted, then the quadratic equation needs to be solved:

[0077] C 2 +bc+(c-ln ( σ ) )=0;

[0078] The goal is to express the concentration C as a logarithmic function of the conductivity: C = q(ln(σ)), where q() is the concentration relationship obtained by fitting and derivation.

[0079] Taking C = q(ln(σ)) as the objective function for concentration curve calibration, the objective function is in the form of:

[0080]

[0081] In the formula, C actual,r represents the actual concentration value at the rth fitting, and t represents the total number of fittings.

[0082] The concentration curve calibration module is used to find the optimal solution of the detection concentration curve calibration objective function using the curve relationship calibration algorithm, and establish the glucose concentration detection calibration curve based on the optimal solution.

[0083] In this embodiment, when using the curve relationship calibration algorithm to find the optimal solution of the detection concentration curve calibration objective function and establishing the glucose concentration detection calibration curve based on the optimal solution, the detection concentration curve calibration objective function can be converted into a relative error least squares function based on the curve relationship calibration algorithm, and then the initial parameters are set to perform Taylor series expansion operation to calculate the Jacobian matrix; the relative error is analyzed based on the Jacobian matrix, and the iterative increment of the expanded detection concentration curve calibration objective function is obtained according to the relative error; the relative error sum of squares is obtained based on the iterative increment and the error analysis algorithm, and the accuracy of the iterative increment is judged by the relative error sum of squares. If the accuracy result does not meet the requirements, the iterative increment is obtained again; if the accuracy result meets the requirements, the iterative increment is substituted into the relative error least squares function as the optimal solution of the objective function, and the glucose concentration detection calibration curve is established based on this.

[0084] In this embodiment, the expression of the error analysis algorithm is:

[0085]

[0086] In the formula, S represents the relative error sum of squares, v represents the total number of iterations, and r u represents the residual error of the u-th iteration, f(y u ,p) represents the relative error least square function corresponding to the calibration objective function of the detection concentration curve at the uth iteration, and p represents the iteration increment.

[0087] It should be explained that the three-electrode chip used in the biological monitoring instrument during the concentration detection process has a good linear relationship within the concentration range of 0.5g / L to 120g / L, so calibration is required before each detection to establish the relationship between concentration, fermentation liquid volume and current. First, when the detection time is up, the instrument will clean the glucose electrochemical detection cell 8, and read the empty current, wait for the current data to stabilize (the current fluctuation is within 3nA within 15 seconds), and record the data as the low point data. Then, the robotic arm sampler 5 draws 50uL of 20g / L standard glucose solution into the sample cell 4, and injects 15uL of glucose solution into the glucose electrochemical detection cell 8. Wait for 40s and then read the current value. If the current value is stable (the current fluctuation is within 3nA within 15 seconds), record the data, repeat three times, and calculate the average value as the high point data, so as to establish a linear straight line.

[0088] Measure the current value when there is no glucose solution and the current when 15uL, 6g / L standard solution is injected, and the concentration-current curve is established to start the detection. If there is no stability within 6 minutes (the current fluctuation is within 3nA within 15 seconds), it will be recalibrated. If the calibration still fails, it is determined that the data of this round is invalid, and the sample retention function is automatically started, the refrigeration function is turned on, and the sampling time is recorded to facilitate subsequent manual detection.

[0089] In order to facilitate understanding of the above technical solution of the present invention, the operation method of calibrating the glucose concentration detection calibration curve in the actual process of the present invention is described in detail below.

[0090] Step 1: Data collection and processing;

[0091] Data collection: The experiment used a solution without glucose (pure solution) and injected glucose solutions with different preset concentrations.

[0092] Record the following data:

[0093] Glucose solution concentration C: including 0.1mol / L, 0.2mol / L, ..., 1.0mol / L;

[0094] Liquid volume V: including 10mL, 20mL, ..., 100mL;

[0095] The current value I in the electrochemical detection cell;

[0096] Measurement results (sample data):

[0097] Concentration C: [0.1, 0.2, 0.3, …, 1.0] mol / L;

[0098] Liquid volume V: [10, 20, 30, ..., 100]mL;

[0099] Current value I (measured value with noise): [0.85, 3.2, 7.1, 12.5, 20.8, 30.1, 42.7, 56.8, 72.4, 89.5] mA.

[0100] Step 2: Analyze the change of current intensity based on the dilution effect;

[0101] Assuming that the current value I, conductivity σ and solution concentration C have the following relationship: σ=k1*C, I=k2*σV; then the effect of the liquid volume change ΔV on the current intensity change ΔI during the dilution process is: ΔI=I(V+ΔV-I(V). Based on the above example: V=50mL, ΔV=10mL, ΔI=I(60mL)-I(50mL)=42.7-30.1=12.6mA. According to the measured data, draw the relationship curve between V and I.

[0102] Step 3: Logarithmic transformation to analyze the relationship between the logarithm of concentration conductivity and current;

[0103] Assume that the relationship between the concentration conductivity logarithm and the current is linear: Ln(I) = aln(C) + b;

[0104] Use the least squares method to perform a linear fit on the data points (ln(C), ln(I)):

[0105] Actual data points: ln(C) = [-2.302, -1.609, -1.204, ..., 0];

[0106] ln(I)=[-0.162, 1.163, 1.960, ..., 4.495].

[0107] Fitting results: a=1.02, b=-0.15;

[0108] The relationship curve between the logarithm of the concentration conductivity and the current is generated: ln(I)=1.02ln(C)-0.15.

[0109] Step 4: derive the relationship between the concentration conductivity logarithm and the glucose concentration;

[0110] Solve for C based on ln(I)=1.02ln(C)-0.15:

[0111]

[0112] Example: For I = 56.8 mA = 4.042, then: lnC = 4.107, C = e 4.107 ≈0.82mol / L.

[0113] Step 5: Using the concentration relationship as the target function for calibrating the detection concentration curve;

[0114] Define the objective function, optimize it based on the relative error least squares, use Taylor series expansion, calculate the Jacobian matrix (partial derivative matrix), and iteratively adjust a and b to minimize the objective function value.

[0115] Based on the optimal solution a=1.02, b=-0.15, a concentration detection calibration curve is drawn, the test data is verified, the error index is calculated, the current value I is recorded through the glucose electrochemical detection cell 8, and the unknown concentration is calculated in real time using the relationship C.

[0116] By combining the dilution effect and the fitting of the logarithmic relationship, the relationship between glucose concentration, current value and conductivity was accurately described, and a concentration detection calibration curve was successfully established, providing a reliable basis for electrochemical concentration detection.

[0117] The drive instruction setting generation unit 1402 is used to determine the detection time and sampling point based on the historical concentration detection data combined with the time segmentation sequence algorithm, and generate the detection drive instruction according to the detection time and sampling point.

[0118] In this embodiment, the drive instruction setting generation unit 1402 includes:

[0119] A time series data acquisition module is used to acquire the corresponding liquid volume, number of liquid fermentation tanks, sampling time interval and sampling time required in the glucose concentration detection process as time series data based on historical concentration detection data;

[0120] The time series processing module is used to normalize the time series data, initialize the segmented sampling time set, and add the initial point and the end point of the sampling time required to the segmented sampling time set;

[0121] A segmented sequence representation module is used to sort the time series data in order according to the sampling time interval, and obtain the segmented sequence representation result of the time series data based on the sorting result and the liquid sampling amount;

[0122] The sampling interval time module is used to analyze the fitting error between the segmented sequence and the time series data through the time segmented sequence algorithm, output the segmented sampling time set based on the fitting error, and determine the sampling drive instruction according to the segmented sampling point set.

[0123] In this embodiment, when the time series data are sorted in sequence according to the sampling time interval and the segmented sequence representation result of the time series data is obtained based on the sorting result and the liquid collection volume, the time series data can be sorted in sequence according to the sampling time interval, and a cumulative segmentation rule based on the liquid collection volume threshold is generated based on the sorting result and the liquid collection volume; the cumulative liquid collection volume is gradually calculated based on the number of liquid collection fermentation tanks and the cumulative segmentation rule, and the time point when the liquid collection volume threshold is reached is recorded, and the time point, the initial point and the end point of the sampling time are combined and recorded as a segmented time set; based on the liquid collection volume threshold, the average liquid collection volume and the time interval change trend in the segmented time set are analyzed, and the segmented sequence representation result is output in combination with the liquid collection volume threshold and the segmented time set.

[0124] In this embodiment, when the fitting error between the segmented sequence and the time series data is analyzed by the time segmented sequence algorithm, the segmented sampling time set is output based on the fitting error, and the sampling drive instruction is determined according to the segmented sampling point set, the linear regression value of the segmented sequence can be analyzed based on the time segmented sequence algorithm, and the linear regression value is subjected to linear interpolation processing to obtain the segmented time series data, and the fitting error between the time series data and the segmented time series data is analyzed; based on the fitting error result, the segmented sequence with the smallest fitting error is obtained for priority segmentation, and the maximum and minimum values ​​of the original curve and the fitting curve of the segmented sequence corresponding to the priority segmentation are analyzed; the segmented sequence corresponding to the priority segmentation is segmented according to the size relationship between the maximum value and the minimum value to form a time series segmentation, and a new sampling time set is output based on the time series segmentation result; the new sampling time is combined with the extracted sampling point to generate a drive instruction containing a time instruction, a liquid extraction volume instruction set and a target tank number for each sampling point.

[0125] It should be explained that the predicted fermentation trend in each fermentation tank (i.e. the predicted value of the corresponding concentration of glucose in the fermentation tank at each time point) can be input into the instrument before the test. Each test will fit and correct the fermentation trend based on the previous test results, and predict the next result (Monod model).

[0126] The Monod model describes the relationship between microbial growth rate and substrate concentration. It assumes that the growth rate of microorganisms is proportional to the substrate concentration. The mathematical expression is: μ = μ max *O / (K s +O), where μ represents the microbial growth rate, μ max represents the maximum growth rate, O represents the substrate concentration, K s Indicates the influence factor of substrate concentration on growth rate. Compare the test result with the predicted result. If the difference between the comparison results is large, re-test immediately. If the result is still large compared with the predicted result, take the average of the test results and save it as the latest test result.

[0127] In order to facilitate understanding of the above technical solution of the present invention, the operation method of generating driving instructions in the actual process of the present invention is described in detail below.

[0128] Step 1: Data collection and preprocessing;

[0129] Based on the historical concentration detection time series data, obtain information on the glucose concentration detection working process:

[0130] Liquid volume: the liquid volume sampled each time (e.g. 10 mL); Number of fermentation tanks for liquid sampling**: the number of fermentation tanks to be sampled (e.g. 3 tanks); Sampling interval: the sampling interval (e.g. sampling every 20 minutes); Sampling time: the time required for the sampling operation (e.g. 5 minutes);

[0131] Then the time series data can be expressed as:

[0132] Time point (minutes): [0, 20, 40, 60, 80, 100, 120];

[0133] Liquid volume (mL): [10, 20, 15, 30, 25, 35, 40];

[0134] Normalize the time series data (to the range of 0-1) and then:

[0135] time: [0, 0.167, 0.333, 0.5, 0.667, 0.833, 1.0];

[0136] Liquid volume: [0.25, 0.5, 0.375, 0.75, 0.625, 0.875, 1.0];

[0137] Initialize the segmented sampling time set: the initial point of the sampling time required: 0 minutes; the end point of the sampling time required: 120 minutes, the initial segment set: [0, 120].

[0138] Step 2: Segment sequence generation;

[0139] Sort the time series by time point (the data is already in order), assume that the liquid extraction threshold for each fermentation tank is 60 mL, and calculate the cumulative liquid extraction:

[0140] Time point 0 minutes: 10 mL of liquid was taken, and the cumulative liquid volume was 10 mL;

[0141] Time point 20 minutes: 20 mL of liquid was collected, and the cumulative liquid volume was 30 mL;

[0142] Time point 40 minutes: 15 mL of liquid was collected, and the cumulative liquid volume was 45 mL;

[0143] Time point 60 minutes: 30 mL of liquid was collected, and the cumulative liquid volume was 75 mL;

[0144] If the threshold is exceeded, the time point is recorded for 60 minutes;

[0145] The new segment time sets are [0, 60] and [60, 120].

[0146] Step 3: Segmented time set analysis;

[0147] Calculate the average liquid volume and time interval change trend within the segment:

[0148] Average liquid volume in segment 1 (0-60 minutes): (10+20+15+30) / 4=18.75 mL; time interval: (60-0) / 4=15 minutes;

[0149] Average liquid volume in segment 2 (60-120 minutes): (25+35+40) / 3=33.33 mL; time interval: (120-60) / 3=20 minutes;

[0150] The segmented sequence representation result is:

[0151] Segment 1: time range 0-60 minutes, average liquid volume 18.75 mL, time interval 15 minutes; Segment 2: time range 60-120 minutes, average liquid volume 33.33 mL, time interval 15 minutes.

[0152] Step 4: Fitting error analysis and priority segmentation;

[0153] Perform linear regression fitting on the segmented sequence and the original time series data:

[0154] Regression equation for segment 1: y = 0.3125x + 10, regression equation for segment 2: y = 0.5x + 25;

[0155] Calculate the fitting error (mean square error):

[0156] Segment 1: MSE = 1.56; Segment 2: MSE = 2.25;

[0157] Since the fitting error of segment 1 is small, segment 1 is selected for optimization first, and the maximum and minimum values ​​of the original curve (actual liquid volume) and the fitting curve (regression equation) of segment 1 are analyzed:

[0158] Maximum value of the original curve: 30 mL (time point 60 minutes);

[0159] Minimum value of the original curve: 10 mL (time point 0 minutes);

[0160] Maximum value of the fitting curve: 26.25 mL (time point 60 minutes);

[0161] Minimum value of fitting curve: 11.25 mL (time point 0 minutes);

[0162] According to the size relationship between the maximum and minimum values, segment 1 is further subdivided:

[0163] Segment 1-1: time range 0-40 minutes, average liquid volume 15 mL;

[0164] Segment 1-2: time range 40-60 minutes, average fluid volume 30 mL;

[0165] Updated sampling time set: [0, 40, 60, 120];

[0166] Extract sampling points using the new set of sampling times:

[0167] At time point 0 minutes, 10 mL of liquid was taken; at time point 40 minutes, 15 mL of liquid was taken; at time point 60 minutes, 30 mL of liquid was taken; at time point 120 minutes, 40 mL of liquid was taken.

[0168] Generate a drive instruction containing time instruction, liquid volume instruction set and target tank number for each sampling point:

[0169] Time point 0 minutes: start sampling, target tank number 1, liquid volume 10mL;

[0170] Time point 40 minutes: start sampling, target tank number 2, liquid volume 15mL;

[0171] Time point 60 minutes: start sampling, target tank number 3, liquid volume 30mL;

[0172] Time point 120 minutes: Start sampling, target tank number 4, liquid volume 40mL.

[0173] In actual operation, sampling operation is performed according to the generated sampling drive instructions, and the actual liquid volume and sampling time are recorded. The sampling drive instructions are applied to the glucose concentration detection process to ensure that sampling is performed according to the predetermined sampling time, liquid volume and target tank number. Based on the actual sampling results, the sampling drive instructions are fed back and optimized to improve sampling efficiency and accuracy.

[0174] Therefore, by combining historical concentration detection data and time segmentation sequence algorithm, accurate sampling drive instructions were successfully generated. The drive instructions not only took into account the sampling time and liquid volume, but also the number of liquid fermentation tanks and the time required for sampling, ensuring the efficiency and accuracy of the sampling process.

[0175] The concentration detection drive execution unit 1403 is used to start the robot arm sampler 5 and the glucose electrochemical detection cell 8 to perform fermentation liquid glucose concentration detection according to the drive instruction, and obtain the fermentation liquid glucose concentration in combination with the concentration calibration curve and the concentration detection circuit 13.

[0176] The concentration result detection and verification unit 1404 is used to analyze the concentration error between the glucose concentration of the fermentation liquid and the concentration prediction result, and to build a verification model based on the error result to analyze the accuracy of the detection concentration calibration curve and the driving instruction.

[0177] In this embodiment, the concentration result detection and verification unit 1404 includes:

[0178] The concentration mean error calculation module is used to analyze the maximum and minimum current values ​​in the measured process and the predicted process based on the fermentation liquid glucose concentration detection results and the concentration prediction results, and calculate the concentration mean error according to the maximum and minimum current values.

[0179] In this embodiment, the calculation formula of the concentration mean error is:

[0180]

[0181] In the formula, F represents the mean concentration error, K M-max Indicates the maximum current during the measurement process, K M-min Indicates the minimum current value during the measurement process, K D-max Indicates the maximum current during the prediction process, K D-min represents the minimum current value during the prediction process, x M Indicates the glucose concentration test result of the fermentation broth, x D represents the concentration prediction result, and i represents the current interval value.

[0182] The accuracy verification model building module is used to analyze the linear relationship between the predicted concentration and the actual concentration based on the concentration mean error results, and to establish an accuracy verification model based on the linear relationship results.

[0183] The accuracy analysis feedback module is used to check the adaptability of the calibration curve to the concentration prediction by using the accuracy verification model, and to analyze the accuracy of the driving instruction to capture the concentration change based on the accuracy verification model.

[0184] It should be explained that the specific implementation steps of the accuracy verification model building module and the accuracy analysis feedback module are as follows:

[0185] Step 1: Use the accuracy verification model to check the adaptability of the calibration curve to the concentration prediction;

[0186] Collect the actual concentration data and predicted concentration data used for calibration in the experiment. The data format should include: actual concentration C actual , calibration curve predicts concentration C pred , using the indicators in the accuracy validation model (e.g., mean error, relative error) to evaluate the predicted concentration of the calibration curve;

[0187] Calibration curve fitness analysis: Plot the predicted concentration against the actual concentration to see if there is a linear trend, check if there is a large error or prediction failure within a specific concentration range, and evaluate whether the calibration curve meets actual testing needs: If the error between the calibration curve and the actual concentration is large, the calibration curve parameters may need to be adjusted or refitted.

[0188] Step 2: Based on the accuracy verification model, analyze the accuracy of the driving instructions to capture the concentration changes;

[0189] Perform concentration detection according to the sampling time point and liquid volume set by the drive instruction, and collect the predicted concentration C of the instruction-driven sampling pred The actual concentration C actual , sampling time point.

[0190] Compare the actual concentration and predicted concentration at the sampling time point to analyze whether the drive instruction accurately captures the concentration change trend. This can be done by drawing a curve of concentration change over time:

[0191] Real change curve C actual (t) and the concentration change curve C predicted by the driving instruction pred (t), use the accuracy verification model to calculate the error at each time point, and analyze the error trend over time: if the error is large at some time points, it may indicate that the drive command failed to respond to the concentration change in time. If the concentration change is not captured accurately, analyze the possible reasons (such as the sampling interval is too large, the sampling volume is insufficient, etc.).

[0192] In order to facilitate understanding of the above technical solutions of the present invention, the working principle or operation mode of the present invention in the actual process is described in detail below.

[0193] In actual application, the instrument can detect substances in four fermentation tanks in turn. Each time it is used, the number of fermentation tanks to be sampled, the sampling interval and the total number of hours required for detection are set on the host computer time setting page at the main controller 14. The machine will automatically calculate each sampling point based on the detection time, and perform sampling detection when the sampling point is reached. After setting the parameters, click the start button to start sampling.

[0194] The fermentation liquid will first flow through the filter sampler 2, which has a built-in nanomembrane. Driven by the feed pump 3, the glucose fermentation liquid passes through the nanomembrane to filter out large particles such as bacteria and fermentation products, leaving only the glucose fermentation liquid and sending it to the sample pool 4. The robotic arm sampler 5 carries a micro-injection pump 7. After extracting 100uL of glucose fermentation liquid from the sample pool 4, it is sent to the glucose electrochemical detection pool 8. Before sending it in, the detection pool will be cleaned and the curve will be calibrated. The current value when there is no glucose solution and the current when 15uL, 6g / L standard glucose fermentation liquid is injected are measured. The concentration and current curve are established to start the detection. First, 5uL of glucose fermentation liquid is sent for initial detection. The injection volume of the second needle is determined according to the initial detection concentration. There are a total of 4 levels: 2uL, 5uL, 15uL and 40uL, corresponding to low concentration, medium concentration, high concentration and ultra-high concentration respectively.

[0195] The concentration of the glucose solution is detected by the concentration detection circuit 13, and the information is fed back to the host computer. Finally, the glucose electrochemical detection cell 8 is cleaned. According to this step, the selected fermentation tanks are detected in turn. After completion, the machine enters a dormant state and waits for the next detection time to arrive. If the value of a certain detection is lower than the set lower limit, a detection will be performed immediately. If both results are lower than the set threshold, it indicates that the glucose substrate in the fermentation tank has been consumed. This fermentation needs to be ended or glucose needs to be supplemented. The instrument will automatically send the alarm threshold to the server through the Internet of Things, and the server will notify the user to go to the fermentation site for processing in the form of text messages.

[0196] It should be explained that the concentration detection circuit 13 includes a three-electrode glucose sensor chip. The three-electrode glucose sensor chip has three electrodes, namely, a working electrode, a counter electrode, and a reference electrode. When the three-electrode glucose chip is working, a constant voltage needs to be applied between the working electrode and the reference electrode to promote the chemical reaction and generate electrons. The generated electrons pass through a transimpedance amplifier to convert the current signal into a voltage signal, which is then sent to the ADC chip for collection.

[0197] In summary, with the help of the above technical scheme of the present invention, the fully automatic online detection instrument proposed by the present invention is designed with a weak signal acquisition system and a concentration detection circuit 13, a robotic arm sampler 5 and other functions, so that the instrument can extract and filter the fermentation liquid from the fermentation tank at regular intervals, and automatically detect the concentration of the fermentation liquid, which greatly solves the problems of high glucose content concentration in the fermentation liquid, difficulty in online real-time detection of the fermentation process, inability to measure the fermentation liquid in time at night, and the need for manual supervision. The present invention can perform detection without supervision, which changes the problem that traditional offline instruments must be manually sampled and manually use injection needles to inject solutions, and realizes automatic sampling and automatic detection. There is no manual participation in the sampling process, and no external artificial interference factors are introduced. There is no need for manual injection to avoid the problem of large manual injection errors and inaccurate noise concentration detection results.

[0198] The present invention integrates the functions of concentration calibration curve generation, drive instruction setting, concentration detection execution and detection result verification by setting a master controller 14. The correlation curve between concentration and fermentation liquid volume can be automatically calibrated by analyzing the relationship between historical concentration data and the curve, and the sampling point and detection time can be automatically set in combination with the time segmentation algorithm and historical data, thereby avoiding the tedious manual planning of the operation path, making full use of the historical concentration detection data, dynamically adjusting the curve model, ensuring that the calibration result matches the actual process changes, accurately planning the detection time and sampling points, ensuring that the detection points are evenly distributed, and reducing the risk of missing key concentration changes. Finally, by analyzing the concentration error, generating a feedback signal, optimizing the calibration curve and the drive instruction, the accuracy and efficiency of the fermentation liquid glucose concentration detection are significantly improved through intelligence, automation and dynamic optimization.

[0199] In the present invention, unless otherwise clearly stipulated and limited, the terms such as "installation", "setting", "connection", "fixation" and "screw-on" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. Unless otherwise clearly defined, ordinary technicians in this field can understand the specific meanings of the above terms in the present invention according to the specific circumstances.

[0200] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A fully automatic instrument for detecting glucose concentration in fermentation liquid, comprising a fermentation tank group (1), characterized in that: A filter sampler (2) is provided on one side of the fermentation tank group (1), a feed pump (3) is provided on one side of the filter sampler (2), a sample pool (4) is provided on one side of the feed pump (3), a robotic arm sampler (5) is provided on one side of the sample pool (4), and a standard sample pool (6) and a micro-injection pump (7) are provided on both sides of the robotic arm sampler (5); A glucose electrochemical detection cell (8) is provided on one side of the microinjection pump (7); one side of the glucose electrochemical detection cell (8) is connected to a cleaning cell (10) via a cleaning pump (9); the bottom end of the glucose electrochemical detection cell (8) is connected to a collection cell (12) via a waste liquid pump (11); a concentration detection circuit (13) is provided inside the glucose electrochemical detection cell (8); and one end of the glucose electrochemical detection cell (8) is connected to a master controller (14).

2. A fully automatic instrument for detecting glucose concentration in fermentation liquid according to claim 1, characterized in that: The master controller (14) is used to combine a preset algorithm with historical concentration detection data to generate a concentration calibration curve and a driving instruction, use the concentration calibration curve and the driving instruction to perform fermentation liquid glucose concentration detection work, and verify the accuracy of the concentration detection according to the detection result; The master controller (14) comprises a detection concentration curve calibration unit (1401), a drive instruction setting generation unit (1402), a concentration detection drive execution unit (1403) and a concentration result detection verification unit (1404); The detection concentration curve calibration unit (1401) is used to analyze the relationship between concentration, fermentation liquid volume and concentration based on historical concentration detection data and curve relationship calibration algorithm, and establish a detection concentration calibration curve according to the relationship result; The drive instruction setting generation unit (1402) is used to determine the detection time and sampling point based on the historical concentration detection data combined with the time segmentation sequence algorithm, and generate the detection drive instruction according to the detection time and sampling point; The concentration detection drive execution unit (1403) is used to start the robot arm sampler (5) and the glucose electrochemical detection cell (8) according to the drive instruction to perform fermentation liquid glucose concentration detection work, and obtain the fermentation liquid glucose concentration in combination with the concentration calibration curve and the concentration detection circuit (13); The concentration result detection and verification unit (1404) is used to analyze the concentration error between the glucose concentration of the fermentation liquid and the concentration prediction result, and to build a verification model based on the error result to analyze the accuracy of the detection concentration calibration curve and the driving instruction.

3. A fully automatic detection instrument for glucose concentration in fermentation liquid according to claim 2, characterized in that: The detection concentration curve calibration unit (1401) comprises: A current information corresponding acquisition module, used for acquiring corresponding current values ​​in the glucose electrochemical detection cell (8) when a glucose-free solution and a glucose solution of a preset concentration and volume are injected based on historical concentration detection data; A relationship curve analysis module is used to describe the intensity change between the current value and the volume of the glucose solution based on the dilution effect relationship, and analyze the relationship curve between the concentration conductivity logarithm and the current according to the intensity change and the fitting algorithm; An objective function establishment module is used to derive a concentration relationship between the logarithm of the concentration conductivity and the concentration of the glucose solution according to the relationship curve, and use the concentration relationship as a target function for calibrating the detection concentration curve; The concentration curve calibration module is used to find the optimal solution of the detection concentration curve calibration objective function using the curve relationship calibration algorithm, and establish the glucose concentration detection calibration curve based on the optimal solution.

4. A fully automatic instrument for detecting glucose concentration in fermentation liquid according to claim 3, characterized in that: The method of using the curve relationship calibration algorithm to find the optimal solution of the detection concentration curve calibration objective function and establishing the glucose concentration detection calibration curve based on the optimal solution includes: After converting the detection concentration curve calibration objective function into a relative error least square function based on the curve relationship calibration algorithm, the initial parameters are set to perform Taylor series expansion operations to calculate the Jacobian matrix; The relative error is analyzed based on the Jacobian matrix, and the iterative increment of the calibration objective function of the expanded detection concentration curve is obtained according to the relative error; The relative error sum of squares is obtained based on the iterative increment and error analysis algorithm, and the accuracy of the iterative increment is judged by the relative error sum of squares. If the accuracy result does not meet the requirements, the iterative increment is obtained again; If the accuracy result meets the requirements, the iterative increment is substituted into the relative error least squares function as the optimal solution of the objective function, and a glucose concentration detection calibration curve is established based on it.

5. A fully automatic instrument for detecting glucose concentration in fermentation liquid according to claim 4, characterized in that: The expression of the error analysis algorithm is: In the formula, S represents the relative error sum of squares, v represents the total number of iterations, and r u represents the residual error of the u-th iteration, f(y u ,p) represents the relative error least square function corresponding to the calibration objective function of the detection concentration curve at the uth iteration, and p represents the iteration increment.

6. The fully automatic detection instrument for glucose concentration in fermentation liquid according to claim 2, characterized in that: The driving instruction setting generation unit (1402) comprises: A time series data acquisition module is used to acquire the corresponding liquid volume, number of liquid fermentation tanks, sampling time interval and sampling time required in the glucose concentration detection process as time series data based on historical concentration detection data; The time series processing module is used to normalize the time series data, initialize the segmented sampling time set, and add the initial point and the end point of the sampling time required to the segmented sampling time set; A segmented sequence representation module is used to sort the time series data in order according to the sampling time interval, and obtain the segmented sequence representation result of the time series data based on the sorting result and the liquid sampling amount; The sampling interval time module is used to analyze the fitting error between the segmented sequence and the time series data through the time segmented sequence algorithm, output the segmented sampling time set based on the fitting error, and determine the sampling drive instruction according to the segmented sampling point set.

7. The fully automatic detection instrument for glucose concentration in fermentation liquid according to claim 6, characterized in that: The step of sorting the time series data in order according to the sampling time intervals and obtaining the segmented sequence representation result of the time series data based on the sorting result and the liquid extraction amount includes: The time series data are sorted in order according to the sampling time interval, and a cumulative segmentation rule based on the liquid volume threshold is generated based on the sorting result and the liquid volume; The cumulative liquid volume is calculated step by step based on the number of liquid fermentation tanks and the cumulative segmentation rule, and the time point when the liquid volume threshold is reached is recorded. The time point, the initial point and the end point of the sampling time are combined and recorded as a segmented time set; Based on the liquid volume threshold, the average liquid volume and time interval change trends within the segmented time set are analyzed, and the segmented sequence representation results are output in combination with the liquid volume threshold and the segmented time set.

8. The fully automatic instrument for detecting glucose concentration in fermentation liquid according to claim 7, characterized in that: The method of analyzing the fitting error between the segmented sequence and the time series data by a time segmented sequence algorithm, outputting a segmented sampling time set based on the fitting error, and determining a sampling drive instruction according to the segmented sampling point set includes: Based on the time segmentation sequence algorithm, the linear regression value of the segmented sequence is analyzed, and the linear regression value is subjected to linear interpolation to obtain the segmented time series data, and the fitting error between the time series data and the segmented time series data is analyzed; Based on the fitting error results, the segmented sequence with the smallest fitting error is obtained for priority segmentation, and the maximum and minimum values ​​of the original curve and the fitting curve of the segmented sequence corresponding to the priority segmentation are analyzed; Segment the segment sequence corresponding to the priority segment according to the size relationship between the maximum value and the minimum value to form a time series segment, and output a new sampling time set based on the time series segmentation result; The new sampling time is combined with the extracted sampling points to generate a driving instruction including a time instruction, a liquid extraction volume instruction set and a target tank number for each sampling point.

9. The fully automatic instrument for detecting glucose concentration in fermentation liquid according to claim 2, characterized in that: The concentration result detection and verification unit (1404) comprises: A concentration mean error calculation module is used to analyze the maximum and minimum current values ​​in the measurement process and the prediction process based on the fermentation liquid glucose concentration detection result and the concentration prediction result, and calculate the concentration mean error according to the maximum and minimum current values; The accuracy verification model building module is used to analyze the linear relationship between the predicted concentration and the actual concentration based on the concentration mean error results, and to establish the accuracy verification model based on the linear relationship results; The accuracy analysis feedback module is used to check the adaptability of the calibration curve to the concentration prediction by using the accuracy verification model, and to analyze the accuracy of the driving instruction to capture the concentration change based on the accuracy verification model.

10. The fully automatic detection instrument for glucose concentration in fermentation liquid according to claim 9, characterized in that: The calculation formula of the concentration mean error is: In the formula, F represents the mean concentration error, K M-max Indicates the maximum current during the measurement process, K M-min Indicates the minimum current value during the measurement process, K D-max Indicates the maximum current during the prediction process, K D-min represents the minimum current value during the prediction process, x M Indicates the test result of glucose concentration in fermentation broth, x D represents the concentration prediction result, and i represents the current interval value.

Images