A multi-parameter intelligent early warning system for oil-in-water suspensions

Abstract

The present application relates to the technical field of water quality monitoring, in particular to a multi-parameter intelligent early warning system for oil suspended matter in water, comprising: a data acquisition module for acquiring fluorescence signals of an ultraviolet fluorescence channel and turbidity signals of an infrared turbidity channel; a data determination module; an anomaly detection module for determining whether a suspected abnormal event occurs according to a turbidity sliding standard deviation of the turbidity signals; an anomaly identification module for determining an event identification state of the suspected abnormal event according to a turbidity peak value decay ratio in response to a suspected abnormality flag; an interference type determination module for determining an abnormal interference type of the abnormal event according to a turbidity fluctuation complexity and a pulse density ratio of the turbidity signals; and an anomaly early warning module for generating a corresponding early warning signal according to the abnormal interference type, thereby improving the detection accuracy of oil suspended matter in water.

Description

Technical Field

[0001] This invention relates to the field of water quality monitoring technology, specifically to a multi-parameter intelligent early warning system for oil suspended solids in water. Background Technology

[0002] With the rapid development of industrial production, the regulatory requirements for industrial wastewater discharge are becoming increasingly stringent. Oil and suspended solids in water are core control indicators for wastewater discharge from industries such as machining, metal smelting, and chemicals. Non-contact optical monitoring technology has become the mainstream technology for online monitoring of oil and suspended solids in water due to its advantages such as good real-time performance, no reagents required, and no secondary pollution.

[0003] Currently, mainstream non-contact online monitoring systems for oil and suspended solids in water generally adopt a dual-channel architecture: using ultraviolet fluorescence to detect oil concentration and infrared transmission to detect suspended solids concentration. This dual-channel technology can obtain relatively accurate measurement results in scenarios with a single interference source. However, in actual industrial settings, emulsified oil droplets in the infrared turbidity channel produce a light scattering effect similar to that of real suspended solids particles, leading to a falsely inflated suspended solids measurement value; real suspended solids particles also produce scattering interference in the ultraviolet fluorescence channel, causing deviations in oil concentration measurements. When the water sample contains both emulsified oil and real suspended solids particles, both types of interference occur simultaneously, forming a complex scenario with multiple sources of interference. Existing dual-channel compensation methods rely on the correlation or fixed ratio between the signals of the two channels to determine the type of interference and perform corrections. However, in scenarios with multiple sources of interference, because the two types of pollution events are highly synchronized in time, even if the signals come from different pollution sources, they will exhibit high correlation, leading to incorrect compensation directions. The corrected oil concentration and suspended solids concentration values ​​both deviate from the true values, resulting in distorted measurement results. Summary of the Invention

[0004] To address this issue, the present invention provides a multi-parameter intelligent early warning system for oil suspension in water, which overcomes the problem that existing technologies do not consider the influence of the fluctuation pattern characteristics of the turbidity signal itself in distinguishing the type of interference, but only rely on the correlation between the two-channel signals for judgment. This results in the inability to accurately distinguish the type of interference in scenarios with multiple sources of interference, causing incorrect compensation direction and distorted measurement results, thus leading to poor detection accuracy of oil suspension in water.

[0005] To achieve the above objectives, the present invention provides a multi-parameter intelligent early warning system for oil suspension in water, comprising: The data acquisition module is used to acquire the fluorescence signal of the ultraviolet fluorescence channel and the turbidity signal of the infrared turbidity channel; The data determination module, which is connected to the data acquisition module, includes a first data determination unit for determining the turbidity sliding standard deviation of the turbidity signal, a second data determination unit for determining the turbidity peak attenuation ratio of the turbidity signal, a third data determination unit for determining the turbidity fluctuation complexity of the turbidity signal, and a fourth data determination unit for determining the pulse density ratio of the turbidity signal. An anomaly detection module, which is connected to the data determination module, is used to determine whether a suspected anomaly event has occurred based on the turbidity sliding standard deviation of the turbidity signal. An anomaly identification module, which is connected to the data determination module and the anomaly detection module respectively, is used to respond to a suspected anomaly flag and determine the event identification state of the suspected anomaly event based on the turbidity peak attenuation ratio of the turbidity signal, wherein the event identification state is an anomaly event or a non-anomaly event. An interference type determination module, which is connected to the data determination module and the anomaly identification module respectively, determines the abnormal interference type of the abnormal event based on the turbidity fluctuation complexity and pulse density ratio of the turbidity signal. The abnormal interference types include emulsified oil interference, silt interference and composite interference. An abnormality warning module, which is connected to the interference type determination module, is used to generate a corresponding warning signal according to the abnormal interference type. The warning signal includes a first warning signal, a second warning signal, a third warning signal, and a fourth warning signal.

[0006] Preferably, the anomaly detection module determines that no anomaly has occurred if the turbidity sliding standard deviation of the turbidity signal is less than a preset turbidity sliding standard deviation. The anomaly detection module responds to a turbidity sliding standard deviation being greater than or equal to a preset turbidity sliding standard deviation by determining that a suspected anomaly event has occurred, recording the current moment as the event starting point, and generating a suspected anomaly flag.

[0007] Preferably, the first data determining unit includes: The window setting subunit is used to set the preset sliding window length and the preset sliding step size. The preset sliding window length is the number of sampling points contained in the window. The standard deviation calculation subunit is used to calculate the standard deviation of the turbidity signal sequence within the window after each sliding window, starting from the first sampling point and using a preset step size, to obtain the turbidity sliding standard deviation.

[0008] Preferably, the anomaly identification module is used to respond to the suspected anomaly flag and determine the event identification status of the suspected anomaly event based on the turbidity peak attenuation ratio of the turbidity signal, wherein, If the turbidity peak attenuation ratio meets the preset attenuation condition, the event identification state is determined to be an abnormal event; if the turbidity peak attenuation ratio does not meet the preset attenuation condition, the event identification state is determined to be a non-abnormal event; the preset attenuation condition is that the turbidity peak attenuation ratio is less than a preset attenuation threshold.

[0009] Preferably, the second data determining unit includes: The time period segmentation subunit is used to set the event start point as the reference point, determine the time period within a first preset duration forward from the reference point as the pre-event steady state period, and determine the time period within a second preset duration backward from the reference point as the event window, and respectively acquire the turbidity signal within the pre-event steady state period and the turbidity signal within the event window. A peak determination subunit is used to determine the turbidity peak value in the turbidity signal within the event window; The baseline determination subunit is used to record the arithmetic mean of the turbidity signal sequence during the steady-state period before the event as the baseline value; A steady-state determination subunit is used to extract a preset third time period from the time corresponding to the turbidity peak within the event window. When the difference between the maximum and minimum values ​​of the turbidity signal within the third time period is less than a preset range value, the arithmetic mean of the turbidity signal sequence within the third time period is determined as the steady-state value. The attenuation ratio calculation subunit is used to calculate the turbidity peak attenuation ratio, which is the ratio of a first difference to a second difference, wherein the first difference is the difference between the turbidity peak and the steady-state value, and the second difference is the difference between the turbidity peak and the baseline value.

[0010] Preferably, the interference type determination module determines the abnormal interference type of the abnormal event based on the turbidity fluctuation complexity of the turbidity signal, wherein, If the turbidity fluctuation complexity is less than the first preset turbidity fluctuation complexity, then the abnormal interference type is determined to be emulsified oil interference. If the turbidity fluctuation complexity is greater than or equal to the first preset turbidity fluctuation complexity and less than the second preset turbidity fluctuation complexity, then the abnormal interference type of the abnormal event is determined for the second time based on the pulse density ratio of the turbidity signal. If the turbidity fluctuation complexity is greater than or equal to the second preset turbidity fluctuation complexity, then the abnormal interference type is determined to be sediment scattering interference. The first preset turbidity fluctuation complexity is less than the second preset turbidity fluctuation complexity.

[0011] Preferably, the third data determining unit includes: The coarse-grained subunit is used to divide the turbidity signal sequence within the event window into several segments of equal length according to a preset length, and to obtain a coarse-grained sequence based on the arithmetic mean of the turbidity signal in each segment. Binarization subunit is used to generate a binary sequence based on the comparison between the arithmetic mean of each segment in the coarse-grained sequence and the median of the coarse-grained sequence, wherein the corresponding position of the segment whose arithmetic mean is greater than or equal to the median takes a first value, and the corresponding position of the segment whose arithmetic mean is less than the median takes a second value. The complexity calculation subunit is used to scan the binary sequence value by value, record the permutation patterns of the binary sequences that have appeared, and increment the sequence complexity by one whenever a permutation pattern of a binary sequence that has not appeared is scanned, and record the permutation pattern as having appeared. The sequence complexity is obtained after the scan is completed. A normalized subunit is used to determine the turbidity fluctuation complexity as the ratio of the sequence complexity to the number of segments.

[0012] Preferably, the interference type determination module is further configured to determine that the abnormal interference type is composite interference and dominated by emulsified oil components when the pulse density ratio is less than the preset pulse density ratio. The interference type determination module is also used to determine that when the pulse density ratio is greater than or equal to the preset pulse density ratio, the abnormal interference type is composite interference and the sediment component is dominant.

[0013] Preferably, the abnormal warning module generates a first warning signal in response to the abnormal interference type being emulsified oil interference. The first warning signal includes an emulsified oil leakage warning and an indication to use demulsification treatment. The abnormal early warning module responds to the abnormal interference type being sediment scattering interference and generates a second early warning signal. The second early warning signal includes an early warning of excessive suspended matter and an indication to use coagulation and sedimentation treatment. The abnormal early warning module responds to the abnormal interference type being a complex interference with emulsified oil components as the main component, and generates a third early warning signal. The third early warning signal includes a first complex pollution warning and an indication that demulsification treatment should be the main treatment and coagulation and sedimentation treatment should be the auxiliary treatment. The abnormal early warning module responds to the abnormal interference type being a complex interference with sediment as the dominant component, and generates a fourth early warning signal. The fourth early warning signal includes a second complex pollution warning and an indication that coagulation and sedimentation treatment should be the main treatment, with demulsification treatment as a supplement.

[0014] Preferably, the fourth data determining unit includes: The differential calculation subunit is used to calculate the first-order difference sequence of the turbidity signal within the event window, wherein the first-order difference sequence is formed by sequentially arranging the differences between adjacent sampling points in the turbidity signal; The pulse threshold determination subunit is used to calculate the standard deviation of the first-order difference sequence of the turbidity signal during the steady-state period before the event, and to determine a preset multiple of the standard deviation as the pulse threshold. The pulse statistics subunit is used to count the number of differences in the first-order difference sequence whose absolute value is greater than the pulse threshold, thereby obtaining the number of pulses. The density ratio calculation subunit is used to determine the pulse density ratio as the ratio of the number of pulses to the total number of differences in the first-order difference sequence.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention achieves quantitative characterization of the multi-dimensional fluctuation characteristics of turbidity signals by simultaneously acquiring fluorescence signals from the ultraviolet fluorescence channel and turbidity signals from the infrared turbidity channel, and uniformly extracting turbidity sliding standard deviation, turbidity peak attenuation ratio, turbidity fluctuation complexity, and pulse density ratio from the turbidity signals; it identifies suspected abnormal events and marks the event starting point based on the turbidity sliding standard deviation, and then eliminates instantaneous disturbances such as bubbles based on the turbidity peak attenuation ratio, achieving reliable screening from signal anomalies to actual pollution events; it classifies turbidity fluctuation complexity as a macro-criteria, directly determining emulsified oil interference or sediment scattering interference when the overall fluctuation pattern can be clearly distinguished, and automatically triggering a micro-level secondary judgment based on the pulse density ratio when the fluctuation pattern falls into the fuzzy range, achieving hierarchical progressive judgment of macro-pattern analysis and micro-pulse statistics; through the anomaly early warning module, it generates differentiated processing instructions with priority distinctions according to different interference types, realizing a closed-loop connection from interference identification to disposal guidance, thereby improving the detection accuracy and early warning reliability of the non-contact water oil suspended solids monitoring system in scenarios with multiple sources of interference.

[0016] 2. This invention achieves initial screening of abnormal events by comparing the turbidity sliding standard deviation with a preset turbidity sliding standard deviation. When the deviation is less than the preset threshold, it is determined that no abnormality has occurred. When the deviation is greater than or equal to the preset threshold, it is determined that a suspected abnormal event has occurred and the event starting point is marked and a suspected abnormality flag is generated simultaneously.

[0017] 3. This invention compares the turbidity peak attenuation ratio with a preset attenuation threshold, and determines an abnormal event when the preset attenuation condition is met and a non-abnormal event when it is not met, thereby achieving automatic identification of true and false anomalies, reducing false alarms, and effectively eliminating instantaneous disturbances such as bubbles.

[0018] 4. This invention improves the judgment efficiency of the early warning system by comparing the turbidity fluctuation complexity with a first preset turbidity fluctuation complexity and a second preset turbidity fluctuation complexity. When the turbidity fluctuation complexity is less than the first preset threshold, it is determined to be emulsified oil interference; when it is greater than or equal to the second preset threshold, it is determined to be mud and sand scattering interference; and when it falls between the two, a secondary judgment based on the pulse density ratio is triggered.

[0019] 5. This invention improves the accuracy of the early warning system in complex interference scenarios by introducing a comparison between the pulse density ratio and a preset pulse density ratio when the turbidity fluctuation complexity falls within the fuzzy range between the first preset threshold and the second preset threshold. When the pulse density ratio is less than the preset threshold, it is determined to be complex interference dominated by emulsified oil components; when it is greater than or equal to the preset threshold, it is determined to be complex interference dominated by silt components. When it is impossible to clearly distinguish, a secondary determination is made from the local pulse density dimension. Attached Figure Description

[0020] The present application will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0021] Figure 1 This is a schematic diagram of the module connection of the multi-parameter intelligent early warning system for oil suspension in water according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating how the turbidity moving standard deviation of the turbidity signal determines whether a suspected abnormal event has occurred, according to an embodiment of the present invention. Figure 3 This is a flowchart illustrating how an event identification state for a suspected abnormal event is determined based on the turbidity peak attenuation ratio of the turbidity signal, according to an embodiment of the present invention. Figure 4 This is a flowchart illustrating how an abnormal interference type of an abnormal event is determined based on the turbidity fluctuation complexity of the turbidity signal, according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figure 1 , Figure 2 , Figure 3 as well as Figure 4The diagrams shown are: a schematic diagram of the module connection of the multi-parameter intelligent early warning system for oil suspension in water according to an embodiment of the present invention; a flowchart of the present invention for determining whether a suspected abnormal event has occurred based on the turbidity sliding standard deviation of the turbidity signal; a flowchart of the present invention for determining the event identification state of a suspected abnormal event based on the turbidity peak attenuation ratio of the turbidity signal; and a flowchart of the present invention for determining the abnormal interference type of the abnormal event based on the turbidity fluctuation complexity of the turbidity signal.

[0024] This invention provides a multi-parameter intelligent early warning system for oil suspension in water, comprising: The data acquisition module is used to acquire the fluorescence signal of the ultraviolet fluorescence channel and the turbidity signal of the infrared turbidity channel; The data determination module, which is connected to the data acquisition module, includes a first data determination unit for determining the turbidity sliding standard deviation of the turbidity signal, a second data determination unit for determining the turbidity peak attenuation ratio of the turbidity signal, a third data determination unit for determining the turbidity fluctuation complexity of the turbidity signal, and a fourth data determination unit for determining the pulse density ratio of the turbidity signal. An anomaly detection module, which is connected to the data determination module, is used to determine whether a suspected anomaly event has occurred based on the turbidity sliding standard deviation of the turbidity signal. An anomaly identification module, which is connected to the data determination module and the anomaly detection module respectively, is used to respond to a suspected anomaly flag and determine the event identification state of the suspected anomaly event based on the turbidity peak attenuation ratio of the turbidity signal, wherein the event identification state is an anomaly event or a non-anomaly event. An interference type determination module, which is connected to the data determination module and the anomaly identification module respectively, determines the abnormal interference type of the abnormal event based on the turbidity fluctuation complexity and pulse density ratio of the turbidity signal. The abnormal interference types include emulsified oil interference, silt interference and composite interference. An abnormality warning module, which is connected to the interference type determination module, is used to generate a corresponding warning signal according to the abnormal interference type. The warning signal includes a first warning signal, a second warning signal, a third warning signal, and a fourth warning signal.

[0025] Specifically, the data acquisition module includes an ultraviolet fluorescence detector and an infrared turbidity detector. The ultraviolet fluorescence detector contains an ultraviolet light source and a fluorescence receiver, while the infrared turbidity detector contains an infrared light source and an infrared receiver. Both detectors employ a non-contact design, meaning that the optical windows do not directly contact the water sample; instead, they are optically coupled to the water sample in the flow cell through a transparent window.

[0026] Specifically, the specific structure of the anomaly detection module, anomaly identification module, interference type determination module, and anomaly early warning module is not limited. They themselves and their units can be composed of logic components, including field-programmable components, computers, or microprocessors in computers.

[0027] Specifically, the anomaly detection module determines that no anomaly has occurred if the turbidity sliding standard deviation of the turbidity signal is less than the preset turbidity sliding standard deviation of 0.12 mg / L. The anomaly detection module responds to a turbidity sliding standard deviation being greater than or equal to a preset turbidity sliding standard deviation by determining that a suspected anomaly event has occurred, recording the current moment as the event starting point, and generating a suspected anomaly flag.

[0028] In this embodiment, the preset turbidity sliding standard deviation is set at 0.12 mg / L. Under normal operating conditions (during periods when no pollution emissions are confirmed), turbidity signals from the infrared turbidity channel are continuously collected for 24 hours at a sampling frequency of 50 Hz. Using a sliding window length of 250 sampling points and a sliding step size of 1 sampling point, the turbidity sliding standard deviation is calculated for each moment within 24 hours, resulting in a sequence of turbidity sliding standard deviation over time. The cumulative distribution of this sequence is statistically analyzed, and the 99th quantile is taken as the preset turbidity sliding standard deviation. Experimental results show that under normal operating conditions, including different time periods (day and night) and different ambient temperatures (15℃ to 35℃), over 99% of the turbidity sliding standard deviations do not exceed 0.10 mg / L. Using 0.12 mg / L as the preset turbidity sliding standard deviation provides a 20% margin beyond the upper limit of normal fluctuations, ensuring both sensitive detection of real pollution events and avoiding false alarms caused by occasional minor fluctuations under normal operating conditions.

[0029] Specifically, under normal operating conditions, the suspended solids content in the water sample within the flow-through tank is at background levels, and the turbidity reading fluctuates slightly and randomly around the baseline value. The numerical differences between sampling points in the turbidity signal sequence within the sliding window are very small, and the turbidity sliding standard deviation remains at a low level. Therefore, when the turbidity sliding standard deviation is less than the preset turbidity sliding standard deviation, it can be determined that no abnormal event has occurred. When wastewater containing emulsified oil or silt is suddenly discharged into the detection tank, a large number of particles instantly enter the infrared light path and produce scattering effects. The turbidity reading rises sharply from the baseline level to a higher level in a short period of time. The sliding window simultaneously includes the low baseline value before the event and the high pollution value after the event. The numerical differences between sampling points in the signal sequence increase sharply, and the turbidity sliding standard deviation rises significantly accordingly. Therefore, when the turbidity sliding standard deviation is greater than or equal to the preset turbidity sliding standard deviation, a suspected abnormal event is determined to have occurred. The current moment is recorded as the event starting point, and a suspected abnormality flag is generated. It should be noted that this step only determines whether a suspected abnormal event has occurred, without distinguishing whether the event is actual pollution or transient disturbances such as bubbles, nor does it distinguish the type of pollution. This is because the turbidity sliding standard deviation is only sensitive to the amplitude of signal fluctuations and cannot distinguish the physical causes of these fluctuations. The ingress of emulsified oil, sediment, or the passage of air bubbles can all cause rapid jumps in the turbidity signal, leading to an increase in the turbidity sliding standard deviation. These situations are uniformly marked as suspected anomalies and passed to subsequent modules for step-by-step identification, ensuring the sensitivity of anomaly detection while avoiding incorrect judgments due to insufficient information in the initial screening stage.

[0030] Specifically, the first data determining unit includes: A window setting subunit is used to set a preset sliding window length of 250 sampling points and a preset sliding step size of 1 sampling point. The preset sliding window length is the number of sampling points contained within the window. In this embodiment, the sampling frequency of the infrared turbidity detector is 50Hz, that is, 50 turbidity readings are collected per second. The preset sliding window length is set to 250 sampling points (corresponding to a 5-second duration), and the preset sliding step size is set to 1 sampling point (i.e., sliding point by point). The selection of the window length needs to achieve a balance between response speed and anti-interference. Comparative experiments were conducted with 150 sampling points (3 seconds), 250 sampling points (5 seconds), and 400 sampling points (8 seconds) to test the response time and false alarm rate of abnormal event detection under the same pollution emission conditions. The results show that the 150-point window significantly increases false alarms due to large particles occasionally passing through the pipeline, and the 400-point window lags the event start point marker time by an average of more than 3 seconds behind the actual pollution start time. The 250-point window achieves the optimal balance between response timeliness and anti-interference. The standard deviation calculation subunit is used to calculate the standard deviation of the turbidity signal sequence within the window after each sliding window, starting from the first sampling point and using a preset step size, to obtain the turbidity sliding standard deviation.

[0031] Specifically, the anomaly identification module is used to respond to the suspected anomaly flag and determine the event identification status of the suspected anomaly event based on the turbidity peak attenuation ratio of the turbidity signal, wherein, If the turbidity peak attenuation ratio meets the preset attenuation condition, the event identification state is determined to be an abnormal event; if the turbidity peak attenuation ratio does not meet the preset attenuation condition, the event identification state is determined to be a non-abnormal event; the preset attenuation condition is that the turbidity peak attenuation ratio is less than the preset attenuation threshold of 0.3.

[0032] Specifically, pure emulsified oil standard water samples (oil concentration 50 ppm) and pure sediment standard water samples (suspended solids concentration 80 mg / L) were prepared separately. Ten independent discharge tests were conducted under the same experimental pipeline and flow rate conditions. For each test, the peak turbidity, steady-state value, and baseline value were recorded, and the peak turbidity attenuation ratio was calculated. The peak turbidity attenuation ratio for the pure emulsified oil sample ranged from 0.04 to 0.15, while that for the pure sediment sample ranged from 0.06 to 0.18. Simultaneously, ten bubble interference tests were conducted by injecting microbubbles into the flow-through tank, and the peak turbidity attenuation ratio was consistently above 0.72. A preset attenuation threshold of 0.3 was used, providing sufficient decision margin between actual pollution events (attenuation ratio ≤ 0.18) and bubble interference (attenuation ratio ≥ 0.72), reliably distinguishing between the two types of events.

[0033] Specifically, when the turbidity peak decay is relatively small (less than the preset decay threshold), it indicates that the difference between the turbidity peak and the steady-state value is small. That is, the turbidity signal does not drop rapidly after reaching the peak, but remains at a high level. This is a typical characteristic of real pollution events, because emulsified oil droplets or sediment particles remain in the water sample after entering the flow tank, continuously scattering infrared light, and the turbidity reading will not recover to the baseline level in a short time. When the turbidity peak decay is relatively large (greater than or equal to the preset decay threshold), it indicates that the turbidity signal drops rapidly after reaching the peak, with a large first difference. The turbidity signal recovers to near the baseline level in a short time. This is a characteristic of instantaneous disturbances such as bubbles, because bubbles pass through the detection tank quickly with the water flow, scattering infrared light only for a very short time. The turbidity reading quickly recovers to the normal level. Identifying such events as non-abnormal events can effectively eliminate bubble interference, avoid initiating subsequent interference type determination processes for invalid events, and reduce false alarms.

[0034] Specifically, the second data determining unit includes: The time period segmentation subunit is used to set the event start point as the reference point, determine the time period within a first preset duration of 10 seconds before the reference point as the pre-event steady state period, and determine the time period within a second preset duration of 60 seconds after the reference point as the event window, and respectively acquire the turbidity signal within the pre-event steady state period and the turbidity signal within the event window. The peak determination subunit is used to determine the turbidity peak value in the turbidity signal within the event window. It scans the turbidity signal sequence within the event window point by point, records the maximum value of the turbidity reading as the turbidity peak value, and records the time corresponding to the maximum value. The baseline determination subunit is used to record the arithmetic mean of the turbidity signal sequence during the steady-state period before the event as the baseline value; A steady-state determination subunit is used to extract a preset third time period of 5 seconds from the moment corresponding to the turbidity peak within the event window. When the difference between the maximum and minimum values ​​of the turbidity signal within the third time period is less than a preset range value, the arithmetic mean of the turbidity signal sequence within the third time period is determined as the steady-state value. In this embodiment, the preset range value is 10% of the baseline value, and the fluctuation amplitude within 5 seconds usually does not exceed 10% of the normal baseline level. Instantaneous disturbances such as bubbles cause the turbidity to drop significantly within 5 seconds after the peak, and the difference between the maximum and minimum values ​​is far greater than 10% of the baseline value. Comparative experiments are conducted using 5%, 10%, and 20% of the baseline value as range values, respectively. Comparative experiments are conducted under known water sample conditions, including 30 real polluted water samples (15 emulsified oil water samples and 15 silt water samples) and 30 water samples with bubble interference. The experimental results show that: under the 5% range condition, in real pollution events, the steady-state segment was not correctly identified 5 times due to normal small fluctuations in the turbidity signal exceeding the baseline value by 5% but not exceeding 10% within 5 seconds, resulting in an identification success rate of 83.3%; under the 10% range condition, all real pollution events were correctly identified, with an identification success rate of 100%, and all bubble interference was correctly eliminated; under the 20% range condition, bubble interference was misjudged as a valid steady-state segment 3 times due to an excessively wide range threshold, with an elimination success rate of 90%; the 10% range condition achieves the optimal balance between the identification success rate of real pollution events and the elimination success rate of bubble interference.

[0035] The attenuation ratio calculation subunit is used to calculate the turbidity peak attenuation ratio, which is the ratio of a first difference to a second difference, wherein the first difference is the difference between the turbidity peak and the steady-state value, and the second difference is the difference between the turbidity peak and the baseline value.

[0036] In this embodiment, the first preset duration of 10 seconds is determined based on the following: Under normal operating conditions, the fluctuation period of the turbidity signal is usually within a few seconds. One hundred sets of normal operating condition samples from different time periods and ambient temperatures were selected, and comparative experiments were conducted at durations of 5 seconds, 10 seconds, and 20 seconds. The experimental results show that: at a duration of 5 seconds, the baseline value fluctuates significantly, with the standard deviation being 1.8 times that of the 10-second duration, making it impossible to obtain a stable and reliable baseline calculation result; at a duration of 20 seconds, the deviation between the baseline value and the 10-second duration is less than 3%, but it significantly increases the system's response delay, leading to a lag in the determination of abnormal events. A duration of 10 seconds achieves the optimal balance between baseline value stability and timely value acquisition.

[0037] The second preset duration of 60 seconds is based on the following: the duration of typical pollution emission events in the machining industry, as targeted in this embodiment, is usually between 30 seconds and several minutes. One hundred actual industrial emission event samples were selected, and comparative experiments were conducted using windows of 40 seconds, 60 seconds, and 90 seconds. The experimental results show that approximately 15% of events under the 40-second window failed to fully capture the steady-state segment, resulting in the inability to accurately calculate the turbidity peak attenuation ratio, thus affecting the accuracy of event identification. The 90-second window showed no significant difference in the extraction effect of core features such as turbidity fluctuation complexity and pulse density ratio compared to the 60-second window, but increased computational load and memory usage by 33%, putting pressure on the real-time performance of embedded devices. The 60-second window achieves the optimal balance between window integrity and computational efficiency.

[0038] Specifically, the interference type determination module determines the abnormal interference type of the abnormal event based on the turbidity fluctuation complexity of the turbidity signal, wherein, If the turbidity fluctuation complexity is less than the first preset turbidity fluctuation complexity of 0.25, then the abnormal interference type is determined to be emulsified oil interference; If the turbidity fluctuation complexity is greater than or equal to the first preset turbidity fluctuation complexity and less than the second preset turbidity fluctuation complexity of 0.55, then the abnormal interference type of the abnormal event is determined for the second time based on the pulse density ratio of the turbidity signal. If the turbidity fluctuation complexity is greater than or equal to the second preset turbidity fluctuation complexity, then the abnormal interference type is determined to be sediment scattering interference. The first preset turbidity fluctuation complexity is less than the second preset turbidity fluctuation complexity.

[0039] In this embodiment, the first preset turbidity fluctuation complexity ranges from [0.20, 0.30], and the second preset turbidity fluctuation complexity ranges from [0.50, 0.60]. Preferably, the first preset turbidity fluctuation complexity is 0.25, and the second preset turbidity fluctuation complexity is 0.55. Pure emulsified oil standard water samples (oil concentration 50 ppm, prepared by diluting commercially available water-soluble cutting fluid at a volume ratio of 1:20) and pure sediment standard water samples (suspended solids concentration 80 mg / L, prepared by dispersing 800-mesh quartz powder in clean water) are prepared separately. Under the same conditions of pipeline flow velocity 0.5 m / s, detection pool inner diameter 25 mm, and sampling frequency 50 Hz, each water sample is continuously discharged at a single discharge volume of 500 mL and a discharge time of approximately 15 seconds, and the test is independently repeated 10 times. The turbidity signal within the event window is recorded for each test, and the turbidity fluctuation complexity is obtained through coarsening, binarization, complexity calculation, and normalization steps. The turbidity fluctuation complexity of pure emulsified oil-water samples is between 0.10 and 0.22, and the turbidity fluctuation complexity of pure sediment-water samples is between 0.58 and 0.75. Accordingly, 0.25 and 0.55 are selected as the preferred values ​​for the first and second preset turbidity fluctuation complexities, respectively. For different types of detectors or pipeline configurations, those skilled in the art can refer to the above test methods and parameters to redetermine the appropriate first and second preset turbidity fluctuation complexities.

[0040] Specifically, emulsified oil droplets are flexible particles that deform under the shear force of the water flow. During the deformation process, they absorb some of the fluid's kinetic energy. The movement of the oil droplets is relatively smooth and orderly, and the fluctuation pattern of the turbidity signal has high repeatability, thus the turbidity fluctuation complexity is low. Sediment particles, on the other hand, are rigid particles that do not deform in the water flow. Their collisions with the fluid and the rigid walls generate a large number of random, high-frequency pulse fluctuations. The turbidity signal is full of spikes and abrupt changes, and the fluctuation pattern is constantly changing, thus the turbidity fluctuation complexity is high.

[0041] Specifically, the third data determination unit includes: The coarse-grained subunit is used to divide the turbidity signal sequence within the event window into several equal-length segments according to a preset length, and obtain a coarse-grained sequence based on the arithmetic mean of the turbidity signal within each segment. In this embodiment, the event window duration is 60 seconds, the infrared turbidity detector sampling frequency is 50Hz, and there are a total of 3000 sampling points within the event window. The preset length is 30 sampling points, that is, the turbidity signal sequence is divided into 100 equal-length segments, and the coarse-grained sequence contains 100 arithmetic means. Binarization subunit is used to generate a binary sequence based on the comparison between the arithmetic mean of each segment in the coarse-grained sequence and the median of the coarse-grained sequence, wherein the corresponding position of the segment whose arithmetic mean is greater than or equal to the median takes a first value of 1, and the corresponding position of the segment whose arithmetic mean is less than the median takes a second value of 0. The complexity calculation subunit is used to scan the binary sequence value by value, record the permutation patterns of the binary sequences that have appeared, and increment the sequence complexity by one whenever a permutation pattern that has not appeared before is scanned, and the permutation pattern is recorded as having appeared. After the scan is completed, the sequence complexity is obtained. The specific implementation steps of this calculation process are as follows: The binary sequence is denoted as S(i), i=1,2,…,K, where K is the number of sub-segments (K=100 in this embodiment). The sequence complexity is initialized to 0, the set of existing patterns is initialized to empty, and the scan starting point p=1 is set. Starting from p, the scan length L=1,2,3,… is gradually increased, checking whether the sub-sequences S(p) to S(p+L-1) are already in the set of existing patterns. If they exist, L is increased; if they do not exist, the sequence complexity is incremented by one, the sub-sequence is added to the set of existing patterns, the scan starting point is updated to p=p+L, L=1 is reset, and the scan continues. When the scan starting point p exceeds K, the scan is completed, and the sequence complexity at this time is the calculation result.

[0042] A normalized subunit is used to determine the turbidity fluctuation complexity as the ratio of the sequence complexity to the number of segments. In this embodiment, the number of segments is 100, and the turbidity fluctuation complexity theoretically ranges from 0 to 1. When the normalized turbidity fluctuation complexity is close to 0, it indicates that the arrangement pattern in the binary sequence is extremely simple, and the turbidity signal fluctuation is highly regular; when it is close to 1, it indicates that the arrangement pattern in the binary sequence is extremely rich, and the turbidity signal fluctuation is highly random.

[0043] Specifically, the interference type determination module is also used to determine that when the pulse density ratio is less than the preset pulse density ratio of 0.08, the abnormal interference type is composite interference and the emulsified oil component is dominant. The interference type determination module is also used to determine that when the pulse density ratio is greater than or equal to the preset pulse density ratio, the abnormal interference type is composite interference and the sediment component is dominant.

[0044] In this embodiment, the preset pulse density ratio was set to 0.08. Mixed water samples with different proportions of emulsified oil and sediment were prepared (emulsified oil volume percentages of 10%, 30%, 50%, 70%, and 90%, with the remainder being sediment; the total suspended solids concentration was uniformly 80 mg / L). Under the same conditions of pipeline flow velocity of 0.5 m / s, detection tank inner diameter of 25 mm, and sampling frequency of 50 Hz, each proportion was independently tested five times with a single discharge volume of 500 mL and a discharge duration of approximately 15 seconds. The experimental results showed that when the emulsified oil content was above 70%, the pulse density ratio was below 0.06; when the emulsified oil content was below 30%, the pulse density ratio was above 0.10; and when the emulsified oil content was between 30% and 70%, the pulse density ratio fluctuated between 0.06 and 0.10. Using 0.08 as the preset pulse density ratio effectively distinguishes between combined interference dominated by emulsified oil (sparse pulses) and sediment-dominated (dense pulses).

[0045] Specifically, when the turbidity fluctuation complexity is greater than or equal to the first preset turbidity fluctuation complexity and less than the second preset turbidity fluctuation complexity, the interference type determination module introduces the pulse density ratio for secondary determination. If the pulse density ratio is less than the preset pulse density ratio, it indicates that high-frequency pulses are sparse in the turbidity signal, rigid collisions are infrequent, the turbidity fluctuation is mainly in a smooth mode, the particulate matter in the water is mainly flexible emulsified oil droplets, whose deformation behavior in the fluid absorbs turbulent energy and is not prone to producing sharp pulse jumps, and rigid silt particles account for only a minority. Based on this, the abnormal interference type is determined to be complex interference with emulsified oil components as the dominant component. If the pulse density ratio is greater than or equal to the preset pulse density ratio, it indicates that high-frequency pulses are dense in the turbidity signal, rigid collisions are frequent, the particulate matter in the water is mainly rigid silt particles, whose hard collisions with the fluid and the wall generate a large number of high-frequency pulse jumps, and flexible emulsified oil droplets account for only a minority. Based on this, the abnormal interference type is determined to be complex interference with silt components as the dominant component.

[0046] Specifically, the abnormal warning module responds to the abnormal interference type being emulsified oil interference and generates a first warning signal. The first warning signal includes an emulsified oil leakage warning and an indication to use demulsification treatment. At this time, the turbidity fluctuation complexity is low, the fluctuation pattern is highly regular, the particulate matter in the water is dominated by flexible emulsified oil droplets, and the false increase in the suspended matter measurement value is mainly caused by emulsified oil. The abnormal early warning module responds to the abnormal interference type as sediment scattering interference and generates a second early warning signal. The second early warning signal includes a suspended solids exceeding the standard warning and an indication to use coagulation and sedimentation treatment. At this time, the turbidity fluctuation is highly complex and the fluctuation pattern is highly random. The particulate matter in the water is dominated by rigid sediment, and the increase in the suspended solids measurement value mainly comes from real suspended solids particles. The abnormal early warning module responds to the abnormal interference type being a complex interference with emulsified oil as the dominant component, and generates a third early warning signal. The third early warning signal includes a first complex pollution warning and an indication that demulsification treatment should be the main treatment and coagulation and sedimentation treatment should be the auxiliary treatment. At this time, the turbidity fluctuation complexity falls into the fuzzy range and the pulse density ratio is low, with local pulses being sparse, indicating that emulsified oil is dominant in the complex interference. The abnormal early warning module responds to the abnormal interference type being a complex interference with sediment as the dominant component, and generates a fourth early warning signal. The fourth early warning signal includes a second complex pollution warning and an indication that coagulation and sedimentation treatment should be the main treatment, with demulsification treatment as a supplement. At this time, the turbidity fluctuation complexity falls into the fuzzy range and the pulse density ratio is high, with local pulses being dense, indicating that sediment is the dominant component in the complex interference.

[0047] Specifically, the fourth data determination unit includes: The differential calculation subunit is used to calculate the first-order difference sequence of the turbidity signal within the event window, wherein the first-order difference sequence is formed by sequentially arranging the differences between adjacent sampling points in the turbidity signal; The pulse threshold determination subunit is used to calculate the standard deviation of the first-order difference sequence of the turbidity signal during the steady-state period before the event, and to determine the preset multiple of the standard deviation, 3, as the pulse threshold. The pulse statistics subunit is used to count the number of differences in the first-order difference sequence whose absolute value is greater than the pulse threshold, thereby obtaining the number of pulses. The density ratio calculation subunit is used to determine the pulse density ratio as the ratio of the number of pulses to the total number of differences in the first-order difference sequence.

[0048] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-parameter intelligent early warning system for oil suspension in water, characterized in that, include: The data acquisition module is used to acquire the fluorescence signal of the ultraviolet fluorescence channel and the turbidity signal of the infrared turbidity channel; The data determination module, which is connected to the data acquisition module, includes a first data determination unit for determining the turbidity sliding standard deviation of the turbidity signal, a second data determination unit for determining the turbidity peak attenuation ratio of the turbidity signal, a third data determination unit for determining the turbidity fluctuation complexity of the turbidity signal, and a fourth data determination unit for determining the pulse density ratio of the turbidity signal. An anomaly detection module, which is connected to the data determination module, is used to determine whether a suspected anomaly event has occurred based on the turbidity sliding standard deviation of the turbidity signal. An anomaly identification module, which is connected to the data determination module and the anomaly detection module respectively, is used to respond to a suspected anomaly flag and determine the event identification state of the suspected anomaly event based on the turbidity peak attenuation ratio of the turbidity signal, wherein the event identification state is an anomaly event or a non-anomaly event. An interference type determination module, which is connected to the data determination module and the anomaly identification module respectively, determines the abnormal interference type of the abnormal event based on the turbidity fluctuation complexity and pulse density ratio of the turbidity signal. The abnormal interference types include emulsified oil interference, silt interference and composite interference. An abnormality warning module, which is connected to the interference type determination module, is used to generate a corresponding warning signal according to the abnormal interference type. The warning signal includes a first warning signal, a second warning signal, a third warning signal, and a fourth warning signal.

2. The multi-parameter intelligent early warning system for oil suspension in water according to claim 1, characterized in that, The anomaly detection module determines that no anomaly has occurred if the turbidity sliding standard deviation of the turbidity signal is less than the preset turbidity sliding standard deviation. The anomaly detection module responds to a turbidity sliding standard deviation being greater than or equal to a preset turbidity sliding standard deviation by determining that a suspected anomaly event has occurred, recording the current moment as the event starting point, and generating a suspected anomaly flag.

3. The multi-parameter intelligent early warning system for oil suspension in water according to claim 2, characterized in that, The first data determining unit includes: The window setting subunit is used to set the preset sliding window length and the preset sliding step size. The preset sliding window length is the number of sampling points contained in the window. The standard deviation calculation subunit is used to calculate the standard deviation of the turbidity signal sequence within the window after each sliding window, starting from the first sampling point and using a preset step size, to obtain the turbidity sliding standard deviation.

4. The multi-parameter intelligent early warning system for oil suspension in water according to claim 3, characterized in that, The anomaly identification module is used to respond to the suspected anomaly flag and determine the event identification status of the suspected anomaly event based on the turbidity peak attenuation ratio of the turbidity signal, wherein... If the turbidity peak attenuation ratio meets the preset attenuation condition, the event identification state is determined to be an abnormal event; if the turbidity peak attenuation ratio does not meet the preset attenuation condition, the event identification state is determined to be a non-abnormal event; the preset attenuation condition is that the turbidity peak attenuation ratio is less than a preset attenuation threshold.

5. The multi-parameter intelligent early warning system for oil suspension in water according to claim 4, characterized in that, The second data determining unit includes: The time period segmentation subunit is used to set the event start point as the reference point, determine the time period within a first preset duration forward from the reference point as the pre-event steady state period, and determine the time period within a second preset duration backward from the reference point as the event window, and respectively acquire the turbidity signal within the pre-event steady state period and the turbidity signal within the event window. A peak determination subunit is used to determine the turbidity peak value in the turbidity signal within the event window; The baseline determination subunit is used to record the arithmetic mean of the turbidity signal sequence during the pre-event steady-state period as the baseline value; A steady-state determination subunit is used to extract a preset third time period from the time corresponding to the turbidity peak within the event window. When the difference between the maximum and minimum values ​​of the turbidity signal within the third time period is less than a preset range value, the arithmetic mean of the turbidity signal sequence within the third time period is determined as the steady-state value. The attenuation ratio calculation subunit is used to calculate the turbidity peak attenuation ratio, which is the ratio of a first difference to a second difference, wherein the first difference is the difference between the turbidity peak and the steady-state value, and the second difference is the difference between the turbidity peak and the baseline value.

6. The multi-parameter intelligent early warning system for oil suspension in water according to claim 5, characterized in that, The interference type determination module determines the abnormal interference type of the abnormal event based on the turbidity fluctuation complexity of the turbidity signal, wherein, If the turbidity fluctuation complexity is less than the first preset turbidity fluctuation complexity, then the abnormal interference type is determined to be emulsified oil interference. If the turbidity fluctuation complexity is greater than or equal to the first preset turbidity fluctuation complexity and less than the second preset turbidity fluctuation complexity, then the abnormal interference type of the abnormal event is determined for the second time based on the pulse density ratio of the turbidity signal. If the turbidity fluctuation complexity is greater than or equal to the second preset turbidity fluctuation complexity, then the abnormal interference type is determined to be sediment scattering interference. The first preset turbidity fluctuation complexity is less than the second preset turbidity fluctuation complexity.

7. The multi-parameter intelligent early warning system for oil suspension in water according to claim 6, characterized in that, The third data determination unit includes: The coarse-grained subunit is used to divide the turbidity signal sequence within the event window into several segments of equal length according to a preset length, and to obtain a coarse-grained sequence based on the arithmetic mean of the turbidity signal in each segment. Binarization subunit is used to generate a binary sequence based on the comparison between the arithmetic mean of each segment in the coarse-grained sequence and the median of the coarse-grained sequence, wherein the corresponding position of the segment whose arithmetic mean is greater than or equal to the median takes a first value, and the corresponding position of the segment whose arithmetic mean is less than the median takes a second value. The complexity calculation subunit is used to scan the binary sequence value by value, record the permutation patterns of the binary sequences that have appeared, and increment the sequence complexity by one whenever a permutation pattern of a binary sequence that has not appeared is scanned, and record the permutation pattern as having appeared. The sequence complexity is obtained after the scan is completed. A normalized subunit is used to determine the turbidity fluctuation complexity as the ratio of the sequence complexity to the number of segments.

8. The multi-parameter intelligent early warning system for oil suspension in water according to claim 7, characterized in that, The interference type determination module is also used to determine that when the pulse density ratio is less than the preset pulse density ratio, the abnormal interference type is composite interference and the emulsified oil component is dominant. The interference type determination module is also used to determine that when the pulse density ratio is greater than or equal to the preset pulse density ratio, the abnormal interference type is composite interference and the sediment component is dominant.

9. The multi-parameter intelligent early warning system for oil suspension in water according to claim 8, characterized in that, The abnormal early warning module responds to the abnormal interference type being emulsified oil interference and generates a first early warning signal, which includes an emulsified oil leakage warning and an indication to use demulsification treatment; The abnormal early warning module responds to the abnormal interference type being sediment scattering interference and generates a second early warning signal. The second early warning signal includes an early warning of excessive suspended matter and an indication to use coagulation and sedimentation treatment. The abnormal early warning module responds to the abnormal interference type being a complex interference with emulsified oil components as the main component, and generates a third early warning signal. The third early warning signal includes a first complex pollution warning and an indication that demulsification treatment should be the main treatment and coagulation and sedimentation treatment should be the auxiliary treatment. The abnormal early warning module responds to the abnormal interference type being a complex interference with sediment as the dominant component, and generates a fourth early warning signal. The fourth early warning signal includes a second complex pollution warning and an indication that coagulation and sedimentation treatment should be the main treatment, with demulsification treatment as a supplement.

10. The multi-parameter intelligent early warning system for oil suspension in water according to claim 9, characterized in that, The fourth data determination unit includes: The differential calculation subunit is used to calculate the first-order difference sequence of the turbidity signal within the event window, wherein the first-order difference sequence is formed by sequentially arranging the differences between adjacent sampling points in the turbidity signal; The pulse threshold determination subunit is used to calculate the standard deviation of the first-order difference sequence of the turbidity signal during the steady-state period before the event, and to determine a preset multiple of the standard deviation as the pulse threshold. The pulse statistics subunit is used to count the number of differences in the first-order difference sequence whose absolute value is greater than the pulse threshold, thereby obtaining the number of pulses. The density ratio calculation subunit is used to determine the pulse density ratio as the ratio of the number of pulses to the total number of differences in the first-order difference sequence.

Images