A method for on-line dynamic filtration of lubricating oil of a fan

Abstract

The application provides a fan lubricating oil online dynamic filtering method, relates to the technical field of data processing, and determines a filtering trigger threshold value by performing performance denaturation analysis of target lubricating oil, trains an online filtering unit in combination with the filtering trigger threshold value, transmits real-time state information to the online filtering unit, and determines an online filtering scheme; performs lubricating oil online filtering processing based on a target filter, and performs feedback decision management and feedback control management of target lubricating oil filtering in combination with filtering records. The technical problem that the filtering control precision of the fan lubricating oil online filtering in the prior art is low and the effectiveness of fan operation stability and fan life maintenance is weak is solved. The technical effects that the filtering effectiveness of the fan lubricating oil online dynamic filtering is improved, the operation stability of the wind turbine generator is ensured, and the risk that the lubricating oil impurity is not filtered in time to induce the shortening of the service life of the unit is avoided are achieved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to an online dynamic filtration method for fan lubricating oil. Background Technology

[0002] Currently, there are some shortcomings in the lubricating oil filtration system of wind turbines, especially in terms of the control accuracy of online filtration. These shortcomings may reduce the stability of wind turbine operation and affect the service life of the wind turbine.

[0003] Specifically, current online lubricating oil filtration technology is not very effective in removing solid particles and liquid impurities from lubricating oil. These impurities include metal shavings generated by wear, hard particles carried in by air, and moisture. The combined effect of these impurities and physical and chemical factors leads to the degradation of lubricating oil performance. If lubricating oil is not filtered in time, it will aggravate mechanical wear, forming a vicious cycle, which may eventually lead to the failure of key components of the wind turbine.

[0004] In summary, the current technology for online filtration of fan lubricating oil has low filtration control accuracy and is not very effective in ensuring stable fan operation and maintaining fan life. Summary of the Invention

[0005] This application provides an online dynamic filtration method for fan lubricating oil, which addresses the technical problems of low filtration control accuracy and weak effectiveness in maintaining stable fan operation and extending fan life in existing online filtration technologies.

[0006] In view of the above problems, this application provides an online dynamic filtration method for wind turbine lubricating oil. The method includes: performing performance degradation analysis on the target lubricating oil to determine a filtration trigger threshold, wherein the filtration trigger threshold is determined based on the over-limit critical value of multiple filtration targets; training an online filtration unit based on the filtration trigger threshold, wherein the filtration unit includes a data detection layer and a filtration management layer connected before and after; reading the real-time status information of the target lubricating oil and transmitting it to the online filtration unit to perform filtration trigger decision and filter management decision to determine an online filtration scheme; performing online filtration processing of the lubricating oil based on the online filtration scheme, wherein the filtration structure of the target filter is adapted to the filtration requirements; and performing feedback decision management and feedback control management of the target lubricating oil filtration based on filtration records.

[0007] The technical solution provided in this application has at least the following technical effects or advantages:

[0008] The method provided in this application determines a filtration trigger threshold by performing performance degradation analysis on the target lubricating oil. This threshold is based on the critical values ​​exceeding the limits of multiple filtration targets. Using this threshold, an online filtration unit is trained. This unit includes a pre- and post-connected data detection layer and a filtration management layer. Real-time status information of the target lubricating oil is read and transmitted to the online filtration unit for filtration trigger and filter management decisions, determining an online filtration scheme. Based on this scheme, online filtration of the lubricating oil using a target filter is performed, where the filter structure of the target filter is adapted to the filtration requirements. Finally, feedback decision management and feedback control management of the target lubricating oil filtration are performed based on filtration records. This method achieves the technical effect of improving the effectiveness of online dynamic filtration of wind turbine lubricating oil, ensuring the operational stability of wind turbine units, and avoiding the risk of shortened unit lifespan caused by untimely filtration of lubricating oil impurities. Attached Figure Description

[0009] Figure 1 This application provides a schematic diagram of an online dynamic filtration method for fan lubricating oil.

[0010] Figure 2 This is a flowchart illustrating the process of determining the filtration trigger threshold in an online dynamic filtration method for fan lubricating oil provided in this application. Detailed Implementation

[0011] This application provides an online dynamic filtration method for wind turbine lubricating oil, addressing the technical problems of low filtration control accuracy and weak effectiveness in maintaining wind turbine operational stability and lifespan in existing online wind turbine lubricating oil filtration systems. It achieves the technical effect of improving the filtration effectiveness of online dynamic filtration of wind turbine lubricating oil, ensuring the operational stability of wind turbine units, and avoiding the risk of shortened unit lifespan caused by untimely filtration of lubricating oil impurities.

[0012] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0013] Example 1

[0014] like Figure 1 As shown, this application provides an online dynamic filtration method for fan lubricating oil, including:

[0015] A100: Perform performance degradation analysis on the target lubricating oil to determine the filtration trigger threshold, wherein the filtration trigger threshold is determined based on the over-limit critical value of the multi-element filtration target;

[0016] In one embodiment, such as Figure 2 As shown, the method step A100 of this application, which involves performing performance degradation analysis on the target lubricating oil to determine the filtration trigger threshold, further includes:

[0017] A110: Determine the multi-element filtration targets, including solid substances and liquid substances;

[0018] A120: Taking the lubrication effect and component life of the wind turbine gear as the target, an impact analysis is performed on the multi-element filtration target to determine the filtration trigger threshold, wherein the impact includes physical and chemical effects;

[0019] A130: Wherein, the filtering trigger threshold has a weight identifier.

[0020] In one embodiment, the filtering trigger threshold has a weight identifier, and the method step A130 provided in this application further includes:

[0021] A131: Determine the hierarchical distribution weights based on the target's intrinsic influence and target content;

[0022] A132: Based on the hierarchical distribution weights, the filtering trigger threshold is identified by hierarchical weights, wherein the first weight layer is the distribution weight value and the second weight layer is the weight distribution rule.

[0023] Specifically, it should be understood that during the operation of wind turbine equipment, the circulating lubricating oil comes into contact with air. When the moisture in the air condenses as the temperature decreases, it may enter the lubricating oil, resulting in liquid impurities (moisture) in the lubricating oil. Moisture in the lubricating oil is harmful to the normal operation of wind turbine equipment, as it will accelerate the oxidation and deterioration of the oil and may also lead to corrosion and wear of metal parts.

[0024] Meanwhile, the friction and operation between mechanical parts during the operation of wind turbine units can lead to wear of metals or non-metals, generating tiny solid particles. These contaminants entering the lubrication system can accelerate the aging of the lubricating oil, reduce its lubrication performance, cause additional wear on mechanical parts, and lead to equipment failure.

[0025] Therefore, this embodiment requires monitoring and filtering of solid and liquid substances in the lubrication oil system of the wind turbine to maintain the stable operation of the wind turbine.

[0026] Specifically, in this embodiment, the multi-element filtration target is determined, which includes solid and liquid substances corresponding to the foregoing.

[0027] The physical effects refer to the fact that solid substances can accelerate the wear of mechanical parts, reduce the lubricating performance of lubricating oil, and lead to equipment failure; the chemical effects refer to the fact that moisture (liquid substances) can promote the oxidation and deterioration of lubricating oil, produce acidic substances, and accelerate the corrosion of metal parts.

[0028] Based on this, this embodiment performs network data retrieval to obtain multiple sample aging and filtration information of multiple sample lubricating oils. Each sample aging and filtration information includes the sample chemical measured value, the sample physical measured value, the sample lubrication performance deviation value, and the sample component life deviation value. The lubrication performance deviation value is the amount of lubrication performance degradation of the sample lubricating oil before filtration, and the component life deviation value is the amount of component life degradation of the wind turbine unit before filtration.

[0029] A model was established based on regression analysis to show the relationship between the content of solid and liquid substances and lubrication performance and component life. The parameters of the obtained model were optimized using the aging filtering information of multiple samples. This allows the prediction of lubricant performance deviations and component life deviations by inputting liquid and solid substance data into the parameter model.

[0030] The parameter model inputs are enumerated and selected until the critical values ​​for solid matter and liquid matter that predict lubricant performance deviation and component life deviation approach zero are obtained. The data combination of the obtained critical values ​​for solid matter and liquid matter is used as the filtration trigger threshold. It should be understood that when the content of either solid matter or liquid matter in the lubricant reaches the filtration trigger value, lubricant filtration in the lubricant system is required.

[0031] The change in lubrication performance deviation per unit change in liquid content minus the change in component life deviation is calculated based on the aging and filtration information of multiple samples; the change in lubrication performance deviation per unit change in solid content minus the change in component life deviation is also calculated based on the aging and filtration information of multiple samples.

[0032] Based on the two sets of changes in lubrication performance deviation and changes in component life deviation, weights are assigned to solid and liquid substances to obtain hierarchical distribution weights. These hierarchical distribution weights reflect the degree of influence of liquid substance changes on lubricating oil-components and the degree of influence of solid substance changes on lubricating oil-components.

[0033] Based on the hierarchical distribution weights, the filtering trigger thresholds are identified by hierarchical weights, where the first weight layer is the distribution weight value and the second weight layer is the weight distribution rule.

[0034] The distribution weight value is the mapping and assignment of two weight values ​​of the hierarchical distribution weight to the critical values ​​of solid matter and liquid matter. The distribution weight value reflects the severity of the impact of the two impurities on the performance of the wind turbine lubrication system.

[0035] The weight distribution rule is a logical rule. Specifically, the weight distribution weight is used to map the two weight values ​​of the hierarchical distribution weight to the critical values ​​of solid and liquid substances, and then multiply them to obtain the updated critical values ​​of solid and liquid substances.

[0036] Specifically, when the measured value of a solid substance reaches the update threshold for solid substances, a pre-warning is triggered; when the threshold for solid substances is reached or exceeded, the filtration program is automatically started. When the measured value of a liquid substance reaches the update threshold for liquid substances, a pre-warning is triggered; when the threshold for liquid substances is reached or exceeded, the filtration program is automatically started.

[0037] This embodiment uses data analysis to determine the filter trigger threshold, thus providing reference information for subsequent online lubricant filtration trigger monitoring.

[0038] A200: Based on the aforementioned filter trigger threshold, train an online filtering unit, the filtering unit including a data detection layer and a filtering management layer connected before and after;

[0039] Specifically, in this embodiment, the updated solid threshold value and the updated liquid threshold value are synchronized to the data detection layer. A large amount of sample solid and liquid material data is called to optimize the comparison performance of the data detection layer, so that the data detection layer can quickly determine the numerical relationship between solid material data, liquid material data, and the updated solid, updated liquid, solid, and liquid material threshold values.

[0040] The interactive target filter obtains multiple sets of sample solid content, sample liquid content, sample filtering control parameters, and sample filtering execution time, and stores these multiple sets of sample solid content, sample liquid content, sample filtering control parameters, and sample filtering execution time based on a knowledge graph, and then synchronizes the storage results to the filtering management layer.

[0041] A300: Reads the real-time status information of the target lubricating oil, transmits it to the online filtration unit, makes filtration triggering decisions and filter management decisions, and determines the online filtration scheme;

[0042] In one embodiment, the transmission to the online filtering unit for filtering triggering decision-making, method step A300 provided in this application includes:

[0043] A310: In conjunction with the data detection layer, threshold exceedance determination is performed on the real-time status information to determine the filter trigger instruction. The exceedance includes a first exceedance based on the filter trigger threshold and a second exceedance based on a preset threshold expansion range. The preset threshold expansion range is determined based on the filtering performance of the filter structure.

[0044] A320: If the filter trigger instruction is empty, end the filter decision.

[0045] In one embodiment, the method steps provided in this application further include:

[0046] A330: If the filter trigger instruction is not empty and is the first over-limit, filter control decision is made in conjunction with the filter management layer to determine the online filter scheme, wherein the filter control decision includes time response dimension and parameter control dimension.

[0047] In one embodiment, the method steps provided in this application further include:

[0048] A340: If the filter trigger instruction is not empty and is the second over-limit, perform multi-layer superimposed filter triggering.

[0049] Specifically, in this embodiment, the real-time status information is the real-time liquid content and real-time solid content of the target lubricating oil obtained based on existing sensor monitoring. After the interactive sensor reads and obtains the real-time status information of the target lubricating oil, the real-time status information is transmitted to the online filtration unit.

[0050] The preset threshold exceedance judgment rule includes a first exceedance based on the filter trigger threshold and a second exceedance based on the preset threshold expansion range.

[0051] The filtration performance is obtained interactively, representing the maximum amount of impurities that the filtration structure can actually handle. Specifically, the filtration performance includes liquid filtration limits and solid filtration limits. The second exceedance occurs when the real-time solid content and / or real-time liquid content in the real-time status information exceeds the liquid filtration limits and / or solid filtration limits of the filtration performance. It should be understood that the liquid filtration limits and solid filtration limits of the filtration performance map to solid and liquid threshold values ​​that are greater than the filtration trigger threshold.

[0052] The first limit is defined as the real-time solid content and / or real-time liquid content in the real-time status information exceeding the critical value of the solid substance and / or the critical value of the liquid substance in the filter trigger threshold, but not exceeding the liquid filtration limit and / or the solid filtration limit.

[0053] In conjunction with the data detection layer, threshold exceedance determination is performed on the real-time status information to determine the filtering trigger instruction.

[0054] Firstly, if the real-time solid content and real-time liquid content in the real-time status information are both less than the updated solid threshold and the updated liquid threshold, or if the real-time solid content and / or real-time liquid content in the real-time status information fall within the data range of the updated solid threshold and the solid substance threshold, and the data range of the liquid substance threshold and the updated liquid threshold, then the filter trigger instruction is empty, the filter decision ends, and there is no need to filter the target lubricating oil at this time.

[0055] Secondly, if the filter trigger command is not empty and is the first limit exceeded, then in the filter management layer, the real-time solid content and real-time liquid content are used to traverse multiple groups of sample solid content-sample liquid content-sample filter control parameters-sample filter execution time to perform filter control decision matching in the time response dimension and parameter control dimension to obtain the online filter scheme. The online filter scheme includes the filter control parameter setting of the filter device and the specific execution time of the set filter control parameters.

[0056] Third, if the filter trigger command is not empty and is the second limit exceeded, multi-layer superimposed filter triggering is performed. The multi-layer superimposed filter is to use real-time solid content and real-time liquid content to traverse multiple groups of sample solid content-sample liquid content-sample filter control parameters-sample filter execution time to match the filter control decision between the time response dimension and the parameter control dimension. After obtaining the online filter scheme, the specific execution time of the filter control parameters in the online filter scheme is doubled to achieve effective filtration of the target lubricating oil based on the target filter.

[0057] This embodiment achieves the technical effect of effectively filtering lubricating oil based on scientific calculations, reducing the dependence of lubricating oil filtration control on human experience.

[0058] A400: Based on the online filtration scheme, perform online filtration of lubricating oil based on the target filter, wherein the filtration structure of the target filter is adapted to the filtration requirements;

[0059] In one embodiment, the filter structure of the target filter is adapted to the filtration requirements, and the method step A400 provided in this application further includes:

[0060] A410: Based on the multi-dimensional filtration objectives, a filter medium is determined, wherein the filter medium includes at least one;

[0061] A420: With filtration energy efficiency as the target, perform ratio optimization and distribution optimization based on the filter media to determine the filter layer scheme, wherein the filter layer scheme is used to configure the filter structure.

[0062] Specifically, in this embodiment, the target filter is a device that specifically performs lubricating oil filtration. The online filtration scheme is executed based on the target filter to complete the online filtration treatment of the target lubricating oil.

[0063] It should be understood that the target filter is capable of performing the online filtration scheme and completing the online filtration treatment of the target lubricating oil on the premise that the filtration structure of the target filter is adapted to the filtration requirements.

[0064] The method to ensure that the filtration structure of the target filter is compatible with the filtration requirements is as follows:

[0065] The multi-element filtration targets include solid substances and liquid substances. Based on the multi-element filtration targets, the filter media are determined, including solid filter media and liquid filter media.

[0066] Solid filter media are used to capture and retain solid particles in lubricating oil and may include metal mesh, microporous membranes, activated carbon, etc. Liquid filter media are used to remove liquid contaminants from the oil, such as demulsifiers, desiccant, or other materials that can react with and separate liquid impurities. Filter media are arranged in different layers of the filter structure, each layer targeting a specific type and size of impurity.

[0067] With filtration efficiency as the target, the ratio optimization based on the filter media is performed to determine the combination ratio of different filter media in order to maximize filtration efficiency and overall system performance, and the ratio of solid filter media and liquid filter media are obtained as the ratio optimization results.

[0068] The distribution optimization of the solid filter media ratio is used to determine the optimal position and order of the solid filter media in the solid filter structure to ensure that the lubricating oil can remove solid impurities when passing through the solid filter media.

[0069] Performing liquid filter media ratio optimization is used to determine the optimal position and order of the liquid filter media in the liquid filter structure to ensure that the lubricating oil can remove liquid-type impurities when passing through the liquid filter media.

[0070] It should be understood that this embodiment uses an enumeration method to determine the optimal position and order of the liquid filter media in the liquid filter structure, and to determine the optimal position and order of the solid filter media in the solid filter structure.

[0071] The optimal position and order of the liquid filter media in the liquid filtration structure, obtained through optimization, are used to determine the optimal position and order of the solid filter media in the solid filtration structure, which is then used as the filtration layer scheme. The filtration structure is then configured based on this scheme. After synchronizing the configured filtration structure to the target filter, the target filter is started to execute the online filtration scheme.

[0072] This embodiment achieves the technical effect of ensuring the effectiveness of the target filter in performing online lubricating oil filtration at the hardware configuration level.

[0073] A500: Based on the filtration records, perform feedback decision management and feedback control management for the target lubricating oil filtration.

[0074] In one embodiment, the method step A500 of this application, which combines filter records to perform feedback decision management for the target lubricating oil filtration, further includes:

[0075] A510: Obtain the filtered records of the predetermined time zone and mine the filtered deviation data, wherein the filtered deviation data meets the preset frequency and includes the filtered deviation feature - deviation feature value;

[0076] A520: Tracing the causes of the filtered deviation data to pinpoint the deviation causes;

[0077] A530: Based on the aforementioned deviation triggers, the online filtering unit performs feedback learning.

[0078] Specifically, it should be understood that the target filter is equipped with a sensor at its output end to measure the state information of the filtered lubricating oil, which includes the content of filtered liquid and the content of filtered solids.

[0079] Obtain filtering records for a predetermined time zone (e.g., the middle 1 / 3 of the time period during which the target filter performs filtering), and mine filtering deviation data. The filtering deviation data meets a preset frequency. Each set of filtering deviation data specifically includes filtering deviation features and deviation feature values. The filtering deviation features include two types: solid and liquid. The deviation feature value is the amount by which the liquid / solid content after filtration exceeds a preset participation threshold.

[0080] For example, if the filtered deviation data obtained from mining a single filtered record does not meet the preset frequency, then data mining of the previous filtered record is performed until the preset frequency is met.

[0081] The cause of the filter deviation data is traced, which includes performing structural effectiveness testing of the filter structure and sensor fault detection to measure real-time solid content and real-time liquid content, in order to locate the cause of the deviation.

[0082] If the deviation is caused by a defect in the effectiveness of the filtering structure, the data of the filtering management layer is updated. If the deviation is caused by a sensor malfunction, the data of the filtering trigger threshold in the data detection layer is updated, thus completing the feedback learning of the online filtering unit.

[0083] This embodiment achieves the technical effect of improving the filtration effectiveness of online dynamic filtration of wind turbine lubricating oil, ensuring the operational stability of wind turbine units, and avoiding the shortened lifespan of units caused by untimely filtration of lubricating oil impurities.

[0084] In summary, any of the methods described above can be stored as computer instructions or programs in various types of computer memory, and the computer instructions or programs can be recognized by various types of computer processors to implement any of the above methods.

[0085] Based on the above specific embodiments of the present invention, any improvements and modifications made to the present invention by those skilled in the art without departing from the principle of the present invention shall fall within the patent protection scope of the present invention.

Claims

1. A method of on-line dynamic filtration of lubricating oil of a fan, characterized in that, The method comprises: performing performance degeneration analysis on target lubricating oil to determine a filter triggering threshold, wherein the filter triggering threshold is determined based on a critical value of a multi-element filter target; training an online filter unit in combination with the filter triggering threshold, the filter unit comprising a data detection layer and a filter management layer connected in front and back; reading real-time state information of the target lubricating oil and transmitting the information to the online filter unit to make filter triggering decisions and filter management decisions and determine an online filter scheme; performing target filter-based online filter processing of lubricating oil based on the online filter scheme, wherein the filter structure of the target filter is adapted to filter requirements; performing feedback decision management and feedback control management of the target lubricating oil filter in combination with filter records; the filter structure of the target filter is adapted to filter requirements, comprising: determining a filter medium based on the multi-element filter target, the filter medium comprising at least one element; performing filter efficiency-based optimization of the filter medium ratio and distribution to determine a filter layer scheme for configuring the filter structure; the feedback decision management of the target lubricating oil filter in combination with filter records comprises: obtaining filter records in a predetermined time zone and mining filter deviation data, wherein the filter deviation data meets a preset frequency and comprises filter deviation features and deviation feature values; performing cause tracing on the filter deviation data to locate deviation causes; performing feedback learning on the online filter unit in combination with the deviation causes; the performance degeneration analysis on target lubricating oil to determine a filter triggering threshold comprises: determining the multi-element filter target, including solid substances and liquid substances; performing influence analysis on the multi-element filter target to determine the filter triggering threshold, with the influence including physical influence and chemical influence, for the purpose of fan gear operation lubrication effect and component service life; wherein the filter triggering threshold has a weight identifier; the filter triggering threshold has a weight identifier, comprising: determining hierarchical distribution weights based on target self-influence degree and target content; performing hierarchical weight identification on the filter triggering threshold based on the hierarchical distribution weights, wherein a first weight layer is a distribution weight value and a second weight layer is a weight distribution rule.

2. The method of claim 1, wherein, the transmission to the online filter unit to make filter triggering decisions comprises: performing threshold overrun judgment on the real-time state information based on the data detection layer to determine filter triggering instructions, wherein the overrun includes first overrun based on the filter triggering threshold and second overrun based on a preset threshold extension interval, the preset threshold extension interval being determined based on filter performance of the filter structure; if the filter triggering instructions are empty, ending filter decisions.

3. The method of claim 2, wherein, if the filter triggering instructions are not empty and are the first overrun, performing filter control decisions in combination with the filter management layer to determine the online filter scheme, wherein filter control decisions include time response dimensions and parameter control dimensions.

4. The method of claim 3, wherein, if the filter triggering instructions are not empty and are the second overrun, performing multi-layer superimposed filter triggering.

Images