Intelligent titration analysis system based on artificial intelligence

Abstract

The invention relates to the technical field of titrimetric analysis automation, in particular to an intelligent titrimetric analysis system based on artificial intelligence, which comprises an error identification module, a temperature adjustment module, a viscosity compensation module, a flow adjustment module and a bubble correction module. In the invention, by independently quantifying the error and identifying the key parameter offset in real time, the data reliability in the titration process is obviously improved, the error interference is reduced, the temperature change is dynamically adjusted and coped, the reagent viscosity and flow fluctuation are adjusted in real time, and the bubble interference and density offset are identified and corrected; the accuracy and the consistency of the test in various environments are ensured, the adaptability, the stability and the repeatability of titration analysis are improved, and a high-precision and traceable experiment result is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of titration analysis automation, and in particular to an intelligent titration analysis system based on artificial intelligence. Background Art

[0002] The field of automated titration analysis technology includes analytical measurement technology based on chemical reactions. By controlling the amount of titration reagent added and combining the endpoint indication method, the concentration or other characteristics of the substance to be measured are determined. The core content of this technical field involves titration methods, endpoint determination methods, reagent control, data acquisition and processing. Titration methods mainly include acid-base titration and redox titration, which analyze different types of substances to be measured respectively. The endpoint determination method covers indicator color change, potentiometric titration and spectral titration, etc., and the titration endpoint is determined by the change of physical or chemical signals. Reagent control involves flow regulation, titration speed control and microinjection device to ensure the accurate addition of reagents. Data acquisition and processing include real-time acquisition and storage of data such as potential, current, absorbance, etc. during the experiment, providing a basis for subsequent analysis. Titration analysis automation technology is widely used in chemical laboratories, environmental monitoring, food testing and pharmaceutical analysis, and realizes efficient and accurate analysis and determination through computer control systems, sensors and automatic titration devices.

[0003] Among them, the intelligent titration analysis system based on artificial intelligence refers to a system that combines artificial intelligence algorithms with titration analysis automation technology to achieve parameter optimization, intelligent endpoint determination and automatic data analysis of the titration process. Aiming at the problems of reagent addition accuracy, endpoint determination error and data processing complexity in the titration process, the system uses a machine learning model to train the titration process parameters, adjusts the reagent titration rate and endpoint determination threshold based on historical experimental data, analyzes the titration curve characteristics through a neural network algorithm to optimize the endpoint recognition accuracy, and uses real-time data stream processing technology to dynamically analyze and detect anomalies in experimental data. The system adjusts the reagent drop rate through a high-precision flow control device, uses potential or spectral signal feedback for adaptive adjustment, and combines the database system to store experimental process data and titration parameters, providing a traceability analysis basis for subsequent experiments.

[0004] The existing technology has limitations in the identification and quantitative analysis of error sources during the titration process. It usually relies on a single parameter or a fixed threshold for adjustment, and it is difficult to independently optimize different error sources, resulting in error superposition that affects the experimental accuracy. The impact of temperature changes is often ignored or compensated with lag, making the reagent flow rate susceptible to interference from ambient temperature fluctuations, causing the titration endpoint to drift. Changes in reagent viscosity are not fully considered, which may lead to unstable flow control and deviations in titration rates in different batches of tests. Short-term flow fluctuations cannot be effectively suppressed, resulting in increased errors in endpoint determination, especially in experiments with high precision requirements. The lack of systematic compensation for density changes and bubble interference can easily lead to unstable factors in the titration process, affecting the accuracy and consistency of the measurement results. These shortcomings limit the adaptability of existing technologies in complex experimental environments and make it difficult to meet the needs of high-precision analysis. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an intelligent titration analysis system based on artificial intelligence.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme: an intelligent titration analysis system based on artificial intelligence comprises:

[0007] The error identification module obtains flow rate, electrode potential offset, endpoint error, temperature, calculates offset, screens significant parameters, analyzes error correlation, screens key error sources, and obtains error impact ranking values;

[0008] The temperature adjustment module calls the error impact ranking value, obtains the environment and reagent temperature, calculates the gradient, determines the rate threshold, adjusts the pump flow rate, corrects the drive voltage, and obtains the flow value after temperature correction;

[0009] The viscosity compensation module calls the temperature-corrected flow value, obtains the reagent viscosity reference and real-time viscosity, calculates the offset and error, adjusts the pump pulse frequency, and obtains the viscosity-corrected flow value;

[0010] The flow adjustment module calls the viscosity-corrected flow value, obtains the flow reference and real-time value, calculates the change gradient, determines the over-limit, calculates the recovery offset value, corrects the titration flow, and obtains the flow stability adjustment value;

[0011] The bubble correction module calls the flow stability adjustment value, obtains the density reference and real-time density, calculates the offset, determines the bubble interference, adjusts the driving voltage, and obtains the analysis correction result.

[0012] As a further scheme of the present invention, the error impact ranking value includes the ranking of significant offset parameters, the ranking of error sources, and the correlation strength index; the temperature-corrected flow value is specifically the flow velocity adjustment value, the corrected driving voltage, and the adjustment effect evaluation; the viscosity-corrected flow value includes the adjusted pulse frequency, the correction effect evaluation, and the viscosity adjustment feedback; the flow stability adjustment value is specifically the flow recovery efficiency, the fluctuation control effect, and the stability improvement index; the titration analysis correction result includes the density correction value, the bubble impact evaluation, and the corrected accuracy improvement.

[0013] As a further solution of the present invention, the error identification module includes:

[0014] The flow and potential offset calculation submodule obtains the titration reagent flow and electrode potential offset, calculates the flow offset and potential offset, calls the ambient temperature data, and calculates the offset ratio based on the error reference range using the formula:

[0015]

[0016] The deviation ratio is calculated and the error of endpoint determination is analyzed;

[0017] Among them, R f represents the deviation ratio, Q m represents the measured titration reagent flow rate, Q r represents the base flow rate, E m represents the measured electrode potential, E r represents the reference potential, T represents the ambient temperature, T r Represents the reference ambient temperature.

[0018] The endpoint determination error analysis submodule calls the offset ratio, analyzes the endpoint determination error, obtains the endpoint determination error offset, calculates the endpoint determination error offset ratio based on the error reference range, and obtains the endpoint error ratio;

[0019] The key error source screening submodule calls the endpoint error ratio, analyzes the change correlation of the differential error factor, screens the significant deviation parameters, identifies the key error sources based on the correlation analysis, and obtains the error impact ranking value.

[0020] As a further solution of the present invention, the temperature adjustment module comprises

[0021] The temperature change gradient calculation submodule calls the error impact ranking value, obtains the laboratory environment temperature and reagent temperature benchmark, calculates the temperature change gradient, obtains the temperature data of multiple time nodes based on the temperature sampling points during the experiment, calculates the temperature change amplitude between adjacent time nodes, and normalizes according to the set time interval to calculate the temperature change gradient using the formula:

[0022]

[0023] Calculate the temperature change gradient and make rate judgment;

[0024] Among them, G θ represents the temperature gradient, Θ j represents the temperature value at the jth time point, Θ j-1 represents the temperature value at the previous time point, s represents the total number of sampling points, Δτ represents the time interval, Θ env represents the laboratory ambient temperature, Θ std Represents the reagent temperature reference;

[0025] The flow adjustment submodule calls the temperature change gradient, sets a threshold range, determines whether the rate exceeds the threshold range, adjusts the flow rate of the titration reagent pump according to the set flow adjustment standard, calculates the adjusted flow offset, obtains the flow adjustment offset, and corrects the titration pump drive voltage;

[0026] The temperature-corrected flow calculation submodule calls the flow adjustment offset, corrects the titration pump driving voltage according to the offset, calculates the temperature-corrected flow value, and obtains the temperature-corrected flow value.

[0027] As a further solution of the present invention, the viscosity compensation module includes:

[0028] The viscosity offset calculation submodule calls the temperature-corrected flow value, obtains the titration reagent viscosity reference and real-time reagent viscosity, calculates the viscosity offset, records the reagent viscosity values ​​at multiple time points based on the measurement data during the experiment, and calculates the deviation from the reference viscosity using the formula:

[0029]

[0030] Calculate the viscosity offset and calculate the flow error;

[0031] Among them, Δη represents the viscosity deviation, η p represents the viscosity of the reagent measured at the pth time point, η b represents the viscosity standard of the reagent, and w represents the number of measurements;

[0032] The flow error adjustment submodule calls the viscosity offset, calculates the flow error caused by the viscosity change, calculates the offset amplitude of the viscosity change on the flow based on the liquid flow characteristics, and obtains the viscosity change flow error through the flow influence ratio correction formula;

[0033] The pulse frequency correction submodule calls the viscosity change flow error, adjusts the titration pump pulse frequency, adjusts the pulse control signal according to the correction amount, dynamically compensates for the flow deviation, and obtains the viscosity-corrected flow value.

[0034] As a further solution of the present invention, the flow adjustment module includes:

[0035] The flow change gradient calculation submodule calls the viscosity-corrected flow value, obtains the standard titration flow benchmark and real-time reagent flow, calculates the flow instantaneous change gradient index, calculates the flow change amplitude at adjacent time points based on the flow measurement values ​​at differentiated time points during the experiment, and calculates the flow change gradient in combination with the time interval normalization, using the formula:

[0036]

[0037] Calculate the instantaneous flow gradient index to make fluctuation judgment;

[0038] Among them, q Represents the instantaneous flow gradient index, Q r represents the flow value at the rth time point, Q r-1 represents the flow value at the previous time point, s represents the total number of sampling points, Δξ represents the time interval, Q std Represents the standard titration flow rate reference, Q mean Represents the average value of real-time reagent flow;

[0039] The flow recovery offset calculation submodule calls the flow instantaneous change gradient index to determine whether the fluctuation exceeds the range, and determines whether the instantaneous flow fluctuation exceeds the set range according to the flow change threshold set in the experiment, and calculates the flow recovery offset value;

[0040] The flow stabilization correction submodule calls the flow recovery offset value, corrects the titration flow, adjusts the output stability of the titration pump, adjusts the pulse drive signal, dynamically compensates for flow fluctuations, and obtains a flow stabilization adjustment value.

[0041] As a further solution of the present invention, the bubble correction module includes:

[0042] The bubble interference determination submodule obtains the reagent density reference and the real-time reagent density based on the flow stability adjustment value, analyzes whether there is abnormal fluctuation in the reagent density by comparing the difference between the real-time reagent density and the reference density, calculates the mean of the density change amplitude, sets the density abnormality threshold, and determines whether there is bubble interference. If the density change amplitude exceeds the set threshold, it is marked that there is bubble interference in the current reagent flow, otherwise it is determined that the reagent flow is normal, and the bubble interference determination result is obtained;

[0043] The density offset calculation submodule calls the flow stability adjustment value, obtains the reagent density benchmark and the real-time reagent density, selects density data within a period of time, compares the measured real-time reagent density with the reagent density benchmark, and calculates the density offset using the absolute deviation mean calculation method, using the formula:

[0044]

[0045] Calculate the density offset and make error correction;

[0046] Among them, Θ D represents the density offset, D i represents the reagent density measured for the i-th time, D ref represents the reagent density reference, and N represents the number of measurements;

[0047] The flow error correction submodule determines the source of the flow error based on the density offset and the bubble interference judgment result. If there is bubble interference, the instantaneous flow error caused by the bubble is calculated. If there is no bubble interference, the instantaneous flow error caused by the density change is calculated, the titration pump driving voltage is corrected, the titration flow is adjusted, and the titration analysis correction result is obtained.

[0048] Compared with the prior art, the advantages and positive effects of the present invention are:

[0049] In the present invention, by independently quantifying errors and identifying key parameter offsets in real time, the data reliability of the titration process is significantly improved and error interference is reduced. Dynamic adjustment is performed to cope with temperature changes, reagent viscosity and flow fluctuations are adjusted in real time, and bubble interference and density offset are identified and corrected, all of which ensure the accuracy and consistency of the test in various environments, improve the adaptability, stability and repeatability of titration analysis, and ensure high-precision and traceable experimental results. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a system flow chart of the present invention;

[0051] Figure 2 is a flow chart of the error identification module of the present invention;

[0052] Figure 3 It is a flow chart of the temperature adjustment module of the present invention;

[0053] Figure 4 is a flow chart of the viscosity compensation module of the present invention;

[0054] Figure 5 It is a flow chart of the flow adjustment module of the present invention;

[0055] Figure 6 It is a flow chart of the bubble correction module of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0057] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are 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 limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0058] See also Figure 1 , an intelligent titration analysis system based on artificial intelligence includes:

[0059] The error identification module obtains the titration reagent flow, electrode potential offset, endpoint judgment error, and ambient temperature, calculates the flow offset, potential offset, and endpoint judgment error offset, calculates the offset ratio based on the error reference range, calls the titration process parameter change trajectory, screens the parameters with significant offset, analyzes the change correlation of the differential error factors, screens the key error sources based on the correlation, and obtains the error impact ranking value;

[0060] The temperature adjustment module calls the error impact sorting value, obtains the laboratory environment temperature and reagent temperature benchmark, calculates the temperature change gradient, sets and determines whether the rate exceeds the threshold range, adjusts the titration reagent pump flow rate, calculates the adjusted flow offset, corrects the titration pump drive voltage according to the offset, and obtains the temperature-corrected flow value;

[0061] The viscosity compensation module calls the temperature-corrected flow value, obtains the titration reagent viscosity benchmark and real-time reagent viscosity, calculates the viscosity offset and flow influence ratio, calculates the flow error caused by the viscosity change, adjusts the titration pump pulse frequency, and obtains the viscosity-corrected flow value;

[0062] The flow adjustment module calls the viscosity-corrected flow value, obtains the standard titration flow reference, real-time reagent flow, calculates the instantaneous flow gradient index, determines whether the fluctuation exceeds the range, calculates the flow recovery offset value, corrects the titration flow, and obtains the flow stability adjustment value;

[0063] The bubble correction module calls the flow stability adjustment value, obtains the reagent density benchmark and real-time reagent density, calculates the density offset, determines whether there is bubble interference, calculates the instantaneous flow error, adjusts the titration pump drive voltage, and obtains the titration analysis correction result.

[0064] The error impact ranking values ​​include the ranking of significant offset parameters, the ranking of error sources, and the correlation strength index. The flow value after temperature correction includes the flow velocity adjustment value, the corrected driving voltage, and the adjustment effect evaluation. The flow value after viscosity correction includes the adjusted pulse frequency, the correction effect evaluation, and the viscosity adjustment feedback. The flow stability adjustment value includes the flow recovery efficiency, the fluctuation control effect, and the stability improvement index. The titration analysis correction results include the density correction value, the bubble impact evaluation, and the corrected accuracy improvement.

[0065] See also Figure 2 , the error identification module includes:

[0066] The flow and potential offset calculation submodule obtains the titration reagent flow and electrode potential offset, calculates the flow offset and potential offset, calls the ambient temperature data, and calculates the offset ratio based on the error reference range using the formula:

[0067]

[0068] The deviation ratio is calculated and the error of endpoint determination is analyzed;

[0069] Among them, R f represents the deviation ratio, Q m represents the measured titration reagent flow rate, Q r represents the base flow rate, E m represents the measured electrode potential, E r represents the reference potential, T represents the ambient temperature, T r Represents the reference ambient temperature.

[0070] First, use a high-precision flowmeter to measure the actual flow rate of the titration reagent during the titration process. For example, if the target flow rate is set to 10.0mL / min and the actual measured value is 10.5mL / min, the flow offset is 0.5mL / min. At the same time, use a precision potentiometer to measure the actual potential of the electrode. Assuming that the reference potential is 200.0mV and the measured value is 205.0mV, the potential offset is 5.0mV. Next, collect the ambient temperature data. Assuming that the laboratory temperature is 25.0℃ and the reference temperature is 20.0℃, the temperature offset is 5.0℃. Based on the error reference range, calculate the offset ratios of each item using the formula:

[0071]

[0072] Among them, R fis the offset ratio, Q m is the measured titration reagent flow rate, Q r is the base flow rate, E m is the measured electrode potential, E r is the reference potential, T is the ambient temperature, T r is the reference ambient temperature, substitute the above values ​​and calculate:

[0073]

[0074] 0.025+2.236=2.311;

[0075] Therefore, the offset ratio R f It is approximately 2.311, which represents the combined deviation of flow, potential and temperature.

[0076] The endpoint determination error analysis submodule calls the offset ratio, analyzes the endpoint determination error, obtains the endpoint determination error offset, calculates the endpoint determination error offset ratio based on the error reference range, and obtains the endpoint error ratio;

[0077] To analyze the endpoint determination error, first determine the theoretical endpoint potential during the titration process, assuming it is 400.0 mV. Then, measure the actual endpoint potential, assuming it is 410.0 mV, and calculate the endpoint potential error to be 10.0 mV. Next, consider the effect of the offset ratio on the endpoint determination. Assuming that the offset ratio is linearly related to the endpoint error, the relationship is:

[0078] E error = k × R f ;

[0079] Among them, E error is the end point potential error, k is the proportionality coefficient, assuming k = 4.0, then:

[0080] E error =4.0×2.311=9.244mV;

[0081] The calculated endpoint potential error is approximately 9.244 mV, which is close to the actual measured value of 10.0 mV, indicating that the offset ratio has a significant effect on the endpoint determination error.

[0082] The key error source screening submodule calls the endpoint error ratio, analyzes the correlation of changes in differential error factors, screens parameters with significant deviations, identifies key error sources based on correlation analysis, and obtains error impact ranking values;

[0083] First, compare the size of each offset. The flow offset is 0.5mL / min, accounting for 5% of the reference flow; the potential offset is 5.0mV, accounting for 2.5% of the reference potential; the temperature offset is 5.0℃, and the corresponding offset ratio contribution is about 2.236, which is significantly higher than the offset ratio contribution of flow and potential. Therefore, the change in ambient temperature is the main source of error. It is recommended to strictly control the ambient temperature during the experiment to reduce the impact on the titration results.

[0084] See also Figure 3 , the temperature control module includes

[0085] The temperature change gradient calculation submodule calls the error impact ranking value, obtains the laboratory environment temperature and reagent temperature benchmark, calculates the temperature change gradient, obtains the temperature data of multiple time nodes based on the temperature sampling points during the experiment, calculates the temperature change amplitude between adjacent time nodes, and normalizes according to the set time interval to calculate the temperature change gradient using the formula:

[0086]

[0087] Calculate the temperature change gradient and make rate judgment;

[0088] Among them, G θ represents the temperature gradient, Θ j represents the temperature value at the jth time point, Θ j-1 represents the temperature value at the previous time point, s represents the total number of sampling points, Δτ represents the time interval, Θ env represents the laboratory ambient temperature, Θ std Represents the reagent temperature reference;

[0089] First, use a high-precision temperature sensor to monitor the laboratory environment temperature at different time points. Assuming that the temperature data is recorded every 10 seconds, the temperature sequence is obtained:

[0090] Θ={23.5℃,23.7℃,24.0℃,24.5℃,25.0℃};

[0091] Then, determine the reagent temperature benchmark. Assuming that the experiment requires the reagent temperature to be maintained at 22.0°C, the specific steps for calculating the temperature change gradient are as follows:

[0092] Calculate the temperature change at adjacent time points:

[0093] ΔΘ 1 =23.7-23.5=0.2℃;

[0094] ΔΘ 2 =24.0-23.7=0.3℃;

[0095] ΔΘ3 =24.5-24.0=0.5℃;

[0096] ΔΘ 4 =25.0-24.5=0.5℃;

[0097] Calculate the temperature gradient G θ The value of is calculated by normalization:

[0098]

[0099] Substitute the known data:

[0100]

[0101] G θ =0.0375+1.58=1.6175;

[0102] Determine whether the temperature change gradient exceeds the allowable range, set the maximum allowable gradient to 1.5, and calculate the value G θ =1.6175 exceeds the threshold range, so the flow rate needs to be adjusted. Finally, the temperature change gradient is obtained for subsequent rate judgment.

[0103] The flow adjustment submodule calls the temperature change gradient, sets the threshold range, determines whether the rate exceeds the threshold range, adjusts the titration reagent pump flow rate according to the set flow adjustment standard, calculates the adjusted flow offset, obtains the flow adjustment offset, and corrects the titration pump drive voltage;

[0104] Call the temperature change gradient, set the threshold range, and determine whether the rate exceeds the threshold range. According to the experimental requirements, set the threshold range to 0.2℃ / min, and check whether the temperature change rate exceeds this range. If it exceeds, adjust the titration reagent pump flow rate, such as the currently calculated G θ It is 0.35℃ / min, which exceeds the set threshold, so the flow rate needs to be adjusted, the adjusted flow offset is calculated, the flow rate is reduced by 5% for correction, and the flow adjustment offset is obtained, which is subsequently used to correct the titration pump drive voltage.

[0105] The temperature-corrected flow calculation submodule calls the flow adjustment offset, corrects the titration pump driving voltage according to the offset, calculates the temperature-corrected flow value, and obtains the temperature-corrected flow value;

[0106] First, obtain the adjusted flow offset to determine whether the current flow meets the stability requirements of reagent titration. If the offset exceeds the set range, for example, the experimental setting allows the flow offset range to be ±0.3mL / min, and the current measured offset is 0.5mL / min, then the drive voltage needs to be adjusted to increase or decrease the pump supply voltage according to the offset direction. For example, if the flow offset is a positive value, reduce the voltage, and if it is a negative value, increase the voltage. At the same time, monitor the adjusted flow and continue to adjust until the flow offset returns to the allowable range. For example, the adjusted flow offset drops to 0.2mL / min, which meets the set standard, and finally obtain the temperature-corrected flow value.

[0107] See also Figure 4 , the viscosity compensation module includes:

[0108] The viscosity offset calculation submodule calls the temperature-corrected flow value, obtains the titration reagent viscosity benchmark and real-time reagent viscosity, calculates the viscosity offset, records the reagent viscosity values ​​at multiple time points based on the measurement data during the experiment, and calculates the deviation from the benchmark viscosity using the formula:

[0109]

[0110] Calculate the viscosity offset and calculate the flow error;

[0111] Among them, Δη represents the viscosity deviation, η p represents the viscosity of the reagent measured at the pth time point, η b represents the viscosity standard of the reagent, and w represents the number of measurements;

[0112] First, monitor the actual viscosity value of the titration reagent, and use a viscometer to record the dynamic viscosity of the reagent at different time points. For example, the base viscosity of a reagent is set to 1.5 mPa·s. During the experiment, the viscosity values ​​at multiple time points are measured, which are 1.52 mPa·s, 1.55 mPa·s, 1.60 mPa·s, 1.65 mPa·s, and 1.70 mPa·s. Then calculate the viscosity offset using the formula:

[0113]

[0114] Substitute the measured data for calculation:

[0115]

[0116]

[0117] The final calculated viscosity offset is 0.104 mPa·s.

[0118] The flow error adjustment submodule calls the viscosity offset to calculate the flow error caused by the viscosity change. Based on the liquid flow characteristics, the flow impact ratio correction formula is used to calculate the offset amplitude of the viscosity change on the flow, and the viscosity change flow error is obtained.

[0119] First, analyze the flow state of the titration reagent. The flow rate of the titration reagent will be affected by the viscosity. The increase in viscosity will increase the flow resistance of the fluid, which will in turn affect the stability of the flow. If the base flow rate of the reagent is set to 5.0mL / min, and the currently measured viscosity offset is 0.104mPa·s, it is necessary to evaluate the impact of the viscosity change on the flow rate and set an experimental measurement threshold range. For example, if the viscosity changes by 0.1mPa·s and causes the flow rate to drop by 0.3mL / min, the current offset may cause a flow deviation of about 0.3-0.4mL / min. The current flow rate measured in the experiment is 4.65mL / min, which is significantly different from the base flow rate of 5.0mL / min. Therefore, the flow error caused by viscosity needs to be corrected, and the error value is input into the control system and compared with the set flow threshold to determine whether it exceeds the allowable error range. For example, the error threshold is set to

[0120] ±0.2mL / min, the current error exceeds this range, so the pulse frequency of the pump needs to be adjusted to compensate for the flow loss, and finally the viscosity change flow error is obtained for subsequent pulse frequency adjustment.

[0121] The pulse frequency correction submodule calls the viscosity change flow error, adjusts the titration pump pulse frequency, adjusts the pulse control signal according to the correction amount, dynamically compensates for the flow deviation, and obtains the viscosity-corrected flow value;

[0122] First, obtain the current pulse frequency setting value of the pump. For example, the base pulse frequency is set to 100Hz. The flow error is usually adjusted by changing the driving pulse frequency of the pump. If a decrease in flow is detected, the pulse frequency needs to be increased to compensate for the flow loss. Assuming the flow error is 0.35mL / min, the pulse frequency can be adjusted to increase by 5-7Hz based on experience. After the adjustment, the pulse frequency of the pump is set to 105Hz, and the titration reagent flow is measured again. If the flow returns to 5.0±0.1mL / min, the adjustment is complete. If the flow still deviates from the set range, continue to fine-tune the pulse frequency until the flow reaches a stable state, and finally obtain the viscosity-corrected flow value.

[0123] See also Figure 5 , the traffic shaping module includes:

[0124] The flow change gradient calculation submodule calls the viscosity-corrected flow value, obtains the standard titration flow benchmark and real-time reagent flow, calculates the instantaneous flow change gradient index, calculates the flow change amplitude at adjacent time points based on the flow measurement values ​​at differentiated time points during the experiment, and calculates the flow change gradient in combination with the time interval normalization, using the formula:

[0125]

[0126] Calculate the instantaneous flow gradient index to make fluctuation judgment;

[0127] Among them, q Represents the instantaneous flow gradient index, Q r represents the flow value at the rth time point, Q r-1 represents the flow value at the previous time point, s represents the total number of sampling points, Δξ represents the time interval, Q std Represents the standard titration flow rate reference, Q mean Represents the average value of real-time reagent flow;

[0128] First, the total number of sampling times s and the measurement interval Δξ are set, and the flow data is standardized. Then, the instantaneous flow change gradient is calculated using the following formula:

[0129]

[0130] In the calculation process, the total sampling times s = 10, the measurement time interval Δξ = 0.5s, and the flow data are collected. It is assumed that the measured reagent flow rates are Q 1 =2.1, Q 2 =2.3, ..., Q 10 = 2.5 (unit: mL / s), calculate the flow rate change at adjacent time points:

[0131]

[0132] Compute the normalized gradient:

[0133]

[0134] Calculate the mean deviation term:

[0135]

[0136] Finally, the instantaneous flow change gradient is calculated:

[0137] Ψ q =0.28+0.07=0.35;

[0138] Set judgment threshold Ψ th =0.3, due to Ψ q=0.35>0.3, the flow fluctuation is judged to be abnormal and needs to be adjusted, and finally the flow instantaneous change gradient index is obtained.

[0139] The flow recovery offset calculation submodule calls the flow instantaneous change gradient index to determine whether the fluctuation exceeds the range. Based on the flow change threshold set in the experiment, it determines whether the instantaneous flow fluctuation exceeds the set range and calculates the flow recovery offset value.

[0140] First, set the allowable flow fluctuation range, for example, ±0.2mL / min. The measured data shows that the maximum fluctuation is 0.8mL / min, which exceeds the set range. Therefore, calculate the flow recovery offset value, determine the corrected target value, and calculate the offset. For example, the base flow is 5.0mL / min, and the measured average value is 5.06mL / min. The offset is calculated as:

[0141] Q offset =5.06-5.0=0.06mL / min;

[0142] Since the offset is small, no adjustment is required for the time being. However, if the offset exceeds 0.2 mL / min, further correction is required to finally obtain the flow recovery offset value.

[0143] The flow stabilization correction submodule calls the flow recovery offset value, corrects the titration flow, adjusts the output stability of the titration pump, adjusts the pulse drive signal, dynamically compensates for flow fluctuations, and obtains the flow stabilization adjustment value;

[0144] First, adjust the pulse drive signal according to the offset. For example, if the current pump pulse frequency is 100Hz and the flow fluctuation is detected to exceed 0.2mL / min, adjust the pulse frequency by ±5Hz. If the flow rate is stabilized at 5.0mL / min±0.1mL / min after adjustment, the adjustment is completed. If there is still a deviation, continue to fine-tune the pulse frequency to finally obtain the stable flow adjustment value.

[0145] See also Figure 6 , the bubble correction module includes:

[0146] The bubble interference determination submodule obtains the reagent density reference and real-time reagent density based on the flow stability adjustment value, analyzes whether there is abnormal fluctuation in reagent density by comparing the difference between the real-time reagent density and the reference density, calculates the mean of the density change amplitude, sets the density abnormality threshold, and determines whether there is bubble interference. If the density change amplitude exceeds the set threshold, it is marked that there is bubble interference in the current reagent flow, otherwise it is determined that the reagent flow is normal and the bubble interference determination result is obtained;

[0147] First, the flow stability adjustment value is called, the change data of the reagent flow in recent times is extracted, and the reagent density benchmark is combined for analysis, the density measurement time window is set, the reagent density data is periodically collected, and the reagent density at different time points is normalized, and the deviation value from the reagent density benchmark is calculated. The density offset is modeled in time series, and its trend characteristics are detected to determine whether the density has a sudden change. If the density change exceeds the set normal offset threshold, there may be bubble interference. On this basis, the bubble influence model is further combined with the flow change situation, and the bubble interference signal is identified by analyzing the relationship between the density offset rate and the instantaneous change rate of the flow in a short time. If the density suddenly changes in a short time and the flow fluctuation rate reaches the set threshold, it is judged as bubble interference, and the interference time point is recorded. The instantaneous flow error is calculated, and the error correction requirements are analyzed. If the error is within the allowable range, no correction is performed. If the error exceeds the set error tolerance, the titration pump drive voltage needs to be adjusted. The adjustment range is determined by the instantaneous flow error, density offset, and historical error correction data, and finally the titration analysis correction result is obtained.

[0148] The density offset calculation submodule calls the flow stability adjustment value, obtains the reagent density benchmark and real-time reagent density, selects density data within a period of time, compares the measured real-time reagent density with the reagent density benchmark, and uses the absolute deviation mean calculation method to calculate the density offset using the formula:

[0149]

[0150] Calculate the density offset and make error correction;

[0151] Among them, Θ D represents the density offset, D i represents the reagent density measured for the i-th time, D ref represents the reagent density reference, and N represents the number of measurements;

[0152] In the process of density offset calculation, it is necessary to determine the measurement period T m , collect reagent density N times during the measurement cycle and record the measurement data D i , use the mean calculation formula to determine the reagent density benchmark D ref , calculate the absolute deviation between each measured value and the reference value, and take the average value as the density offset:

[0153]

[0154] Among them, Θ D represents the density offset, D i represents the reagent density measured for the i-th time, D ref represents the reagent density standard, and N represents the number of measurements.

[0155] Calculation steps:

[0156] Set the measurement period T m :Choose a suitable measurement period T m , such as taking 10 samples within 5 seconds, with each measurement taking place at an interval of 0.5 seconds;

[0157] Collect reagent density data: Perform N = 10 density measurements within the set period. Assume that the measured density values ​​are D 1 =1.03,D 2 =1.05,…,D 10 =1.08 (unit: g / cm3);

[0158] Calculate the reagent density benchmark: Take the average of all measured values ​​as the density benchmark:

[0159]

[0160] Calculate the density offset: Use the above formula:

[0161]

[0162] Determine whether the density offset exceeds the threshold: set the offset threshold Θ th =0.02, then Θ D =0.018 does not exceed the threshold, and the density deviation is considered to be within the normal range.

[0163] The flow error correction submodule determines the source of the flow error based on the density offset and bubble interference judgment results. If there is bubble interference, it calculates the instantaneous flow error caused by the bubble. If there is no bubble interference, it calculates the instantaneous flow error caused by the density change, corrects the titration pump drive voltage, adjusts the titration flow, and obtains the titration analysis correction result;

[0164] First, the error correction threshold is set. If the instantaneous flow error is lower than the set threshold, no adjustment is made. If it exceeds the threshold, the correction amplitude is determined according to the error correction model. By adjusting the increase or decrease of the titration pump driving voltage, the titration flow rate gradually returns to a stable state. During the adjustment process, the error monitoring window is set to periodically detect the error correction status. If the error still exceeds the allowable range after adjustment, the driving voltage continues to be adjusted until the error converges to the set range.

[0165] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent titration analysis system based on artificial intelligence, characterized in that: The system comprises: The error identification module obtains flow rate, electrode potential offset, endpoint error, temperature, calculates offset, screens significant parameters, analyzes error correlation, screens key error sources, and obtains error impact ranking values; The temperature adjustment module calls the error impact ranking value, obtains the environment and reagent temperature, calculates the gradient, determines the rate threshold, adjusts the pump flow rate, corrects the drive voltage, and obtains the flow value after temperature correction; The viscosity compensation module calls the temperature-corrected flow value, obtains the reagent viscosity reference and real-time viscosity, calculates the offset and error, adjusts the pump pulse frequency, and obtains the viscosity-corrected flow value; The flow adjustment module calls the viscosity-corrected flow value, obtains the flow reference and real-time value, calculates the change gradient, determines the over-limit, calculates the recovery offset value, corrects the titration flow, and obtains the flow stability adjustment value; The bubble correction module calls the flow stability adjustment value, obtains the density reference and real-time density, calculates the offset, determines the bubble interference, adjusts the driving voltage, and obtains the analysis correction result.

2. The intelligent titration analysis system based on artificial intelligence according to claim 1, characterized in that: The error impact ranking value includes the ranking of significant offset parameters, the ranking of error sources, and the correlation strength index; the temperature-corrected flow value includes the flow rate adjustment value, the corrected driving voltage, and the adjustment effect evaluation; the viscosity-corrected flow value includes the adjusted pulse frequency, the correction effect evaluation, and the viscosity adjustment feedback; the flow stability adjustment value includes the flow recovery efficiency, the fluctuation control effect, and the stability improvement index; the titration analysis correction result includes the density correction value, the bubble impact evaluation, and the corrected accuracy improvement.

3. The intelligent titration analysis system based on artificial intelligence according to claim 1, characterized in that: The error identification module comprises: The flow and potential offset calculation submodule obtains the titration reagent flow and electrode potential offset, calculates the flow offset and potential offset, calls the ambient temperature data, and calculates the offset ratio based on the error reference range using the formula: The deviation ratio is calculated and the error of endpoint determination is analyzed; Among them, R f represents the deviation ratio, Q m represents the measured titration reagent flow rate, Q r represents the base flow rate, E m represents the measured electrode potential, E r represents the reference potential, T represents the ambient temperature, T r Represents the reference ambient temperature; The endpoint determination error analysis submodule calls the offset ratio, analyzes the endpoint determination error, obtains the endpoint determination error offset, calculates the endpoint determination error offset ratio based on the error reference range, and obtains the endpoint error ratio; The key error source screening submodule calls the endpoint error ratio, analyzes the change correlation of the differential error factor, screens the significant deviation parameters, identifies the key error sources based on the correlation analysis, and obtains the error impact ranking value.

4. The intelligent titration analysis system based on artificial intelligence according to claim 1 is characterized in that: The temperature adjustment module includes The temperature change gradient calculation submodule calls the error impact ranking value, obtains the laboratory environment temperature and reagent temperature benchmark, calculates the temperature change gradient, obtains the temperature data of multiple time nodes based on the temperature sampling points during the experiment, calculates the temperature change amplitude between adjacent time nodes, and normalizes according to the set time interval to calculate the temperature change gradient using the formula: Calculate the temperature gradient and make rate judgment; Among them, G θ represents the temperature gradient, Θ j represents the temperature value at the jth time point, Θ j-1 represents the temperature value at the previous time point, s represents the total number of sampling points, Δτ represents the time interval, Θ env represents the laboratory ambient temperature, Θ std Represents the reagent temperature reference; The flow adjustment submodule calls the temperature change gradient, sets a threshold range, determines whether the rate exceeds the threshold range, adjusts the flow rate of the titration reagent pump according to the set flow adjustment standard, calculates the adjusted flow offset, obtains the flow adjustment offset, and corrects the titration pump drive voltage; The temperature-corrected flow calculation submodule calls the flow adjustment offset, corrects the titration pump driving voltage according to the offset, calculates the temperature-corrected flow value, and obtains the temperature-corrected flow value.

5. The intelligent titration analysis system based on artificial intelligence according to claim 1 is characterized in that: The viscosity compensation module comprises: The viscosity offset calculation submodule calls the temperature-corrected flow value, obtains the titration reagent viscosity reference and real-time reagent viscosity, calculates the viscosity offset, records the reagent viscosity values ​​at multiple time points based on the measurement data during the experiment, and calculates the deviation from the reference viscosity using the formula: Calculate the viscosity offset and calculate the flow error; Among them, Δη represents the viscosity deviation, η p represents the viscosity of the reagent measured at the pth time point, η b represents the viscosity standard of the reagent, and w represents the number of measurements; The flow error adjustment submodule calls the viscosity offset, calculates the flow error caused by the viscosity change, calculates the offset amplitude of the viscosity change on the flow based on the liquid flow characteristics, and obtains the viscosity change flow error through the flow influence ratio correction formula; The pulse frequency correction submodule calls the viscosity change flow error, adjusts the titration pump pulse frequency, adjusts the pulse control signal according to the correction amount, dynamically compensates for the flow deviation, and obtains the viscosity-corrected flow value.

6. The intelligent titration analysis system based on artificial intelligence according to claim 1 is characterized in that: The flow adjustment module comprises: The flow change gradient calculation submodule calls the viscosity-corrected flow value, obtains the standard titration flow benchmark and real-time reagent flow, calculates the flow instantaneous change gradient index, calculates the flow change amplitude at adjacent time points based on the flow measurement values ​​at differentiated time points during the experiment, and calculates the flow change gradient in combination with the time interval normalization, using the formula: Calculate the instantaneous flow gradient index to make fluctuation judgment; Among them, q Represents the instantaneous flow gradient index, Q r represents the flow value at the rth time point, Q r-1 represents the flow value at the previous time point, s represents the total number of sampling points, Δξ represents the time interval, Q std Represents the standard titration flow rate reference, Q mean Represents the average value of real-time reagent flow; The flow recovery offset calculation submodule calls the flow instantaneous change gradient index to determine whether the fluctuation exceeds the range, and determines whether the instantaneous flow fluctuation exceeds the set range according to the flow change threshold set in the experiment, and calculates the flow recovery offset value; The flow stabilization correction submodule calls the flow recovery offset value, corrects the titration flow, adjusts the output stability of the titration pump, adjusts the pulse drive signal, dynamically compensates for flow fluctuations, and obtains a flow stabilization adjustment value.

7. The intelligent titration analysis system based on artificial intelligence according to claim 1 is characterized in that: The bubble correction module comprises: The bubble interference determination submodule obtains the reagent density reference and the real-time reagent density based on the flow stability adjustment value, analyzes whether there is abnormal fluctuation in the reagent density by comparing the difference between the real-time reagent density and the reference density, calculates the mean of the density change amplitude, sets the density abnormality threshold, and determines whether there is bubble interference. If the density change amplitude exceeds the set threshold, it is marked that there is bubble interference in the current reagent flow, otherwise it is determined that the reagent flow is normal, and the bubble interference determination result is obtained; The density offset calculation submodule calls the flow stability adjustment value, obtains the reagent density benchmark and the real-time reagent density, selects density data within a period of time, compares the measured real-time reagent density with the reagent density benchmark, and calculates the density offset using the absolute deviation mean calculation method, using the formula: Calculate the density offset and make error correction; Among them, Θ D represents the density offset, D i represents the reagent density measured for the i-th time, D ref represents the reagent density reference, and N represents the number of measurements; The flow error correction submodule determines the source of the flow error based on the density offset and the bubble interference judgment result. If there is bubble interference, the instantaneous flow error caused by the bubble is calculated. If there is no bubble interference, the instantaneous flow error caused by the density change is calculated, the titration pump driving voltage is corrected, the titration flow is adjusted, and the titration analysis correction result is obtained.

Images