PH value online detector based on artificial intelligence

Abstract

The invention belongs to the technical field of PH detection, particularly relates to a sauce PH detection technology, and particularly discloses an online PH value detector based on artificial intelligence, sauce distribution uniformity monitoring is synchronously achieved in the sauce pH detection process, vibration control is implemented based on sauce distribution uniformity feedback, it is ensured that detection is in a full mixing state, and the detection accuracy is improved. Meanwhile, through linkage analysis of pH sudden change recognition and the uniformity state, process disturbance and probe faults can be effectively distinguished under the continuous monitoring condition, false alarm and invalid maintenance are avoided, the intelligent level of system diagnosis is improved, and stable and reliable online pH detection is achieved; and meanwhile, liquid level offset monitoring is introduced when vibration control is implemented based on sauce uniformity feedback, and vibration adjustment is performed in combination with a dynamic coupling relationship between a uniformity index and liquid level offset, so that stirring is ensured to operate within a structural safety range, excessive vibration is avoided, and the system stability and operation safety are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of PH detection, particularly relates to sauce PH detection technology, and specifically discloses a PH value on-line detector based on artificial intelligence. BACKGROUND

[0002] With the increasing attention of consumers to the quality of condiments, as a widely used compound condiment, the quality stability of chili sauce is directly related to the taste experience and food safety of consumers. In the production process of chili sauce, pH value is one of the important parameters for measuring the acid-base balance, preservative performance, fermentation degree and microbial safety of products. Therefore, realizing on-line real-time detection of pH value has become an indispensable technical link in modern chili sauce production lines.

[0003] However, due to the high solid content and easy stratification characteristics of chili sauce, when the stratification phenomenon occurs after the sauce is stationary during the monitoring process of the traditional pH probe, the probe only contacts the upper liquid phase, and cannot accurately reflect the real pH value of the whole, resulting in a large deviation of the measurement result.

[0004] In addition, in the actual production environment, due to factors such as uneven mixing of sauce and abnormal operation of equipment, the output signal of the pH probe may suddenly change, resulting in distorted data collection. In this case, due to the lack of synchronous sauce distribution detection means, it is difficult to accurately identify whether these fluctuations are caused by process abnormalities, and normal fluctuations may be misjudged as equipment failure, which on the one hand triggers unnecessary probe calibration or downtime maintenance and other invalid fault handling operations, increasing the operation and maintenance burden; on the other hand, it may lead to false shielding of effective data, affecting the integrity and accuracy of pH value continuous monitoring.

[0005] In order to more accurately identify the output fluctuation of pH value, some schemes propose to take out the pH probe for manual calibration or cleaning, but this will cause interruption of detection, and cannot realize real-time on-line detection. Such interruption not only affects the continuity of production, but also reduces the response speed and reliability of the system. In addition, frequent manual intervention increases the operation complexity and maintenance cost, which is not conducive to the efficient operation of automatic production. SUMMARY

[0006] In view of this, the present application aims to provide a PH value on-line detector based on artificial intelligence, which synchronously detects the uniformity of sauce distribution during the detection of sauce PH, and constructs a dynamic detection compensation based on distribution state recognition, effectively solving the problems existing in the prior art.

[0007] The purpose of the present application can be achieved by the following technical scheme: a PH value on-line detector based on artificial intelligence, comprising the following modules: a uniformity detection module for monitoring the turbidity value in the partition of the tank by using symmetrically arranged optical turbidity sensors, and analyzing the uniformity of sauce distribution by calculating the turbidity difference in the region.

[0008] The vibration control module is used to trigger gradient vibration control when uneven distribution of the sauce is identified, and to monitor the sauce distribution uniformity and sauce liquid level deviation synchronously during the vibration process, and then to adjust the vibration according to the dynamic combination of the sauce distribution uniformity and the liquid level deviation.

[0009] The data acquisition module is used to acquire the real-time pH value of the sauce by using the immersion pH probe, and to identify the pH mutation signal of the sauce by comparing the real-time pH value of the sauce with the historical sliding average value.

[0010] The dynamic compensation module is used to mark the detection state of the pH probe as a shielding period or a failure period according to the identified pH mutation signal of the sauce, to pause the pH value recording when the detection state of the pH probe is the shielding period, to start the sliding filter to output the effective pH value when the pH probe is in the effective detection state, and then to compensate for the missing pH data of the shielding period according to the effective pH value, and to trigger a failure warning when the detection state of the pH probe is the failure period.

[0011] Compared with the prior art, the present application has the following beneficial effects: 1. The present application synchronously realizes the monitoring of the sauce distribution uniformity during the sauce pH detection process, and implements vibration control based on the feedback of the sauce distribution uniformity, ensuring that the detection is in a fully mixed state, significantly improving the pH measurement accuracy, and through the linkage analysis of the pH mutation identification and the uniformity state, the process disturbance and the probe failure can be effectively distinguished under continuous monitoring conditions, avoiding false alarms and invalid maintenance, improving the intelligent level of system diagnosis, and realizing stable and reliable online pH detection.

[0012] 2. The present application introduces liquid level deviation monitoring when implementing vibration control based on the feedback of the sauce uniformity, and adjusts the vibration in combination with the dynamic coupling relationship between the uniformity index and the liquid level deviation, ensuring that the vibration operates within the structural safety range, avoiding excessive vibration to cause equipment damage or severe liquid level fluctuation, and improving the system stability and operation safety. BRIEF DESCRIPTION OF DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creating any creative labor.

[0014] Figure 1 It is a schematic diagram of the module composition in the present application.

[0015] Figure 2 It is a logic schematic diagram of vibration adjustment according to the dynamic combination of the sauce distribution uniformity and the liquid level deviation in the present application.

[0016] Figure 3 The flow chart is used for dynamic compensation of PH data when the detection state of the PH probe is a shielding period in the application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the protection scope of the application.

[0018] Referring to Figure 1 As shown in the figure, the application provides a PH value online detector based on artificial intelligence, which comprises a uniformity detection module, a vibration control module, a data acquisition module and a dynamic compensation module, wherein the uniformity detection module is connected with the vibration control module, and the data acquisition module is connected with the dynamic compensation module.

[0019] The uniformity detection module is used for monitoring the partition turbidity value in the tank by using the symmetrically arranged optical turbidity sensors, and analyzing the sauce uniformity by calculating the regional turbidity difference.

[0020] In an optional implementation of the above scheme, the process of monitoring the partition turbidity value in the tank by using the symmetrically arranged optical turbidity sensors is as follows: the optical turbidity sensors are symmetrically installed in the tank body axial isosurface plane, and the installation positions of the sensors divide the internal space of the tank body into a plurality of monitoring regions.

[0021] It should be noted that the optical turbidity sensor detects the turbidity of the sauce in the tank by using the sensor to emit a beam of light through the liquid, and the suspended particles in the liquid will cause the light to scatter. The intensity of the scattered light is proportional to the concentration of the particles in the liquid, so as to reflect the turbidity of the liquid.

[0022] Each optical turbidity sensor is used for detecting the liquid phase turbidity of the sauce in the region where the optical turbidity sensor is located, so as to obtain the turbidity values of different regions in the tank body.

[0023] Turbidity is a macroscopic manifestation of the concentration of suspended particles in a liquid, and its change directly reflects the spatial distribution of the material. The uneven distribution of particles in a high-viscosity mixture such as sauce often manifests as local turbidity difference. By detecting the turbidity of different regions, it can be indirectly judged whether the material is mixed uniformly.

[0024] It should be noted that the axial direction of the tank refers to the vertical direction along the center of the tank, and the axial isometric plane refers to a horizontal section parallel to the bottom surface at a fixed height inside the tank. The optical turbidity sensors are symmetrically installed on the axial isometric plane because the solid particles in the sauce are prone to settle vertically due to gravity, and a concentration gradient is usually formed at the bottom or a local area of the tank. The symmetrical arrangement of the sensors in the same isometric plane can identify whether eccentric settling occurs laterally, and the fluid motion in the same horizontal section has similar dynamics during subsequent vibration control. The axial isometric plane can effectively capture the turbidity distribution difference in the radial direction to determine whether there is local precipitation or segregation. Therefore, the symmetrical arrangement of multiple optical turbidity sensors in the axial isometric plane of the tank can ensure representative sampling of the liquid in multiple directions and areas of the tank. By symmetrical arrangement, the turbidity difference between the symmetrical areas can be compared to determine whether there is precipitation or other uneven distribution of the sauce, thereby providing a trigger basis for subsequent vibration control.

[0025] In a further optional implementation of the above scheme, the analysis of the uniformity of the sauce distribution by calculating the turbidity difference of the regions includes the following content: according to the symmetrical installation positions of the multiple optical turbidity sensors on the axial isometric plane of the tank, the liquid space in the tank is divided into a plurality of symmetrically distributed monitoring regions.

[0026] At any sampling time, the turbidity values of the regions measured by each sensor are synchronously acquired, and the turbidity values of each symmetrical region are compared pairwise to calculate the turbidity difference.

[0027] The standard deviation of the turbidity difference between all symmetrical regions is calculated as an evaluation index of the uniformity of the sauce.

[0028] It should be explained that in the static state, the distribution of the material in the tank tends to be balanced under the action of gravity and diffusion, and the concentration field has spatial consistency in the geometrically symmetrical region. Since the sensors are arranged at axial symmetrical positions, the measured turbidity values of the regions should be basically consistent. When significant turbidity difference occurs in actual detection, it indicates that the symmetry of the system is broken, and there may be non-uniform phenomena such as local precipitation, stratification, or insufficient mixing. By calculating the standard deviation of the turbidity difference between all symmetrical regions, the overall distribution dispersion can be effectively quantified. This index as an evaluation parameter of the uniformity of the sauce has good sensitivity and repeatability, and can objectively reflect the mixing uniformity level of the material in the tank.

[0029] The vibration control module is used to trigger gradient vibration control when the uneven distribution of the sauce is identified, and to synchronously monitor the uniformity of the sauce distribution and the sauce liquid level deviation during the vibration process, and then to adjust the vibration according to the dynamic combination of the uniformity of the sauce distribution and the liquid level deviation.

[0030] Optionally, when the uneven distribution of the sauce is identified, the gradient vibration control is triggered as follows: The sauce uniformity evaluation index is compared with the preset uniformity threshold value, and the gradient vibration control is started when the sauce uniformity evaluation index exceeds the uniformity threshold value, as follows: The controllable vibration frequency range of the probe vibration rod is divided into several discrete vibration frequency levels according to the set frequency step, forming a vibration frequency sequence with increasing characteristics.

[0031] The uniformity threshold value reflects the maximum acceptable non-uniformity of the sauce in spatial distribution. Specifically, for samples with known uniform sauce distribution, multiple-point turbidity measurement is performed, and the standard deviation of the turbidity difference in the symmetric region is taken as the uniformity threshold value.

[0032] The explanation of the above operation is that when the sauce distribution is uneven, external mechanical vibration needs to be applied to generate shear force and fluid disturbance inside the liquid, promote particle resuspension and interphase mixing, and achieve homogenization. However, the sauce usually has high viscosity, multi-phase mixing, and easy sedimentation, and a single fixed frequency vibration may not effectively break the local precipitation. If high-frequency vibration is directly applied, it may cause local shear thinning of high-viscosity chili sauce, oil phase separation, or gas bubble entrapment, which may damage the system stability, even cause liquid surface to fluctuate violently or structural resonance, so high-intensity vibration is not suitable. Therefore, the controllable vibration frequency range is divided into several discrete levels according to the set frequency step, realizing gradual excitation from low frequency to high frequency, and ensuring that the vibration intensity increases smoothly.

[0033] As a further explanation of the above explanation, the frequency step should not be too large when dividing the increasing frequency sequence, otherwise the vibration energy jump between adjacent levels will be significant, reducing the control accuracy and affecting the homogenization efficiency.

[0034] The frequency step can be calibrated based on vibration response characteristics experiments. Specifically, step frequency tests are performed on the sauce under actual or simulated conditions, and the mixing uniformity improvement rate and liquid surface displacement response under different frequency increments are observed. The smallest effective frequency change that can cause significant mixing improvement but not excessive disturbance is selected as the reference step.

[0035] The control probe vibration rod starts low-intensity vibration in the tank from the initial vibration frequency in the controllable vibration frequency range, and automatically increases the vibration frequency level every set time interval until the maximum vibration frequency in the controllable vibration frequency range is reached.

[0036] It is also necessary to explain that a certain time interval needs to be maintained after each vibration frequency level before increasing to the next level, the purpose of which is to ensure that the vibration energy is fully transmitted and responded in the sauce system, so that the material completes effective flow reconstruction and particle redistribution under this excitation intensity, and avoids lagging in the mixing process due to too fast frequency switching. The time period from the start of vibration to the significant decrease in turbidity can be measured by step response experiment as the maintenance time interval.

[0037] Further optionally, the sauce liquid level offset amount refers to the monitoring process as follows: a laser displacement sensor is arranged on the side wall of the sauce tank, the optical axis of which is aligned with the center area of the liquid level in the tank, for non-contact real-time measurement of the liquid level during vibration.

[0038] The distance signal output by the laser displacement sensor is continuously collected during vibration, and compared with the initial liquid level measured before starting vibration, to obtain the real-time offset amount of the sauce liquid level relative to the initial liquid level.

[0039] It should be noted that the axial force generated by mechanical excitation and the fluid disturbance caused by bubble rupture will cause the dynamic offset of the sauce liquid level relative to the initial static position. If the offset amplitude is too large, it may cause liquid splashing, material loss or pollution of the surrounding environment, or even abnormal pressure fluctuation in the tank. Therefore, by monitoring the liquid level offset amount in real time through the laser displacement sensor, the fluid stability of the vibration process can be quantitatively evaluated to identify whether there is an overrun disturbance, so as to realize safe closed-loop control of the vibration intensity, avoid blindly increasing the vibration intensity and ignoring the safety risks of structure and process, and ensure stable and reliable operation of the system.

[0040] Continuing further optionally, referring to Figure 2 According to the dynamic combination of sauce distribution uniformity and liquid level offset amount, the vibration adjustment is as follows: during the vibration process, the real-time monitored sauce uniformity evaluation index is compared with the uniformity threshold value, and the sauce liquid level offset amount is compared with the safe offset amount.

[0041] (1) When the sauce uniformity evaluation index is still higher than the uniformity threshold value at a certain time and the sauce liquid level offset amount meets the safe offset amount, the vibration operation is continued according to the established gradient frequency increasing strategy.

[0042] The safe offset amount in the above reflects the maximum fluctuation boundary of the sauce liquid level during vibration. For example, the maximum offset value that does not cause overflow, splashing or probe exposure can be set as the safe offset amount based on the static liquid level in the tank and combined with the liquid level fluctuation test data under actual working conditions.

[0043] Another example, it can be set according to industry standards or recommended values of equipment manufacturers.

[0044] It should be noted that the sauce uniformity evaluation index is represented by the standard deviation of the turbidity difference of the symmetric regions, and the numerical value directly reflects the discrete degree of spatial distribution. The larger the standard deviation, the more significant the difference in particle concentration between each monitoring region, and the worse the mixing uniformity.

[0045] Applied to case (1), it is characterized that when the sauce uniformity evaluation index is still higher than the preset uniformity threshold, it can be determined that the sauce is in an insufficient mixing state, and when the sauce liquid level deviation meets the safe deviation, it means that the sauce liquid level deviation is less than or equal to the safe deviation, which represents that there is no overrun disturbance, and the vibration can continue to be implemented.

[0046] (2) When the sauce uniformity evaluation index at a certain time falls below the uniformity threshold and the sauce liquid level deviation meets the safe deviation, switch to the micro-pulse vibration mode.

[0047] Applied to case (2), it is characterized that the material has been fully mixed and there is no risk of structural disturbance, and by switching to the micro-pulse vibration mode, local settlement is inhibited and short-term stability of the system is maintained.

[0048] (3) When the sauce uniformity evaluation index at a certain time falls below the uniformity threshold and the sauce liquid level deviation is higher than the safe deviation, maintain the micro-pulse vibration mode.

[0049] Applied to case (3), it is characterized that when the sauce uniformity evaluation index has fallen below the threshold, it means that the material has reached the target mixing uniformity, and there is no need for further implementation of intensive vibration; at the same time, the liquid level deviation exceeds the safe limit, indicating that the current fluid disturbance is strong and the dynamic load of the tank body is large, so it is not appropriate to continue to apply conventional or high-intensity vibration to avoid intensifying the liquid level fluctuation, causing splashing or structural resonance risk. Under this double constraint, the system does not need to increase the vibration intensity nor completely stop the vibration, so the micro-pulse vibration mode is adopted, which can maintain the uniform state that has been achieved while providing moderate disturbance to inhibit local settlement, with low energy input and small mechanical stress, effectively balancing mixing stability and equipment operation safety.

[0050] (4) When the sauce uniformity evaluation index at a certain time is still higher than the uniformity threshold and the liquid level deviation is higher than the safe deviation, the vibration frequency level is gradually reduced from the preset vibration frequency sequence based on the current vibration frequency until the liquid level deviation falls to the safe deviation.

[0051] Enter a execution duration at the target frequency level where the liquid level falls to the safe deviation to monitor the sauce uniformity.

[0052] If the sauce uniformity evaluation index falls below the uniformity threshold within the execution duration, maintain the micro-pulse vibration mode.

[0053] If not reduced to the homogeneity threshold, the frequency adjustment step is reduced to a small increment, and a fine frequency is obtained based on the target frequency level. Then, the liquid level deviation and homogeneity change are continuously monitored at the frequency, until the dual goals of homogeneity and liquid level stability are achieved, and the fine pulse vibration mode is finally maintained.

[0054] In the example implementation of the above case, the reduction of the frequency adjustment step can be specifically configured as half of the original set step, that is, a half-step increment is used for fine adjustment.

[0055] Applied to case (4), which represents a contradictory state that the system is in a state of insufficient mixing but the vibration intensity has exceeded the limit, that is, the current vibration energy is not enough to achieve homogenization, but has caused excessive deviation of the liquid level, which poses a risk to the safety of the equipment and the stability of the process. At this time, if the frequency continues to increase, it will intensify the disturbance, which may lead to splashing, structural fatigue or sensor malfunction; while directly reducing and stopping cannot complete the mixing task. Therefore, the system starts the vibration degradation mechanism with safety priority, and gradually reduces the vibration level based on the current frequency until the liquid level deviation returns to the safe range, ensuring the controllability of the fluid dynamics.

[0056] After the liquid level is stabilized, a preset execution time window is entered at the target frequency, and the homogeneity evolution trend is continuously monitored: if the homogeneity index falls below the threshold within this period, it means that the current frequency is low but still has effective mixing effect, and the system enters the fine pulse vibration mode to maintain the homogeneity state; if it does not meet the standard, it means that the mixing efficiency at this frequency is insufficient, but the frequency cannot be increased significantly due to the safety of the liquid level, so the system enters the fine adjustment mode and gradually increases the frequency by reducing the frequency step, to explore the optimal vibration intensity within the safe boundary.

[0057] This strategy realizes adaptive optimization control under the constraint of structural safety, avoiding high-risk vibration and gradually approaching the mixing target under the limited condition, finally achieving the dual closed loop of homogeneity and liquid level stability, ensuring the reliability of the material state before pH detection.

[0058] The data acquisition module is used to acquire the real-time sauce PH value by using the immersion PH probe, and to identify the sauce PH mutation signal by comparing the real-time sauce PH value with the historical sliding average.

[0059] As a preferred implementation of the above scheme, the sauce PH mutation signal is identified as follows: the PH probe continuously acquires and records the sauce pH value sequence within a set time window.

[0060] The real-time pH value at the current time is compared with the sliding average of the historical pH value within the time window, and the instantaneous deviation between the two is calculated.

[0061] It is to be noted that the sliding average processing of the historical pH values within the set time window can effectively suppress random noise and short-time disturbance, and reflect the reference level of the pH values. The average value represents the expected stable pH level under the current process state.

[0062] Further, the length of the set time window needs to ensure that a sufficient number of pH sampling points are included to support the statistical stability of the sliding average. The window length can be determined according to the required number of sampling points and the sampling frequency (i.e., the detection interval) of the pH probe.

[0063] The deviation amount is compared with the configured pH mutation judgment limit value. If the deviation amount between the real-time pH value at the current time and the sliding average value exceeds the pH mutation judgment limit value, it is determined that a sauce pH mutation signal is detected.

[0064] The pH mutation judgment limit value in the above operation reflects the recognition threshold of the system for abnormal changes in the pH value, and is used to distinguish between normal process fluctuations and significant mutation events. The limit value can be determined by an experimental calibration method: under the condition that the sauce is uniformly distributed and the process is stable, collect a sequence of pH values in multiple consecutive time windows, calculate the deviation amount between the instantaneous pH value and the sliding average value in each window according to the same algorithm, and statistically analyze the distribution characteristics of the obtained deviation amount. The average value or the average value plus several times the standard deviation is used as the pH mutation judgment limit value to ensure that the threshold has good representativeness and noise resistance, thereby improving the accuracy and reliability of mutation recognition.

[0065] The present application does not use the difference between adjacent pH values as the basis for judgment, but rather determines the deviation between the current real-time pH value and the historical sliding average value. This is because the adjacent pH difference method is extremely sensitive to measurement noise, which can easily misjudge random fluctuations as real mutations, resulting in an increase in false positive rate and difficulty in distinguishing between slow drift and instantaneous abnormalities. In contrast, the sliding average method effectively suppresses high-frequency noise interference through smoothing of local historical data, while retaining the response capability to significant step changes. This method achieves high robustness in monitoring the dynamic changes of the pH process, ensuring detection sensitivity while improving stability and reliability, and has good real-time performance and engineering realizability.

[0066] Referring to Figure 3 As shown in the figure, the dynamic compensation module is used to mark the detection state of the pH probe as a shielding period or a failure period according to the identified sauce PH mutation signal. When the detection state of the pH probe is a shielding period, the pH value recording is suspended, and the sliding filter is started to output the effective pH value when the pH probe is in an effective detection state. Then, the effective pH value is used to compensate for the missing PH data in the shielding period. When the detection state of the pH probe is a failure period, a failure warning is triggered.

[0067] In the manner capable of being realized in the above-mentioned solution one, the detection state of the PH probe is marked as the shielding period or the failure period according to the identified sauce PH mutation signal, and the sauce uniformity evaluation index within a time window before the time is analyzed back when the sauce PH mutation signal is identified.

[0068] If the sauce uniformity evaluation index has been below the uniformity threshold value within the time window, it is indicated that the pH mutation is not caused by the uneven mixing, and it can be determined that the mutation is caused by the abnormal response of the probe, and at this time, the detection state of the current pH probe is marked as the failure period.

[0069] On the contrary, if the sauce uniformity evaluation index is higher than the uniformity threshold value within the time window, it is preliminarily judged that the pH mutation may be caused by the non-homogenization of the sauce ingredients, and further enters the tracking stage, and whether the pH mutation signal still exists within the time window in which the sauce uniformity is restored to the threshold value or below is continuously tracked in the subsequent monitoring process, if the pH mutation signal disappears and does not appear again, it is determined that the previous mutation is the normal dynamic response in the mixing process, and at this time, the detection state of the current pH probe is marked as the shielding period, otherwise, the pH mutation signal still continuously appears in the stable state in which the sauce is uniform, and it is determined that the pH probe drifts, is contaminated or the system abnormally reacts, and is marked as the failure period.

[0070] The application backtracks the material uniformity state before the pH mutation occurs when the pH mutation is identified, judges whether the mutation is in a reasonable dynamic process window, if the mutation occurs in the uniform state, it is more likely to be abnormal, if the mutation occurs in the non-uniform state, the tracking stage is set, and whether the pH is restored to be stable after the material is uniform is observed, so as to verify whether the mutation is recoverable, further distinguish the temporary disturbance and the persistent failure, the state association and the dynamic verification are combined, the accuracy and the diagnosis reliability of the pH mutation discrimination are significantly improved, and the fine distinction between the temporary disturbance and the real failure is realized.

[0071] It should be pointed out that the effective detection state mentioned above is the period in which the sauce uniformity evaluation index is below the uniformity threshold value and the pH mutation signal is not detected.

[0072] In the manner capable of being realized in the above-mentioned solution one, the sliding filter output effective PH value is started when the PH probe is in the effective detection state, and then the PH data missing in the shielding period is compensated according to the effective PH value as follows: when the PH probe is in the effective detection state, the duration of the current effective detection state is recorded, and the effective acquisition window is intercepted at the end of the effective detection maintenance period, and then the latest PH data in the effective acquisition window is collected.

[0073] It needs to be understood that the effective acquisition window is intercepted at the end of the effective detection maintenance period, which is essentially focused on the tail interval of the stable state, which has deviated from the initial transient response and can better reflect the steady-state pH level under the current process conditions, avoiding the influence of early dynamic fluctuations on the representativeness of the data.

[0074] Further, the effective acquisition window is intercepted at the end of the effective detection maintenance period, which is intended to ensure that a sufficient number of continuous pH detection data points are included in the window to support subsequent statistical analysis, avoid the influence of random noise, measurement fluctuations or temporary interference caused by relying only on a single instantaneous measurement value, and reduce the representativeness and reliability of the data. For example, the time length of the effective acquisition window can be dynamically determined according to the detection frequency of the pH probe, i.e., the sampling interval, and the preset minimum required analysis data amount.

[0075] The standard deviation of the latest PH data in the effective acquisition window is calculated and compared with the fluctuation threshold. If the standard deviation is lower than the fluctuation threshold, it indicates that the PH value fluctuation is small and the data is concentrated. The arithmetic mean value in the effective window is calculated as the effective PH value. If the standard deviation exceeds the fluctuation threshold, it indicates that the data dispersion is high. At this time, the histogram is used to analyze the distribution of the latest PH data in the effective acquisition window, and the interval median effective PH value with the highest frequency is extracted.

[0076] In the above operation, the fluctuation threshold of the standard deviation is used to represent the maximum dispersion degree of the pH measurement data in the effective acquisition window under stable working conditions, which is a benchmark for distinguishing whether the data fluctuation is within a normal range. The threshold can be set in the following way: in multiple production batches with known uniform sauce distribution and stable process operation, collect pH sequences of several effective detection periods to ensure that the data comes from a reliable state without external disturbance and sensor abnormality; calculate the standard deviation of each stable pH sequence to form a normal fluctuation sample set. Further calculate the mean and standard deviation of the sample set, so as to set the fluctuation threshold as the mean + 3 times the standard deviation. Thus, based on the statistical process control principle, most of the normal fluctuation scenarios can be covered, while the influence of extreme values is effectively suppressed, improving the robustness and criterion reliability of the threshold.

[0077] The effective pH value output by the pH probe in the last effective detection state before the shielding period starts is filled to compensate for the shielding period.

[0078] Specifically, the above-mentioned filling and compensation of the shielding period means that the effective pH value is extended to the entire shielding period on the time axis.

[0079] It should be noted that when the pH probe enters the shielding period, the detection signal is unreliable, at which time the formal record of the pH value is suspended. However, if the data output is completely interrupted, the continuity of online monitoring will be affected. Therefore, the high-confidence pH value obtained in the previous effective detection state is used to compensate for the data filling of the shielding period to maintain the integrity of the timing data and ensure the continuity and representativeness of subsequent process analysis, trend judgment, and quality tracing. As for the failure period, the failure warning mechanism is triggered preferentially rather than data compensation, and the fundamental reason is that the failure period is usually caused by hardware or performance abnormalities such as probe drift, contamination, response delay, or electrolyte leakage, indicating that the sensor has lost its basic measurement capability. At this time, if data filling is performed, the real failure may be covered up, leading to incorrect process control decisions and even affecting product safety. Therefore, the failure period should be clearly identified and alarmed to prompt maintenance personnel to calibrate or replace the probe in a timely manner.

[0080] The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, the above-described embodiments can be implemented in whole or in part in the form of a computer program product.

[0081] Those of ordinary skill in the art can realize that the modules and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are implemented in hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0082] In addition, the functional modules in each embodiment of the present application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0083] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0084] Finally, the above is merely preferred embodiments of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An artificial intelligence-based online pH detector, characterized in that ,include: The uniformity detection module is used to monitor the turbidity values ​​of different zones in the tank using symmetrically arranged optical turbidity sensors, and analyze the uniformity of sauce distribution by calculating the regional turbidity differences; A vibration control module is used to trigger gradient vibration control when uneven sauce distribution is detected, and synchronously monitor the sauce distribution uniformity and sauce liquid level offset during the vibration process, and then adjust the vibration according to the dynamic combination of sauce distribution uniformity and liquid level offset; The data acquisition module is used to obtain the pH value of the sauce in real time using an immersion pH probe and identify pH mutation signals of the sauce by comparing the real-time pH value of the sauce with the historical sliding mean; The dynamic compensation module is used to mark the detection status of the pH probe as a shielding period or a fault period based on the identified pH mutation signal of the sauce. When the detection status of the pH probe is the shielding period, the pH value recording is suspended, and when the pH probe is in the effective detection state, the sliding filter is started to output the effective pH value, and then the missing pH data in the shielding period is compensated according to the effective pH value. When the detection status of the pH probe is the fault period, a fault warning is triggered.

2. The artificial intelligence-based online pH detector according to claim 1, characterized in that: The process of using symmetrically arranged optical turbidity sensors to monitor the turbidity values ​​of the partitions in the tank is as follows: Optical turbidity sensors are symmetrically installed in the axial plane of the tank. The installation position of each sensor divides the internal space of the tank into several monitoring areas. Each optical turbidity sensor is used to detect the turbidity of the sauce liquid phase in the area where it is located, thereby obtaining the turbidity values ​​of different areas in the tank.

3. The artificial intelligence-based online pH detector according to claim 2, characterized in that: The method for analyzing the sauce distribution uniformity by calculating regional turbidity differences includes the following: According to the symmetrical installation positions of multiple optical turbidity sensors on the axial plane of the tank body, the liquid space in the tank is divided into several symmetrically distributed monitoring areas; At any sampling moment, the regional turbidity values ​​measured by each sensor are synchronously obtained, and the turbidity values ​​of each symmetrical area are compared pairwise to calculate the turbidity difference; The standard deviation was calculated based on the turbidity difference between all symmetrical areas as an evaluation index of sauce uniformity.

4. The artificial intelligence-based online pH detector according to claim 3, characterized in that: The gradient vibration control is triggered when the sauce distribution is identified to be uneven as follows: The sauce uniformity evaluation index is compared with the preset uniformity threshold. When the sauce uniformity evaluation index exceeds the uniformity threshold, the gradient vibration control is activated as follows: The controllable vibration frequency range of the probe vibrating rod is divided into a number of discrete vibration frequency levels according to the set frequency step size, forming a vibration frequency sequence with increasing characteristics; The control probe vibrator starts to implement low-intensity vibration in the tank body from an initial vibration frequency within a controllable vibration frequency range, and automatically increases a vibration frequency level every set time interval until the maximum vibration frequency within the controllable vibration frequency range is reached.

5. The artificial intelligence-based online pH detector according to claim 1, characterized in that: The sauce liquid level deviation is shown in the following monitoring process: A laser displacement sensor is installed on the side wall of the sauce tank, with its optical axis aligned with the center of the liquid level in the tank, for non-contact real-time measurement of the liquid level during vibration. During the vibration period, the distance signal output by the laser displacement sensor is continuously collected and compared with the initial liquid level height measured before the vibration is started to obtain the real-time offset of the sauce liquid level relative to the initial liquid level height.

6. The artificial intelligence-based online pH detector according to claim 4, characterized in that: The vibration adjustment is performed according to the dynamic combination of sauce distribution uniformity and liquid level offset as follows: During the vibration process, the sauce uniformity evaluation index monitored in real time is compared with the uniformity threshold, and the sauce liquid level deviation is compared with the safety deviation; When the sauce uniformity evaluation index is still higher than the uniformity threshold at a certain moment and the sauce liquid level offset meets the safety offset, the vibration operation continues to be performed according to the established gradient frequency increase strategy; When the sauce uniformity evaluation index drops below the uniformity threshold at a certain moment and the sauce liquid level offset meets the safety offset, it switches to the micro-pulse vibration mode; When the sauce uniformity evaluation index drops below the uniformity threshold at a certain moment and the sauce liquid level deviation is higher than the safe deviation, the micro-pulse vibration mode is maintained; When the sauce uniformity evaluation index is still higher than the uniformity threshold and the liquid level offset is higher than the safety offset at a certain moment, the vibration frequency level is gradually reduced from the preset vibration frequency sequence based on the current vibration frequency until the liquid level offset falls back to the safety offset; When the liquid level drops back to the target frequency level of the safety offset, an execution period is entered to continuously monitor the sauce distribution uniformity; If the sauce uniformity evaluation index drops below the uniformity threshold within the execution time, the micro-pulse vibration mode is maintained; If it does not fall below the uniformity threshold, the frequency adjustment step is reduced and slightly increased to obtain a fine-tuning frequency based on the target frequency level. The liquid level offset and uniformity changes are then continuously and synchronously monitored at this frequency until the dual goals of uniformity and liquid level stability are achieved, and finally the micro-pulse vibration mode is maintained.

7. The artificial intelligence-based online pH detector according to claim 1, characterized in that: The identification of sauce pH mutation signal is performed as follows: Use the pH probe to continuously collect and record the pH value sequence of the sauce within the set time window; Compare the current real-time pH value with the sliding mean of the historical pH values ​​within the time window and calculate the instantaneous deviation between the two; The deviation is compared with the configured pH mutation judgment limit. If the deviation between the real-time pH value and the sliding mean at the current moment exceeds the pH mutation judgment limit, it is determined that a pH mutation signal of the sauce is detected.

8. The artificial intelligence-based online pH detector according to claim 3, characterized in that: The specific operation of marking the detection state of the pH probe as the shielding period or the fault period based on the identified pH mutation signal of the sauce is as follows: When a pH mutation signal of the sauce is identified, the sauce uniformity evaluation index within a time window before the moment is traced back and analyzed; If the sauce uniformity evaluation index is lower than the uniformity threshold within the time window, the current detection state of the pH probe is marked as a fault period; On the contrary, if the sauce uniformity evaluation index is higher than the uniformity threshold within the time window, the tracking phase is entered. In the subsequent monitoring process, it is continuously tracked whether there is still a pH mutation signal within the time window when the sauce uniformity recovers to below the threshold. If the pH mutation signal disappears and does not reappear, the current detection status of the pH probe is marked as the shielding period, otherwise it is marked as the fault period.

9. The artificial intelligence-based online pH detector according to claim 3, characterized in that: The effective detection state is a period when the sauce uniformity evaluation index is lower than the uniformity threshold and no pH mutation signal is detected.

10. The artificial intelligence-based online pH detector according to claim 1, characterized in that: When the pH probe is in a valid detection state, the sliding filter is started to output the effective pH value, and then the pH data missing during the shielding period is compensated according to the effective pH value as follows: When the pH probe is in the effective detection state, the duration of the current effective detection state is recorded, and the effective collection window is intercepted at the end of the effective detection maintenance period, and the latest pH data within the effective collection window is collected; The standard deviation of the latest pH data within the effective acquisition window is calculated and compared with the fluctuation threshold. If the standard deviation is lower than the fluctuation threshold, the arithmetic mean within the effective window is calculated as the effective pH value; if the standard deviation exceeds the fluctuation threshold, the high-frequency value is determined by histogram statistics as the effective pH value; For the shielding period entered by the pH probe, filling compensation is performed with the effective pH value output under the most recent effective detection state before the start of the shielding period.

Images