Lithium battery monorail crane oil quality online detection method

Abstract

The present application relates to a kind of lithium battery monorail crane oil quality online detection method, and oil sample is obtained in the hydraulic oil return of monorail crane online, and the oil sample is pressure and flow regulation;The oil sample after regulation is divided into crude oil sample and reference sample;The reference sample is removed from interference component processing, to remove or weaken the influence of moisture or solid particles on detection, to obtain reference oil;Under the same detection condition, the oil sample and the reference oil are detected respectively, and the crude oil detection signal and reference detection signal are obtained;Based on the crude oil detection signal and the reference detection signal, calculate discriminant, and the discriminant is compared with preset threshold value;When the discriminant is greater than the preset threshold value, output oil quality abnormal alarm;Otherwise, output oil quality normal result.Improve the detection precision and anti-interference ability, avoid false alarm and result fluctuation, improve the continuity and reliability of system output.

Description

Technical Field

[0001] This invention relates to the field of oil condition monitoring and fault diagnosis technology, specifically to an online detection method for oil quality in lithium battery monorail cranes. Background Technology

[0002] Existing technologies, such as patent application CN114755396A, while achieving online access to the oil detection system and comprehensive detection of oil moisture, density, viscosity, dielectric constant, water activity, temperature, and abrasive properties through a six-in-one sensor and a particle counting sensor, and combined with impurity removal components for oil filtration, possess basic online monitoring and early warning capabilities from an overall structural perspective. However, their technical solutions still have significant limitations and shortcomings. First, the detection logic of existing technologies is essentially a single-channel direct detection mode, where the oil flows through various sensors sequentially along a single flow path and outputs detection data. The judgment relies mainly on the detection parameters themselves or simple threshold judgments, lacking an effective internal comparison mechanism. This makes it impossible to distinguish between changes in the detection signal caused by actual oil deterioration and changes caused by interference factors such as moisture and particles, leading to misjudgments or insufficient sensitivity under complex operating conditions. Especially in environments such as downhole monorail hoists, oil is often simultaneously affected by the combined effects of water intrusion and particle contamination, and existing solutions do not separate and analyze interfering components, making it difficult for the detection results to accurately reflect the true quality state of the oil. Secondly, while existing technologies incorporate impurity removal components for oil filtration, this filtration is performed after detection or as a means of improving oil quality. They lack a "before-and-after comparison" detection system and the technical approach of dividing the same oil sample into original and treated samples for comparative testing under identical conditions. Therefore, they cannot utilize differential or ratio methods to eliminate systematic errors and environmental disturbances, nor can they generate stable and reliable discriminant quantities. Detection results are easily affected by sensor drift, temperature fluctuations, and flow rate changes, leading to deviations. Thirdly, existing technologies do not consider time drift during detection. The sensor output signal inevitably experiences zero-point drift or sensitivity changes during long-term operation. However, the system lacks mechanisms for acquiring reference signals before and after detection, calculating drift, and determining drift thresholds. Furthermore, it does not provide retesting or resampling methods to distinguish between instantaneous disturbances and continuous drift. Consequently, the stability and reliability of detection results are difficult to guarantee under conditions of equipment vibration, sudden flow changes, or environmental disturbances.

[0003] Furthermore, existing technologies lack multi-level discrimination and result confirmation mechanisms. Their alarm logic is relatively simple; once the detected parameters exceed the set range, a warning is directly output. They do not introduce strategies based on joint judgment using multiple discriminant parameters, nor do they consider verifying the authenticity of an anomaly through repeated detection or reference signal consistency verification. This easily leads to frequent false alarms or insufficient alarm reliability in practical applications. In addition, existing technologies do not differentiate between different types of contamination. Although they can detect moisture and particulate parameters, they have not established a discrimination model between moisture-dominated and particulate-dominated anomalies, failing to provide clear fault-pointing information for subsequent maintenance and reducing the system's diagnostic value and engineering application significance. Simultaneously, the oil in their system structure always flows along a single path, lacking a design for oil diversion. This prevents the simultaneous construction of multiple detection branches based on the same oil source for comparative analysis, making it difficult for the system to achieve high-precision, interference-resistant detection results. In summary, while existing technologies can achieve basic online oil monitoring and filtration functions, they still have significant shortcomings in terms of anti-interference ability, discrimination accuracy, drift suppression ability, anomaly confirmation mechanism, and contamination type differentiation, making it difficult to meet the demand for high-reliability online detection of oil quality under complex working conditions. Summary of the Invention

[0004] The purpose of this invention is to provide an online detection method for the quality of oil in a lithium battery monorail crane, thereby addressing some of the drawbacks and shortcomings mentioned in the background art.

[0005] The present invention addresses the aforementioned technical problems by employing the following technical solution: an online detection method for the quality of hydraulic fluid in a lithium battery monorail crane, comprising: acquiring an oil sample online in the hydraulic return oil of the monorail crane, and adjusting the pressure and flow rate of the oil sample; dividing the adjusted oil sample into a crude oil sample and a reference sample; and performing interference component removal treatment on the reference sample to remove or reduce the influence of moisture or solid particles on the detection, thereby obtaining reference oil.

[0006] Under the same detection conditions, the crude oil sample and the reference oil are respectively tested for oil properties to obtain crude oil detection signal and reference detection signal; a discrimination value is calculated based on the crude oil detection signal and the reference detection signal, and the discrimination value is compared with a preset threshold: when the discrimination value is greater than the preset threshold, an oil quality abnormality alarm is output; otherwise, a normal oil quality result is output.

[0007] Furthermore, under the same detection conditions, a first reference signal of the reference oil is first acquired, then the crude oil signal of the crude oil sample is acquired, and then a second reference signal of the reference oil is acquired; the discrimination quantity is obtained by performing a difference operation or a ratio operation on the mean or weighted combination of the crude oil signal and the first reference signal and the second reference signal.

[0008] Further, the interference component removal process includes: first performing a moisture removal process to obtain a first reference oil; then performing a solid particle removal process on the first reference oil to obtain a second reference oil; calculating a first discrimination value based on the crude oil sample and the first reference oil, and calculating a second discrimination value based on the crude oil sample and the second reference oil; and outputting a moisture-dominated anomaly alarm or a particle-dominated anomaly alarm based on the difference, ratio, or comparison result of the first discrimination value and the second discrimination value with the corresponding threshold.

[0009] Furthermore, when the discrimination quantity is greater than the preset threshold, the interference component removal process is re-executed and the reference oil detection signal is re-acquired; if the difference between the two reference oil detection signals exceeds the drift threshold, the abnormal alarm is suppressed and a confidence insufficient prompt is output; otherwise, the abnormal oil quality alarm is confirmed and output.

[0010] Furthermore, before calculating the discrimination quantity, a reference drift quantity is first calculated based on the first reference signal and the second reference signal; when the reference drift quantity exceeds a preset drift threshold, it is determined that the current detection does not meet the discrimination condition, and the discrimination result is suppressed, or the first reference signal, crude oil signal and second reference signal are reacquired before the discrimination quantity is calculated.

[0011] Furthermore, the discrimination quantity includes a differential discrimination quantity and a ratio discrimination quantity; the differential discrimination quantity and the ratio discrimination quantity are compared with the corresponding thresholds respectively, and when both meet the preset alarm conditions, an oil quality abnormality alarm is output; otherwise, a normal oil quality result is output.

[0012] Furthermore, when the difference between the two consecutive reference oil detection signals exceeds the drift threshold and an abnormal alarm is suppressed, the current discrimination value is frozen and the previous valid detection result is maintained; the abnormal alarm confirmation output is restored only when the difference between the reference oil detection signals does not exceed the drift threshold within a preset number of consecutive times.

[0013] Furthermore, the difference between the two reference oil detection signals is determined by a combination of differential difference and ratio difference; when any difference exceeds the corresponding drift threshold, an abnormal alarm is suppressed; when both the differential difference and the ratio difference do not exceed the corresponding drift threshold and the discrimination value is still greater than the preset threshold, an abnormal oil quality alarm is confirmed and output.

[0014] Furthermore, when the reference drift exceeds the preset drift threshold, a reference signal retest is first performed to distinguish between instantaneous disturbances and continuous drift; if the reference drift obtained from the retest still exceeds the preset drift threshold, the first reference signal, crude oil signal, and second reference signal are reacquired; if the reference drift obtained from the retest does not exceed the preset drift threshold, the discrimination quantity is recalculated and the discrimination result is output.

[0015] Furthermore, the reference signal retest includes: continuously acquiring at least two reference signals without changing the reference oil generation conditions, and calculating the difference between two adjacent reference signals; when any difference exceeds the preset drift threshold, it is determined to be a continuous drift and triggers the reacquisition of the first reference signal, the crude oil signal, and the second reference signal; when each difference does not exceed the preset drift threshold, it is determined to be an instantaneous disturbance and the discriminant quantity is recalculated.

[0016] The beneficial effects of this invention are as follows: By acquiring oil samples online from the hydraulic return oil of a monorail crane and performing a split-flow process on the samples, the crude oil sample is compared with a reference oil after interference components have been removed under the same detection conditions. This effectively reduces the interference of moisture and solid particles on the detection results, allowing the discriminant to more accurately reflect changes in oil quality. Compared with traditional single-signal detection methods, this invention significantly improves detection accuracy and anti-interference capability by constructing a reference benchmark and performing differential or ratio analysis, enabling stable and reliable online monitoring of oil conditions under complex operating conditions.

[0017] Furthermore, this invention effectively suppresses the impact of sensor drift, transient disturbances, and environmental fluctuations on detection results by introducing front and rear reference signal acquisition, reference drift judgment, retesting mechanism, and joint judgment using both differential and ratio criteria. Simultaneously, by graded removal of moisture and solid particles and the construction of separate discriminant quantities, the invention achieves differentiation of anomaly sources. Combined with anomaly confirmation, freeze, and recovery mechanisms, it avoids false alarms and result fluctuations, improving the continuity and reliability of system output. Attached Figure Description

[0018] Figure 1 This is a functional relationship diagram of the online oil quality detection of the present invention.

[0019] Figure 2 This is a timing and discrimination margin diagram of the detection sequence signal in Embodiment 1 of the present invention.

[0020] Figure 3 This is an enhanced reference drift retest and recovery discrimination diagram in Embodiment 1 of the present invention.

[0021] Figure 4 This is a comparison chart of the two-level reference oil discrimination quantity and threshold in Embodiment 2 of the present invention.

[0022] Figure 5 This is a diagram showing the output status of failure confirmation and freeze in Embodiment 2 of the present invention.

[0023] Figure 6 This is a diagram showing the stable recovery and final confirmation during the freezing period in Embodiment 2 of the present invention. Detailed Implementation

[0024] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0025] Combined with appendix Figure 1 This invention discloses an online method for detecting the quality of hydraulic fluid in a lithium battery monorail crane. An online sampling unit is installed on the hydraulic return oil line of the lithium battery monorail crane, allowing the hydraulic oil to continuously enter the detection branch during operation. To ensure stable test results, the pressure and flow rate of the oil sample entering the detection branch are first regulated, ensuring the sample passes smoothly through the detection system within a set pressure and flow rate range, avoiding signal distortion caused by return oil pulsation, instantaneous impact, or flow rate fluctuations. The regulated oil sample enters a splitting unit, where it is simultaneously divided into a crude oil sample and a reference sample. The crude oil sample retains its original composition in the return oil, characterizing the current actual oil quality. The reference sample then enters an interference component removal unit for processing, reducing the influence of moisture or solid particles on subsequent testing, thereby obtaining reference oil.

[0026] Interfering component removal can be achieved using dehydration components, filtration components, or combinations thereof. Dehydration components remove free water, emulsified water, or reduce the influence of moisture, while filtration components trap metal particles, dust particles, and other suspended impurities. The reference oil obtained after this treatment has a lower interference level than the crude oil sample and can be used as a benchmark for discrimination. To ensure effective comparison, the crude oil sample and reference oil are tested for oil properties under the same testing conditions, including the same detection devices, the same testing time period, the same temperature conditions, and the same flow state. During testing, crude oil detection signals and reference detection signals are obtained separately. These detection signals characterize the oil's electrical, dielectric, impedance, optical, or other properties that reflect quality changes.

[0027] After obtaining the two types of detection signals mentioned above, the control unit calculates a discrimination value based on the crude oil detection signal and the reference detection signal. The discrimination value can be expressed as a difference to characterize the absolute degree of deviation between the two signals, or as a ratio to characterize the relative degree of change. Since the interference from water or solid particles in the reference oil has been removed or reduced, the difference between the crude oil sample and the reference oil can more directly reflect quality changes such as oil contamination, deterioration, or abnormal mixing. When the discrimination value is compared with a preset threshold, if the discrimination value is greater than the preset threshold, it indicates that the crude oil sample has undergone a significant abnormal deviation relative to the reference oil, and the system outputs an oil quality abnormality alarm to prompt maintenance personnel to check for water ingress, wear, contamination, or oil failure in the hydraulic system. If the discrimination value is not greater than the preset threshold, the current oil quality is determined to be within the normal range, and the system outputs a normal oil quality result. By combining online sampling, diversion comparison, removal of interfering components, and detection under the same conditions, crude oil samples are compared with reference oil in real time. This effectively suppresses the influence of simple environmental fluctuations and sensor drift on the detection results, improves the accuracy and practicality of online monitoring of hydraulic oil in monorail cranes, and is suitable for continuously operating equipment under complex downhole conditions.

[0028] To further suppress time drift and instantaneous fluctuations during the detection process, multiple sets of signals were acquired sequentially under the same detection conditions. First, reference oil, after interference component removal treatment, was introduced into the detection unit to acquire the first reference signal. Then, the system switched to the crude oil sample for detection, acquiring the crude oil signal. Finally, the reference oil was reintroduced into the detection unit to acquire the second reference signal. These three detections were performed continuously under the same temperature, flow rate, and pressure conditions to ensure signal comparability.

[0029] After obtaining the first reference signal, the crude oil signal, and the second reference signal, the two reference signals are combined to obtain a reference baseline signal. The reference baseline signal can be the average of the first and second reference signals, or a weighted combination value can be formed by setting weights according to time sequence or system stability, thereby reflecting the baseline change trend during the detection process. Subsequently, the crude oil signal and the reference baseline signal are used to perform calculations to obtain the discrimination quantity.

[0030] The discriminant can be expressed as a differential operation to characterize the absolute deviation of the crude oil signal relative to the reference signal, or as a ratio operation to characterize its relative change. By introducing two reference signals before and after crude oil detection, and using their combined result as a benchmark, the effects of sensor drift, slight temperature changes, or flow fluctuations during the detection process can be effectively eliminated, thus improving the stability and reliability of the discriminant.

[0031] Before calculating the discrimination value, the stability of the reference signal is evaluated. Specifically, a reference drift is calculated based on the first and second reference signals. The reference drift characterizes the degree of change of the same reference oil between two consecutive tests. This drift can be expressed as a difference or a ratio to reflect the baseline fluctuation of the detection system over a short period of time. The reference drift is compared with a preset drift threshold. When the reference drift does not exceed the preset drift threshold, it indicates that the detection conditions are stable, and the subsequent discrimination value calculation process can proceed. When the reference drift exceeds the preset drift threshold, it indicates that there is significant drift or disturbance during the detection process, and the current detection is determined not to meet the discrimination conditions.

[0032] If the discrimination criteria are not met, the system suppresses the output of the current discrimination result to avoid misjudgment. Simultaneously, a re-detection process can be triggered as needed, i.e., the first reference signal, crude oil signal, and second reference signal are reacquired, and the discriminant is recalculated under a new detection sequence, thereby improving the reliability of the detection results.

[0033] After verification using the reference drift, discriminant parameters are further calculated. These parameters include differential discriminant and ratio discriminant. The differential discriminant characterizes the absolute difference between the crude oil signal and the reference signal, while the ratio discriminant characterizes the relative change between the two. The differential discriminant and the ratio discriminant are compared to their corresponding thresholds. Only when both parameters simultaneously meet preset alarm conditions is an abnormality in oil quality identified and an alarm is output. If either parameter fails to reach its corresponding threshold, the oil quality is considered within the normal range, and a normal result is output. This dual-criteria approach effectively reduces the risk of misjudgment associated with a single indicator.

[0034] When the reference drift calculated based on the first and second reference signals exceeds a preset drift threshold, a reference signal retesting process is executed first to distinguish between instantaneous disturbances and continuous drift, in order to avoid misjudgment due to occasional disturbances. Specifically, while keeping the reference oil generation conditions unchanged, without altering the oil circuit state, temperature conditions, and flow and pressure parameters, at least two reference signals are continuously acquired, and the difference between two adjacent reference signals is calculated to obtain the drift situation during the retesting stage.

[0035] During the retest, if the difference between any two adjacent reference signals exceeds a preset drift threshold, the current detection system is deemed to have persistent drift or insufficient stability. In this case, a re-detection process is triggered, involving the reacquisition of the first reference signal, crude oil signal, and second reference signal. The discriminant is then recalculated under the new detection sequence to ensure the reliability of the detection benchmark. If the difference between any two reference signals during the retest does not exceed the preset drift threshold, the previous drift is considered a transient disturbance, indicating overall stability of the detection system. In cases where a transient disturbance is identified, the discriminant calculation process resumes, calculating the discriminant based on the acquired crude oil signal and reference signal combination, and outputting the corresponding discriminant result.

[0036] To further differentiate the sources of oil quality anomalies, the reference sample underwent multi-stage interference component removal treatment. First, the reference sample was introduced into a moisture removal unit, where free water and emulsified water were removed from the oil or their effects were significantly reduced through a dehydration component, thus obtaining the first reference oil. This first reference oil, while maintaining the basic components of the crude oil, reduced moisture interference and can be used as a benchmark reflecting the suppression of moisture factors.

[0037] The first reference oil is then introduced into the solid particle removal unit, where metal abrasive particles, dust particles, and other suspended impurities are removed through a precision filtration assembly to obtain the second reference oil. This second reference oil, with reduced moisture and particle interference, further approximates the ideal baseline state.

[0038] During the detection process, detection signals from the crude oil sample and the first reference oil were acquired under the same detection conditions, and a first discriminant was calculated to reflect the signal difference before and after moisture removal. Simultaneously, detection signals from the crude oil sample and the second reference oil were acquired, and a second discriminant was calculated to reflect the combined difference after both moisture and particles were removed. By comparing the first and second discriminant values, the influence of different interfering components on the detection results can be analyzed.

[0039] Specifically, the judgment is made based on the difference or ratio between the first and second discriminant values, or by comparing them with the corresponding thresholds. When the first discriminant value is significantly greater than the second discriminant value or exceeds the corresponding threshold first, it indicates that moisture removal has a more significant impact on signal changes, and a moisture-dominated anomaly can be identified, with a corresponding alarm output. When the second discriminant value still shows a large change relative to the first discriminant value or exceeds the corresponding threshold, it indicates that solid particles contribute significantly to the signal, and a particle-dominated anomaly can be identified, with a corresponding alarm output.

[0040] When the discrimination value calculated based on the crude oil signal and the reference signal exceeds a preset threshold, the system initially determines that the oil quality is abnormal. To avoid false alarms due to unstable reference standards or insufficient interference removal, a confirmation process is performed before outputting an abnormality alarm. Specifically, the reference sample is re-processed to remove interfering components, and the reference oil detection signal is acquired again under the same detection conditions to form two reference oil detection results.

[0041] Subsequently, a difference analysis is performed on the two reference oil detection signals to assess the stability of the reference benchmark. The difference can be calculated in the form of differential or ratio and compared with a preset drift threshold. When the difference between the two reference oil detection signals exceeds the drift threshold, it indicates that there are unstable factors in the reference oil generation or detection process. At this time, the output of the current abnormal alarm is suppressed, and a confidence deficiency prompt is sent to the upper-level system to indicate that the current detection result does not meet the reliable judgment conditions.

[0042] When the difference between two consecutive reference oil detection signals does not exceed the drift threshold, it indicates that the reference standard is stable and the removal of interfering components is consistent. At this point, the anomaly reflected by the aforementioned discrimination quantity is confirmed as a true and valid result, and the system outputs an oil quality anomaly alarm. By introducing a re-verification mechanism, false alarms caused by reference fluctuations can be effectively reduced.

[0043] When the difference between two reference oil detection signals obtained through re-inspection exceeds a preset drift threshold and abnormal alarms are suppressed accordingly, the current discrimination value is frozen to prevent unstable system output. Simultaneously, the valid detection result from the previous cycle is retained as the current output. Freezing means not updating the discrimination result and alarm status within the current detection cycle, thereby avoiding frequent output jumps due to reference baseline fluctuations and improving the continuity and stability of system operation.

[0044] During subsequent testing, the system continuously performs reference oil detection and calculates the difference between two consecutive reference oil detection signals. When the difference between the reference oil detection signals does not exceed the drift threshold for a preset number of consecutive tests, the system is deemed to have returned to a stable state. At this point, the frozen state is lifted, the confirmation output logic for abnormal alarms is restored, and the abnormality is re-evaluated based on the latest discrimination result.

[0045] Furthermore, to improve the reliability of reference signal stability determination, the difference between two consecutive reference oil detection signals is determined using a combination of differential difference and ratio difference. Differential difference characterizes the absolute change in the signal, while ratio difference characterizes the relative change. Both differences are compared to their corresponding drift thresholds. If either difference exceeds its corresponding threshold, the reference signal is determined to have abnormal fluctuations, and an abnormal alarm is suppressed. When both differential difference and ratio difference do not exceed their corresponding drift thresholds, and the discrimination value is still greater than a preset threshold, it indicates that the reference benchmark is stable and the abnormality determination is valid. The system then confirms and outputs an oil quality abnormality alarm. This combination of dual difference determination and a freeze-recovery mechanism significantly improves the stability and reliability of the detection results.

[0046] Example 1:

[0047] In this embodiment, a lithium battery monorail transport line for transferring hydraulic support components is arranged in the east wing transport roadway of a coal mine. The monorail uses a hydraulically driven clamping and braking mechanism. After running continuously for 6 hours during the night shift, the equipment began to perform a series of alternating transport tasks on uphill and curved sections. During this shift, a total of 18 transports of support connectors and pin assemblies were carried out, with a single load ranging from 0.8t to 1.1t. At 6 hours and 12 minutes into the shift, the driver reported a short-term impact when the vehicle passed through the rail joint, accompanied by increased pulsation in the return oil pipeline, a local humidity increase to 89% underground, and a rise in hydraulic oil temperature from 42.6℃ at the beginning of the shift to 46.2℃. To avoid false alarms caused by such instantaneous disturbances in the online monitoring results, this embodiment employs a method of same-source sample diversion, interference-free reference, reference drift retesting, and dual-criteria joint judgment to conduct online detection of the hydraulic return oil quality of the monorail.

[0048] In this embodiment, an online sampling branch is installed on the hydraulic return oil main pipe. The sampling location is between the front end of the return oil filter and the oil tank return inlet, so that the sampled oil can accurately reflect the return oil status. During the 6th hour and 12 minutes to the 6th hour and 15 minutes of the night shift, the measured pressure of the return oil main pipe fluctuated between 0.47 MPa and 0.52 MPa. After the sampling branch passes through the flow limiting valve, pressure stabilizing valve, and micro buffer chamber in sequence, the pressure of the detection branch is stabilized at 0.21 MPa, the flow rate of the detection branch is stabilized at 0.78 L / min, and the flow rate fluctuation is less than 3.0%. Under these stable pressure and flow conditions, the oil sample continuously enters the detection system, thereby reducing signal distortion caused by return oil pulsation, instantaneous impact, and flow rate fluctuations.

[0049] After pressure and flow stabilization, the oil sample enters the splitting unit and is simultaneously divided into a crude oil sample and a reference sample. The crude oil sample, maintaining its actual composition in the return oil, is directly sent to the detection unit. The reference sample first passes through a dehydration component to remove free water and reduce the influence of emulsion water, and then through a fine filtration component to reduce interference from suspended particles, yielding reference oil. The detection unit uses the same set of dielectric impedance composite sensors, with continuously set detection periods. The detection cell temperature is maintained between 46.0℃ and 46.3℃ through a constant temperature jacket to ensure that the crude oil sample and reference oil are under the same detection conditions. The system completes a full detection sequence in the order of the first reference signal, the crude oil signal, and the second reference signal, with the total detection time controlled within 65 seconds.

[0050] In this detection sequence, the first reference signal is obtained when the reference oil is introduced for the first time. Then switch to crude oil sample to obtain crude oil signal. Then the reference oil is reintroduced into the detection unit to obtain a second reference signal. Considering that the first reference detection is closer to the moment of the track joint impact and is prone to residual pulsation disturbances, while the second reference detection is closer to the moment of crude oil detection and can more accurately reflect the baseline state corresponding to this judgment, this embodiment assigns a higher weight to the second reference signal and selects a weighting coefficient. The reference signal is calculated using the following formula:

[0051]

[0052] Substituting the data from this embodiment, we get:

[0053]

[0054]

[0055] Therefore, the reference signal corresponding to this detection is 1.30372. This value is closer to the second reference signal. For a clearer visual representation of the relationships between the first reference signal, crude oil signal, second reference signal, and reference signal in this detection sequence, please refer to the appendix. Figure 2 From the appendix Figure 2 It can be seen that although the crude oil signal is higher than the two reference signals, the reference baseline signal is still closer to the second reference signal, reflecting that more emphasis is placed on the reference value at the later time adjacent to the crude oil detection time in terms of timing. The deviation of the crude oil signal relative to the reference baseline has not exceeded the differential threshold safety zone, and the corresponding ratio discrimination value is still lower than the preset ratio threshold.

[0056] Before calculating the oil quality discrimination metric, the stability of the two reference signals is evaluated. The reference drift is calculated using the following formula:

[0057]

[0058] The corrected quantity in the formula is taken Substituting the data from this embodiment, we get:

[0059]

[0060]

[0061] Right now

[0062]

[0063] This embodiment presets a drift threshold. Therefore, the reference drift of the detection sequence in this round, 2.569%, is slightly higher than the drift threshold of 2.40%. If the detection condition is directly judged to be unstable and an abnormal correlation conclusion is output at this time, there is a risk of mistaking the instantaneous disturbance for a continuous drift. Therefore, the system does not immediately output the discrimination result, but enters the reference signal retest process to distinguish between instantaneous disturbance and continuous drift.

[0064] During the retest, the reference oil generation conditions remained unchanged, and the oil circuit structure, detection branch pressure, and detection branch flow rate were not altered. Reference signals were continuously acquired under the same temperature conditions. The first retest yielded a reference signal of 1.301, the second yielded 1.309, and the third yielded 1.304. The system calculated the drift between adjacent retest signals. The drift of the first group of adjacent retest signals was taken as the ratio of its absolute difference to the corresponding equilibrium benchmark, with a calculated result of approximately 0.613%. The drift of the second group of adjacent retest signals was approximately 0.383%, both significantly lower than the preset drift threshold of 2.40%. This indicates that the large difference between the first and second reference signals primarily stemmed from short-term benchmark drift caused by the impact of the track joint superimposed on the instantaneous fluctuations in downhole humidity, rather than continuous instability of the detection system. Therefore, this embodiment classifies this drift as an instantaneous disturbance and resumes the discriminant calculation process without triggering the reacquisition of the complete first reference signal, crude oil signal, and second reference signal sequences. The entire process described above, from the initial drift exceeding the threshold, entering the retest phase, to continuous retesting, recovery to stability, and re-entering the discrimination phase, can be found in the appendix. Figure 3 From the appendix Figure 3 As can be seen, the initial drift was above the threshold, while the drift in the subsequent two sets of retests fell back below the threshold. This indicates that the system identified and processed the instantaneous disturbance and did not directly determine the short-term offset caused by a one-time impact as continuous instability, thus ensuring the reliability of subsequent judgments.

[0065] After recalculating the discriminant, this embodiment uses a combination of differential discriminant and ratio discriminant to determine the oil quality status. Based on the aforementioned detection results, the absolute difference between the crude oil signal and the reference signal is 0.03228, which is used as the differential discriminant. Comparing this with the differential threshold of 0.045, it can be seen that 0.03228 is less than 0.045, and the differential alarm condition is not met. Further, dividing this absolute difference by the reference signal yields the ratio discriminant.

[0066]

[0067] That is, 2.476%. In this embodiment, the ratio threshold is set to 3.20%, therefore 2.476% does not meet the ratio alarm condition. Since the differential discrimination quantity and the ratio discrimination quantity do not simultaneously meet the preset alarm conditions, no oil quality abnormality alarm is output in this round of detection. (Appendix) Figure 2 The results were also visualized using differential margin and ratio margin. As can be seen from the figure, although there is an identifiable offset between the crude oil signal and the reference benchmark, the offset is still within the differential threshold safety zone and the allowable range of ratio discrimination. Therefore, the detection results of this round should be judged as normal rather than abnormal.

[0068] To further illustrate the rationality of the dual-criteria conclusion, this embodiment introduces a joint discriminant index as an auxiliary analytical measure. The joint discriminant index is calculated using the following formula:

[0069]

[0070] in , ,and Substituting the data from this embodiment, we first calculate the first term:

[0071]

[0072]

[0073] Calculate the second term:

[0074]

[0075]

[0076] Therefore, we can conclude that:

[0077]

[0078]

[0079] Right now

[0080]

[0081] In this embodiment, the joint indicator threshold ,therefore This result further illustrates, from a comprehensive deviation perspective, that the deviation of the crude oil sample from the reference benchmark is still within the normal fluctuation range. It should be noted that this joint discrimination index is only used to assist in explaining why neither the differential discrimination quantity nor the ratio discrimination quantity reached the alarm threshold in this round of testing. The final output logic still requires both the differential discrimination quantity and the ratio discrimination quantity to simultaneously meet the threshold conditions to constitute an abnormal alarm. (See attached...) Figure 2 The distribution of the detection sequence signal and the discriminant margin relationship shown can more clearly show that the fundamental reason for the small joint index is that the absolute and relative deviations between the crude oil signal and the reference benchmark do not exceed the judgment boundaries set in the embodiment. Therefore, the auxiliary analysis results corroborate the final conclusion of the dual criteria, rather than replacing the dual criteria themselves.

[0082] Based on the on-site operation records, the main reason for the slightly higher reference drift in the first test was the short-term impact load caused by the fully loaded monorail passing through the rail joint, which instantaneously amplified the return oil pulsation. Simultaneously, the localized high humidity environment in the roadway caused a slight shift in the output benchmark after the reference sample was processed to remove interference. Directly rejecting the test results based solely on the difference between the first and second reference signals, or ignoring drift verification and making a rough comparison between the crude oil signal and the offset reference signal, could lead to misjudgment. This embodiment, by continuing to perform at least two reference signal retests after the reference drift exceeds the threshold, confirms that the differences between subsequent adjacent reference signals have all returned to within the threshold, thereby accurately identifying the drift as an instantaneous disturbance rather than a continuous drift, avoiding false triggering of re-detection and false alarms.

[0083] Example 2:

[0084] In this embodiment, a lithium battery-powered monorail transport system is installed in the West Third Transport Roadway of a mine. This system is used to transport cable reels, support components, and connecting parts between the fully mechanized mining preparation area and the material storage point. The system has been running continuously for four shifts, accumulating over 31 hours of operation. After the night shift handover, during a heavy-load uphill transport operation, the monorail, under near-rated load, pulled a large-diameter cable reel and two sets of support components from the lower platform to the higher junction. During this operation, the load on the hydraulic clamping and braking circuits continuously increased. The shift's inspection personnel discovered slight emulsification of the oil in the return oil observation window, and a slight increase in fine metal particles adhering to the magnetic return oil collection component compared to the previous shift. Therefore, it is necessary to determine whether the current abnormal oil quality is primarily due to moisture or solid particles, thus providing a basis for subsequent maintenance.

[0085] In this embodiment, an online sampling unit is installed on the main hydraulic return oil pipe of the monorail crane, with the sampling point located between the main return oil pipe and the oil tank return inlet. During the corresponding detection period, the pressure of the main return oil pipe fluctuates between 0.54 MPa and 0.59 MPa. After the sampled oil enters the detection branch, it is successively regulated by a buffer throttling component, a pressure stabilizing component, and a constant flow component. The pressure of the detection branch is stabilized at 0.24 MPa, the flow rate of the detection branch is stabilized at 0.92 L / min, and the oil temperature in the detection pool is maintained between 49.1℃ and 49.6℃. After the above adjustments, the oil sample is divided into a crude oil sample and a reference sample. The crude oil sample directly retains the actual return oil state, while the reference sample enters a multi-stage interference component removal process.

[0086] The reference sample first enters the moisture removal unit, where the influence of free water and emulsified water on the detection signal is weakened by a hydrophobic membrane dehydration component and a coalescing separation structure, resulting in the first reference oil. Subsequently, the first reference oil continues to the solid particle removal unit, passing sequentially through a precision filter component to remove metal abrasive particles, dust particles, and other suspended impurities, resulting in the second reference oil. Since the detection signal used in this embodiment is a normalized composite output after coupling dielectric and impedance responses, increased moisture significantly increases the polarization component, while metal abrasive particles generate additional disturbances to the impedance channel. Therefore, under actual operating conditions, the deviation between the first reference oil after dehydration and the crude oil signal reflects the degree of change after the moisture factor is weakened. The deviation between the second reference oil obtained after further particle removal and the crude oil signal reflects the overall change after both moisture and particles are weakened. Based on this, the discriminant relationship between the crude oil sample and the two-stage reference oil can be used to identify the source of anomalies.

[0087] During this testing cycle, the testing unit acquired crude oil signals, a first reference oil signal, and a second reference oil signal under the same temperature, pressure, flow rate, and circulation conditions. The measured crude oil signal... First reference oil signal Second reference oil signal Correct measurement The first and second discriminants are calculated using the following formulas:

[0088]

[0089]

[0090] Substituting the data from this embodiment into the formula for the first discriminant, we obtain:

[0091]

[0092]

[0093] Right now

[0094]

[0095] Substituting into the formula for the second discriminant, we get:

[0096]

[0097]

[0098] Right now

[0099]

[0100] In this embodiment, the first discrimination threshold is set to 8.80%, and the second discrimination threshold is set to 7.20%. It can be seen that the first discrimination value of 10.778% exceeds its corresponding threshold, and the second discrimination value of 7.715% also exceeds its corresponding threshold. However, the first discrimination value is significantly greater than the second discrimination value, indicating that in this comprehensive detection response, the signal change caused by removing moisture is more significant, while the additional change caused by further removing solid particles is relatively small. Therefore, it is preliminarily determined that the oil quality anomaly is mainly caused by moisture, with particle factors being a secondary rather than the primary factor. For a more intuitive demonstration of the discrimination relationship between the two levels of reference oil, please refer to the appendix. Figure 4 Appendix Figure 4 The first discriminant, the second discriminant, and their corresponding thresholds were compared and displayed, as shown in the figure. and All exceeded their respective thresholds, but Significantly higher This more intuitively illustrates that the contribution of moisture factors was more prominent in this anomaly.

[0101] After the initial determination is made, the system does not immediately output an anomaly alarm. Instead, it re-executes the multi-stage interference component removal process of the reference sample and re-acquires the reference oil detection signal using the second reference oil as the confirmation benchmark. The second reference oil is chosen as the confirmation object because it has simultaneously reduced moisture and particle interference, making it closer to the ideal reference state and suitable as the benchmark oil sample for the anomaly confirmation stage. In this embodiment, the two confirmation detection signals after re-preparing the second reference oil are denoted as follows: and To characterize the overall variability between two confirmatory tests, a reference confirmatory difference index is constructed as follows:

[0102]

[0103] Where the weighting coefficient is taken , Substituting the data from this embodiment into the above formula yields:

[0104]

[0105]

[0106]

[0107]

[0108] In this embodiment, the drift threshold is confirmed. Therefore, there is This indicates that the re-prepared second reference oil showed significant fluctuations between the two confirmation tests, indicating instability of the reference standard. Furthermore, the system simultaneously determined the difference based on both differential difference and ratio difference. The differential difference between the two confirmation signals was 0.048, exceeding the differential drift threshold of 0.030, and the ratio difference was...

[0109]

[0110] That is, 3.598%, which is greater than the ratio drift threshold of 2.60%. Since the combined judgment results of the difference and ratio difference both point to instability of the reference benchmark, this anomaly alarm is suppressed. The process from initial anomaly detection to failure confirmation and output freeze can be found in the appendix. Figure 5 Appendix Figure 5 This comprehensively illustrates the relationship between system state evolution, joint confirmation difference indicators, differential differences, ratio differences, and corresponding thresholds. (See appendix.) Figure 5 As can be seen, the joint confirmation difference index C=0.04345 is higher than the threshold of 0.032, the differential difference of 0.048 is higher than the threshold of 0.030, and the ratio difference of 3.598% is also higher than the threshold of 2.60%. This indicates that the confirmation stage does not meet the conditions for stable output of abnormal alarms. Therefore, the processing logic of the system entering the frozen state is reasonable.

[0111] While suppressing abnormal alarms, the system freezes the current discrimination value and retains the output result of the previous valid detection cycle as the current output. In this embodiment, the result of the previous valid detection cycle was that the oil quality was normal. Therefore, although the first and second discrimination values ​​have initially indicated a water-dominated abnormality in this cycle, the system does not update the current alarm status due to the instability of the second reference oil confirmation signal, and continues to use the normal output of the previous cycle, thereby avoiding frequent jumps in monitoring results between adjacent cycles due to fluctuations in the reference benchmark. The on-site cause of this confirmation failure, after subsequent tracing and verification of operating conditions, was mainly related to the shear heating of the return oil after heavy uphill loading, the incomplete stabilization of the emulsion state, and the short-term flow redistribution within the filter branch. Appendix Figure 5The state changes shown also indicate that this embodiment does not directly output an anomaly when the initial judgment result meets the alarm conditions. Instead, it actively suppresses the abnormal output after the baseline is found to be unstable during the confirmation process, and maintains the valid result of the previous cycle by freezing, thereby reducing the risk of false alarms.

[0112] Even in a frozen state, the system does not cease detection but continues to monitor the stability of the newly prepared second reference oil. To determine whether the system has stabilized, this embodiment continuously acquires the second reference oil detection signal four times, denoted as follows: , , and The number of consecutive stability determinations is set to 3, and the difference between two adjacent signals is denoted as follows: , and The stable recovery index is calculated using the following formula:

[0113]

[0114] Substituting the data from this embodiment, we get:

[0115]

[0116]

[0117]

[0118]

[0119] Right now

[0120]

[0121] In this embodiment, the recovery threshold is set to 0.30%. Because... This indicates that the second reference oil signal, continuously detected during the freeze period, has returned to a stable state, and the detection system meets the conditions for unfreezing. Accordingly, the system unfreezes, resumes the abnormal alarm confirmation output logic, and enters a new confirmation detection cycle. The stable changes of the second reference oil signal during the freeze period and the relationship between the recovery index and the threshold can be found in the appendix. Figure 6 Appendix Figure 6 The second reference oil detection signal was analyzed four times consecutively during the freezing period, along with adjacent stable differences and recovery indicators. A comprehensive presentation was given. (Attached) Figure 6 It is evident that the fluctuations of the second reference oil signal detected continuously during the freezing period were small, and the calculated recovery index was only 0.195%, which is significantly lower than the recovery threshold of 0.30%. This indicates that the detection system has recovered from the unstable state of the previous stage to a stable state where the output can be reconfirmed.

[0122] In the next complete detection cycle after the freeze was lifted, the crude oil signal and the two-level reference oil signals were reacquired, and the results were consistent with the aforementioned trend. During this cycle, the crude oil signal was 1.460, the first reference oil signal was 1.320, and the second reference oil signal was 1.357. Using the same method, the first discriminant was calculated as follows:

[0123]

[0124] That is, 10.463%, and the second discriminant is

[0125]

[0126] That is, 7.491%. Both still exceed the corresponding thresholds, and the first discriminant remains significantly greater than the second discriminant, consistently pointing to a moisture-dominated anomaly. Subsequently, two confirmation tests were performed on the newly prepared second reference oil, obtaining two confirmation signals as follows: and At this point, the difference is 0.002, which is lower than the difference drift threshold of 0.030, and the ratio difference is...

[0127]

[0128] That is, 0.148%, which is lower than the ratio drift threshold of 2.60%. If we continue to calculate according to the aforementioned confirmed difference index, then...

[0129]

[0130]

[0131] As can be seen, this value is also far below the drift confirmation threshold of 0.032, indicating that the second reference oil benchmark is stable at this time, meeting the anomaly confirmation conditions. (Appendix) Figure 6 The confirmation result after recovery also provides an intuitive reflection: after the conditions for unfreezing are met through continuous stability monitoring during the recovery phase, the system re-enters the confirmation output logic and, in the next detection cycle, it again shows that the first discriminant is significantly greater than the second discriminant and the confirmation difference index is far below the threshold. Therefore, it can support the final output of a moisture-dominated anomaly alarm.

[0132] Therefore, this embodiment, through online sampling, pressure and flow stabilization, separation of crude oil samples and reference samples, multi-stage removal of moisture and solid particles from the reference samples, and the formation of a first reference oil and a second reference oil, distinguishes the source of anomalies based on the discrimination relationship between crude oil and the two-stage reference oil. This enables effective identification of moisture-dominated and particle-dominated anomalies under conditions of continuous downhole operation and heavy load disturbances. Furthermore, by adding a second reference oil confirmation detection after the initial anomaly detection, and suppressing anomaly alarms, freezing the current discrimination quantity, and using the previous valid output when the confirmation signal is unstable, and then unfreezing and restoring the confirmation output after several consecutive stable second reference oil signals, the risk of false alarms caused by reference benchmark fluctuations can be effectively reduced.

[0133] Therefore, this embodiment ultimately confirmed that the output moisture was the primary cause of the abnormal alarm. Combined with subsequent on-site maintenance results, it was found that the aging of the fuel tank breather seal allowed high-humidity air to enter, which, combined with the increased oil temperature caused by the heavy uphill operation during that shift, made moisture in the oil the main source of the abnormality. While metal abrasive particles increased, they did not yet constitute the dominant factor.

Claims

1. A method for online detection of oil quality in a lithium battery monorail crane, characterized in that... include: Online oil samples are obtained from the hydraulic return oil of the monorail crane, and the pressure and flow rate of the oil samples are adjusted. The adjusted oil samples were divided into crude oil samples and reference samples; The reference sample is subjected to interference component removal treatment to remove or reduce the influence of moisture or solid particles on the detection, thereby obtaining reference oil; Under the same detection conditions, the crude oil sample and the reference oil were respectively tested for oil properties to obtain crude oil detection signal and reference detection signal; A discrimination value is calculated based on the crude oil detection signal and the reference detection signal, and the discrimination value is compared with a preset threshold: when the discrimination value is greater than the preset threshold, an oil quality abnormality alarm is output; otherwise, a normal oil quality result is output.

2. The method for online detection of oil quality in a lithium battery monorail crane according to claim 1, characterized in that: Under the same detection conditions, a first reference signal of the reference oil is first acquired, then the crude oil signal of the crude oil sample is acquired, and then a second reference signal of the reference oil is acquired; the discrimination quantity is obtained by performing a difference operation or a ratio operation on the mean or weighted combination of the crude oil signal and the first and second reference signals.

3. The method for online detection of oil quality in a lithium battery monorail crane according to claim 1, characterized in that: The interference component removal process includes: first performing a moisture removal process to obtain a first reference oil; then performing a solid particle removal process on the first reference oil to obtain a second reference oil; calculating a first discrimination value based on the crude oil sample and the first reference oil, and calculating a second discrimination value based on the crude oil sample and the second reference oil; and outputting a moisture-dominated anomaly alarm or a particle-dominated anomaly alarm based on the difference, ratio, or comparison result of the first discrimination value and the second discrimination value with the corresponding threshold.

4. The method for online detection of oil quality in a lithium battery monorail crane according to claim 1, characterized in that: When the discrimination value is greater than the preset threshold, the interference component removal process is re-executed and the reference oil detection signal is re-acquired; if the difference between the two reference oil detection signals exceeds the drift threshold, the abnormal alarm is suppressed and a confidence insufficient prompt is output; otherwise, the abnormal oil quality alarm is confirmed and output.

5. The method for online detection of oil quality in a lithium battery monorail crane according to claim 2, characterized in that: Before calculating the discrimination value, the reference drift value is first calculated based on the first reference signal and the second reference signal; When the reference drift exceeds the preset drift threshold, it is determined that the current detection does not meet the discrimination condition, and the discrimination result is suppressed, or the first reference signal, crude oil signal and second reference signal are reacquired and the discrimination amount is recalculated.

6. The method for online detection of oil quality in a lithium battery monorail crane according to claim 2, characterized in that: The discrimination quantity includes a differential discrimination quantity and a ratio discrimination quantity; the differential discrimination quantity and the ratio discrimination quantity are compared with the corresponding thresholds respectively. When both meet the preset alarm conditions, an oil quality abnormality alarm is output; otherwise, a normal oil quality result is output.

7. The method for online detection of oil quality in a lithium battery monorail crane according to claim 4, characterized in that: When the difference between the two consecutive reference oil detection signals exceeds the drift threshold and an abnormal alarm is suppressed, the current discrimination value is frozen and the previous valid detection result is maintained; the abnormal alarm confirmation output is restored only when the difference between the reference oil detection signals does not exceed the drift threshold within a preset number of consecutive times.

8. The method for online detection of oil quality in a lithium battery monorail crane according to claim 4, characterized in that: The difference between the two reference oil detection signals is determined by a combination of differential difference and ratio difference; when any difference exceeds the corresponding drift threshold, the abnormal alarm is suppressed; when both the differential difference and the ratio difference do not exceed the corresponding drift threshold and the discrimination value is still greater than the preset threshold, an abnormal oil quality alarm is confirmed and output.

9. The method for online detection of oil quality in a lithium battery monorail crane according to claim 5, characterized in that: When the reference drift exceeds the preset drift threshold, a reference signal retest is performed first to distinguish between instantaneous disturbances and continuous drift. If the reference drift obtained from the retest still exceeds the preset drift threshold, the first reference signal, crude oil signal and second reference signal are reacquired; if the reference drift obtained from the retest does not exceed the preset drift threshold, the discrimination quantity is recalculated and the discrimination result is output.

10. The method for online detection of oil quality in a lithium battery monorail crane according to claim 9, characterized in that: The reference signal retest includes: continuously acquiring at least two reference signals without changing the reference oil generation conditions, and calculating the difference between two adjacent reference signals; when any difference exceeds the preset drift threshold, it is determined to be a continuous drift and triggers the reacquisition of the first reference signal, the crude oil signal and the second reference signal; when each difference does not exceed the preset drift threshold, it is determined to be an instantaneous disturbance and the discriminant is recalculated.

Images