Total acid value detection method, steam turbine, device, medium and product

Abstract

The embodiment of the invention provides a total acid value detection method, a steam turbine, a device, a medium and a product. The detection method comprises the following steps: acquiring real-time parameter information of EH fire-resistant oil on a branch connected with a steam turbine equipment oil tank through at least one detection sensor; obtaining a real-time predicted total acid value according to the real-time parameter information through a pre-trained total acid value detection model; the corresponding temperature compensation sub-model is matched through the temperature value, the moisture characterization parameter and the pollution characterization parameter in the total acid value detection model are corrected through the matched temperature compensation sub-model, and the real-time predicted total acid value is obtained based on the corrected moisture characterization parameter and the pollution characterization parameter. A detection sensor is arranged on a branch connected with a steam turbine equipment oil tank, real-time parameter information (temperature, moisture characterization parameters and pollution characterization parameters) of EH fire-resistant oil in operation is directly obtained, and continuous and real-time monitoring of the total acid value is achieved.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of steam turbine equipment, in particular to a total acid value detection method, a steam turbine, an apparatus, a medium and a product. BACKGROUND

[0002] In large steam turbine equipment, the EH (electro-hydraulic regulation) system widely uses phosphate ester type fire-resistant oil as working medium. The quality of EH fire-resistant oil, especially its total acid value (TAN), is a key indicator for evaluating the aging degree of oil and judging whether it meets the operation requirements. Too high total acid value will accelerate the deterioration of the oil itself, cause sludging, and seriously corrode the metal parts (such as servo valves, oil cylinders, pipelines) in the system, leading to major equipment failures such as seal failure, control precision decline, and even servo valve jamming, which threatens the safe and stable operation of the unit. Therefore, it is crucial to detect the total acid value of EH fire-resistant oil in a timely and accurate manner.

[0003] Currently, the conventional detection method of total acid value of EH fire-resistant oil in steam turbine mainly relies on laboratory offline analysis. The operator needs to manually sample from the oil tank or circulating pipeline regularly, send the oil sample to the laboratory, and perform titration analysis according to the standard method (such as ASTM D664 or GB / T7304). This method has significant limitations: it usually takes several hours or even several days from sampling, sample sending to obtaining results, which cannot reflect the real-time acid value state of the oil. During this period, the acid value of the oil may have changed significantly, making it difficult to discover abnormal situations such as sudden increase in acid value in a timely manner, delaying the processing opportunity. It cannot provide continuous acid value change trend data, which is not conducive to state early warning and predictive maintenance. SUMMARY

[0004] In view of the above-mentioned shortcomings of the prior art, the purpose of the present disclosure is to provide a total acid value detection method, a steam turbine, an apparatus, a medium and a product, which solve the problems in the related art.

[0005] The first aspect of the present disclosure provides a total acid value detection method, wherein the detection method is arranged in an oil tank of a steam turbine equipment for detecting EH fire-resistant oil; the detection method comprises:

[0006] acquiring real-time parameter information of the EH fire-resistant oil in the oil tank of the steam turbine equipment by at least one detection sensor; wherein the real-time parameter information includes temperature value, moisture representation parameter and pollution representation parameter;

[0007] obtaining real-time predicted total acid value according to the real-time parameter information through a pre-trained total acid value detection model; wherein, the pre-trained total acid value detection model comprises a plurality of temperature compensation sub-models corresponding to different temperature gradient ranges respectively; the corresponding temperature compensation sub-model is matched through the temperature value, and the moisture representation parameter and the pollution representation parameter in the total acid value detection model are corrected through the matched temperature compensation sub-model, and the real-time predicted total acid value is obtained based on the corrected moisture representation parameter and the pollution representation parameter.

[0008] In the embodiments of the first aspect, the real-time predicted total acid value and the historical data are input into a pre-trained time sequence prediction model to predict future predicted total acid values at future preset time points or time periods.

[0009] An alarm information is generated based on the future predicted total acid values.

[0010] In the embodiments of the first aspect, the prediction of the future predicted total acid values at the future preset time points or time periods further comprises updating the historical data with the real-time predicted total acid values obtained at each time point to update the future predicted total acid values at the preset time points or time periods through the time sequence prediction model.

[0011] In the embodiments of the first aspect, the generation of the alarm information based on the future predicted total acid values comprises:

[0012] A total acid value warning event is determined and an intervention action is triggered based on consistency indication of the future predicted total acid value and a real-time predicted total acid value change trend in a specified time period before a future time point corresponding to the future predicted total acid value.

[0013] In the embodiments of the first aspect, the consistency indication satisfies at least one of the following conditions:

[0014] 1) the real-time predicted total acid value continuously rises in a preset time range, or the difference between the real-time predicted total acid value and the future predicted total acid value is less than a preset threshold value;

[0015] 2) the real-time predicted total acid value reaches a preset risk value at a preset time length away from the future preset time point; wherein, the preset risk value is a preset proportion value of the future predicted total acid value;

[0016] 3) when the real-time predicted total acid value does not reach the future predicted total acid value at the future preset time point, it is determined that the total acid value warning event does not occur, and the future predicted total acid value at the next future preset time point is continuously predicted.

[0017] In the embodiments of the first aspect, the temperature compensation sub-model is established by the following method:

[0018] Obtain moisture characterization parameters, pollution characterization parameters and measured total acid value of EH fire-resistant oil samples under different preset temperature gradients;

[0019] For each temperature gradient, train an independent sub-model to establish the mapping relationship between the moisture molecule parameters, pollution characterization parameters and total acid value at the temperature.

[0020] The temperature gradient is an equidistantly divided temperature interval, or a non-equidistant temperature interval based on the working temperature range of the EH fire-resistant oil.

[0021] In an embodiment of the first aspect, a detection branch is further provided, two ends of the detection branch being connected to the oil pipe of the oil tank of the steam turbine equipment; the sensor is arranged in the detection branch.

[0022] The second aspect of the present disclosure provides a computer device, which comprises:

[0023] a processor and a memory;

[0024] The memory stores program instructions.

[0025] The processor is configured to execute the program instructions to perform the detection method of any one of the first aspect.

[0026] The third aspect of the present disclosure provides a computer-readable storage medium, which stores program instructions, and the program instructions are executed to perform the detection method of any one of the first aspect.

[0027] The fourth aspect of the present disclosure provides a computer program product, which comprises program instructions for performing the detection method of any one of the first aspect.

[0028] The present disclosure has the following advantages: by arranging a detection sensor on the branch connected to the oil tank of the steam turbine equipment, real-time parameter information (temperature, moisture characterization parameters, pollution characterization parameters) of the EH fire-resistant oil in operation is directly obtained, the dependence on offline analysis in the laboratory is completely eliminated, and continuous and real-time monitoring of the total acid value is realized. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 The figure shows the overall flowchart of the detection method of the total acid value in an embodiment of the present disclosure.

[0030] Figure 2 The figure shows the overall flowchart of the detection method of the total acid value in another embodiment of the present disclosure.

[0031] Figure 3 The figure shows the construction method flowchart of the temperature compensation sub-model in an embodiment of the present disclosure.

[0032] Figure 4 A structural diagram of a computer device in an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0033] The present disclosure is described below by way of specific examples, which are not intended to limit the scope of the present disclosure. The advantages and features of the present disclosure can be easily understood by those skilled in the art from the messages disclosed in the present disclosure. The present disclosure can also be implemented or applied by other different embodiments or modules, and the details in the present disclosure can be modified or changed in various ways without departing from the spirit of the present disclosure. It should be noted that the embodiments in the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0034] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings, so that those skilled in the art can easily implement the present disclosure. The present disclosure can be embodied in various different forms, and is not limited to the embodiments described herein.

[0035] In the present disclosure, the expressions of "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like mean that the specific features, structures, materials or characteristics expressed in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. Moreover, the specific features, structures, materials or characteristics expressed can be combined in any one or a group of embodiments or examples in a suitable manner. In addition, the skilled in the art can combine and combine the different embodiments or examples expressed in the present disclosure and the features of the different embodiments or examples without conflict.

[0036] In addition, the terms "first", "second" are used only for the purpose of expression, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features limited by "first", "second" can explicitly or implicitly include at least one of the features. In the present disclosure, the meaning of "a group" is two or more, unless otherwise specifically limited.

[0037] In order to clearly illustrate the present disclosure, devices irrelevant to the description are omitted, and the same or similar constituent elements throughout the description are assigned the same reference numerals.

[0038] Throughout the description, when it is said that a device is "connected" to another device, it not only includes the case of "direct connection", but also includes the case of "indirect connection" in which other elements are placed therebetween. In addition, when it is said that a device "includes" a certain constituent element, unless otherwise specifically stated, other constituent elements are not excluded, but it means that other constituent elements can also be included.

[0039] Although the terms first, second, etc. can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first interface and a second interface, etc. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises", "comprising", "includes" and / or "including", when used herein, specify the presence of stated features, steps, operations, elements, modules, items, components, and / or groups thereof, but do not preclude the presence or addition of one or more other features, steps, operations, elements, modules, items, components, and / or groups thereof. As used herein, the terms "or" and "and / or" are construed to be inclusive, or mean any one or any combination of the listed items. Thus, "A, B or C" or "A, B and / or C" means any of the following: A; B; C; A and B; A and C; B and C; A, B and C. Exceptions to this definition are only present when items are grouped in conjunction with the phrase "one of X, Y or Z" or "one of X, Y and / or Z" unless otherwise stated.

[0040] The professional terms used herein are used only to refer to specific embodiments and are not intended to limit the present disclosure. The singular form used herein, unless the context clearly indicates otherwise, also includes the plural form. The meaning of "include" used in the specification is to specify a particular characteristic, region, integer, step, operation, element, and / or component, and does not exclude the presence or addition of other characteristics, regions, integers, steps, operations, elements, and / or components.

[0041] Although not differently defined, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. Terms defined in commonly used dictionaries are to be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and the present disclosure, and should not be interpreted in an idealized or overly formal sense unless clearly defined otherwise.

[0042] Anti-flame oil (typified by EH anti-flame oil) bears multiple core roles in large power equipment systems such as steam turbines, and its functions are directly related to the safe operation and control accuracy of the equipment: as the hydraulic working medium of the electro-hydraulic regulation system (EH system) of the steam turbine, it precisely controls the opening of key components such as the main steam valve and the speed regulating valve through the conversion and transmission of hydraulic energy, realizes the real-time regulation of the steam turbine speed and load with stable pressure output, ensures the parameter matching of the unit under all operating conditions, and is the core carrier of the "power regulation center" of the equipment; at the same time, it has a very high ignition point (> 230℃) and is not easy to burn in the high-temperature environment near the steam turbine cylinder, which can effectively avoid the fire risk caused by leakage of traditional mineral hydraulic oil, and is especially suitable for scenes such as power plants with strict fire safety requirements, and the changes in its physical and chemical properties (such as total acid number, viscosity, and water content) can directly reflect the degree of oil deterioration, and potential system failures can be warned in advance through monitoring, providing a basis for equipment maintenance.

[0043] Detecting total acid number (TAN) is a key indicator for evaluating the deterioration and corrosion of oil (such as EH anti-flame oil, lubricating oil, etc.). During use, oil is affected by high temperature, oxygen, moisture, and other factors, and undergoes oxidation, hydrolysis, and other reactions to generate corrosive substances such as organic acids and acidic esters. The total amount of these substances is directly reflected by the total acid number. An increase in the total acid number means an increase in acidic substances in the oil, which not only intensifies the corrosion of metal components (destroys the passivation film and triggers electrochemical reactions), but also can lead to changes in oil viscosity, a decrease in lubrication performance, and even accelerate the aging and failure of seals. By regularly detecting the total acid number, the deterioration trend of the oil can be grasped in a timely manner, and it can be determined whether the oil needs to be replaced or purified, thereby preventing equipment corrosion, wear, and other failures and ensuring the safe and stable operation of the system.

[0044] In related technologies, there are obvious deficiencies in oil deterioration monitoring and equipment protection: first, the detection means relies mainly on offline analysis, which requires laboratory testing after sampling, and has a lag, making it difficult to reflect the dynamic deterioration trend of the oil in real time and easily missing the best intervention opportunity; second, related technologies mainly focus on a single indicator and lack collaborative analysis of multiple parameters such as viscosity, moisture, and particle size, making it difficult to comprehensively evaluate the overall performance of the oil; in addition, the early warning mechanism for early deterioration is not perfect, and intervention is often triggered only when the indicator is significantly out of range, at which time the equipment may have already experienced hidden corrosion or wear, increasing maintenance costs and failure risks.

[0045] To solve the above problems, an embodiment of the present disclosure provides a method for detecting total acid number, wherein the method is arranged in an oil tank of a steam turbine equipment and is used for detecting EH anti-flame oil; in the method, Figure 1In the embodiment, by setting detection sensors on the oil tank branch of the steam turbine equipment, the temperature value, moisture representation parameter and pollution representation parameter of the EH fire-resistant oil are obtained in real time, realizing the transition from offline detection to online real-time monitoring, eliminating the hysteresis of the traditional method, and dynamically reflecting the oil deterioration state; the pre-trained total acid value detection model is used to predict the total acid value based on multi-parameter (temperature, moisture, pollution) collaborative analysis, breaking through the limitation of single index evaluation and improving the judgment accuracy of the comprehensive performance of the oil.

[0046] Specifically, the detection method comprises:

[0047] Step S1: acquiring real-time parameter information of EH fire-resistant oil in the oil tank of the steam turbine equipment through at least one detection sensor, ensuring that the real-time collection and analysis of the oil state are realized without affecting the normal operation of the main system. The real-time parameter information includes temperature value, moisture representation parameter and pollution representation parameter.

[0048] The moisture representation parameter includes at least one of moisture content value, oil density, oil kinematic viscosity, resistivity and water activity.

[0049] Specifically, in the steam turbine equipment, the moisture content of EH fire-resistant oil and its related characteristics are one of the key factors affecting the health status of the oil. The moisture representation parameter includes moisture content value, oil density, oil kinematic viscosity, resistivity and water activity, etc. Among them, the moisture content value directly affects the oil deterioration process, because the water contacts the metal parts in the oil and generates corrosion reaction to generate acidic substances, thereby increasing the total acid value. Although the oil density itself does not directly reflect the change of the total acid value, it can indirectly indicate the aging degree of the oil. The change of the oil composition during the aging process may lead to the change of the density, and then affect the total acid value. The change of the oil kinematic viscosity is usually accompanied by oil oxidation. With the accumulation of oxidation products, the viscosity rises, which often indicates the increase of the total acid value. The resistivity reflects the conductivity performance of the oil. Low resistivity indicates that there are more conductive impurities in the oil, such as water decomposition products or metal ions, which may indirectly lead to the increase of the total acid value. Finally, the water activity refers to the proportion of dissolved water in the oil. High water activity will exacerbate the emulsification tendency of the oil and promote the oxidation reaction, further increasing the total acid value.

[0050] The pollution representation parameter includes at least one of particle pollution degree, metal ion concentration, oil sludge content and oxidation product concentration.

[0051] The higher the particle contamination level, the more likely the oil will undergo mechanical wear and oxidation, especially metal particles as catalysts accelerate the oxidation process of the oil, resulting in an increase in total acid number. The presence of metal ion concentration (such as iron, copper, etc.) can catalyze the oxidation reaction of the oil, producing more acidic substances, so the increase in metal ion concentration will significantly accelerate the oxidation rate of the oil, thereby increasing the total acid number. The oil sludge content is composed of insoluble byproducts produced by oil oxidation, which not only affects the flowability of the oil, but also adsorbs moisture and pollutants, forming a vicious cycle, and an increase in oil sludge content is usually accompanied by an increase in total acid number. The concentration of oxidation products directly reflects the degree of oxidation of the oil, and oxidation products are acidic substances and other byproducts produced during the aging process of the oil, so there is a direct positive correlation between the total acid number.

[0052] The total acid number is a comprehensive reflection of the degree of oxidation and deterioration of the oil, and the above moisture and pollution characterization parameters are the key factors affecting the oxidation process of the oil. Moisture promotes oxidation and corrosion reactions, and pollutants provide a catalytic environment or directly participate in chemical reactions, both of which work together to accelerate the deterioration of the oil, leading to the accumulation of acidic substances. Therefore, the changes in these parameters not only have a statistical correlation with the total acid number, but also form a causal relationship in the chemical reaction mechanism. By monitoring and analyzing these parameters in real time, the trend of the total acid number can be accurately predicted without chemical titration, and the online evaluation and early warning of the EH anti-flame oil state can be realized.

[0053] In some embodiments, a detection branch is also provided, and the two ends of the detection branch are connected to the oil pipe of the oil tank of the steam turbine equipment, so that the anti-flame oil can flow through the branch without affecting the operation of the main oil circuit system.

[0054] The detection branch guides part of the oil to flow through the sensor area without affecting the normal operation of the main oil circuit system, realizing non-invasive online monitoring. In the detection branch, the sensor can be installed in two ways according to its structure: one is that the sensor has a pipeline design, which can be directly integrated into the detection branch, so that the oil flows through the inside of the sensor for measurement, with the advantages of easy installation, fast response, good sealing, etc.; the other is to install the sensor in an independent oil tank, and the oil tank is connected to the main oil circuit through the detection branch, realizing local circulation sampling of the oil, which is suitable for scenes where the sensor needs to avoid direct high pressure or needs to perform complex parameter analysis. The above two installation methods can effectively ensure the real-time and accurate monitoring of the sensor on the oil condition, provide data support for oil health management, and improve the safety and intelligent level of the operation of the steam turbine equipment.

[0055] In some embodiments, the following are the types of temperature sensors, moisture characterization parameters (such as moisture content sensors for moisture content values, oil density sensors for oil density, kinematic viscosity sensors for oil kinematic viscosity, resistivity sensors for resistivity, water activity sensors for water activity), and pollution characterization parameters (such as particle pollution level sensors for particle pollution level, metal ion concentration sensors for metal ion concentration, sludge content sensors for sludge content, oxidation product concentration sensors for oxidation product concentration) for temperature values and their working principles, or in some embodiments, composite sensors can also be used to collect the above parameter values.

[0056] In some embodiments, the temperature sensor can use a thermocouple or an RTD (Resistance Temperature Detector)

[0057] Its working principle:

[0058] Thermocouples are based on the Seebeck effect, which is the generation of a voltage difference at the junction of two different metals at different temperatures. By measuring this voltage difference, the temperature can be calculated.

[0059] RTD uses the property of certain materials (usually platinum) that their electrical resistance changes with temperature. By precisely measuring this change in resistance, an accurate temperature reading can be obtained.

[0060] In some embodiments, the moisture content sensor can use a capacitive humidity sensor.

[0061] Its working principle: Capacitive humidity sensors determine moisture content by measuring the change in dielectric constant caused by water in the oil. When water is present in the oil, its dielectric constant changes, affecting the capacitance value of the internal capacitor in the sensor. By precisely measuring and calibrating the capacitance value, the moisture content in the oil can be obtained.

[0062] In some embodiments, the oil density sensor can use an oscillating tube densitometer.

[0063] Working principle: Oscillating tube densitometers are based on the relationship between resonant frequency and liquid density. When oil flows through an oscillating tube, the vibration frequency of the tube changes according to the density of the oil. By measuring this frequency change, the density of the oil can be accurately calculated.

[0064] In some embodiments, the kinematic viscosity sensor can use an ultrasonic viscosity sensor

[0065] Working principle: Ultrasonic viscosity sensors use the relationship between the propagation speed of ultrasonic waves in a liquid and the viscosity of the liquid. By emitting and receiving ultrasonic signals through the oil, and analyzing the time delay and attenuation of the signals, the viscosity of the oil can be calculated.

[0066] In some embodiments, the resistivity sensor can employ a four-probe resistivity sensor.

[0067] Working principle: The four-probe resistivity sensor calculates the resistivity by applying a small current to the oil and measuring the resulting voltage drop. This method effectively eliminates the effects of contact resistance, providing highly accurate resistivity measurements.

[0068] In some embodiments, the water activity sensor can employ a relative humidity sensor.

[0069] Working principle: Relative humidity sensors typically use capacitive or resistive elements to measure the proportion of dissolved water in the oil. These sensors indirectly reflect the water activity level in the oil by detecting changes in the relative humidity of the environment.

[0070] In some embodiments, the particulate contamination level sensor can employ a laser particle counter,

[0071] Working principle: Laser particle counters identify and count particulate matter in the oil by emitting a beam of laser light through the oil sample and detecting the intensity and distribution of scattered light. Different sizes of particles cause different scattering patterns, allowing the number and size distribution of particles to be statistically determined.

[0072] Metal ion concentration sensors can employ Atomic Absorption Spectrometry (AAS) or X-ray Fluorescence Spectrometry (XRF).

[0073] Working principle:

[0074] AAS converts the sample into gaseous atoms and irradiates them with light of a specific wavelength. The unabsorbed light is measured by a detector, providing the concentration of metal elements.

[0075] XRF, on the other hand, uses X-rays to excite the inner-shell electrons of atoms in the sample, causing characteristic X-rays to be emitted. By analyzing the energy and intensity of these characteristic X-rays, the presence and concentration of metal ions can be determined.

[0076] In some embodiments, the sludge content sensor can employ an infrared spectrometer,

[0077] Working principle: Infrared spectrometers analyze the chemical composition of the oil by measuring its absorption of infrared light. Since sludge is mainly composed of oxidation products and other insoluble byproducts, it has specific infrared absorption peaks, so the content of sludge can be determined through spectral analysis.

[0078] In some embodiments, the oxidation product concentration sensor can employ a UV-Vis spectrophotometer,

[0079] Working principle: The UV-visible spectrophotometer evaluates the concentration of oxidation products in oil by measuring the absorption of ultraviolet and visible light. Many oxidation products have distinct absorption peaks in the ultraviolet region. By analyzing the position and intensity of these absorption peaks, the concentration of oxidation products in the oil can be quantitatively analyzed.

[0080] Step S2: Obtain real-time predicted total acid value according to the real-time parameter information through a pre-trained total acid value detection model.

[0081] Specifically, real-time prediction of total acid value has significant advantages over traditional laboratory direct testing, mainly in terms of efficiency, cost, and intelligent management. By real-time collection of oil temperature, moisture, and pollution characterization parameters through sensors, and comprehensive analysis using a pre-trained model, online continuous monitoring of oil condition during equipment operation can be achieved, timely detection of oil degradation trend can be achieved, and sudden failure caused by lagging detection can be avoided. This real-time and continuity not only improves the system response speed, but also supports state-based preventive maintenance strategy, reduces unnecessary oil change frequency and maintenance cost. At the same time, automatic data collection and analysis process reduces the demand for manual operation, reduces human error and improves the intelligent level of operation and maintenance. In addition, multi-parameter fusion analysis provides more comprehensive and accurate oil condition evaluation than single indicator detection, adapts to complex operating environment, and ensures higher detection accuracy and reliability. Overall, real-time prediction of total acid value provides strong technical support for the safe and stable operation of steam turbine equipment, effectively reducing the overall operating cost.

[0082] In some embodiments, the establishment process of the total acid value detection model includes: under the control of external environmental conditions, collecting EH anti-flame oil samples with different total acid value levels, and obtaining their measurement values under multiple parameters through sensors, including temperature, moisture content, resistivity, viscosity, oxidation product concentration, and other key indicators.

[0083] Subsequently, the collected data is modeled and analyzed using multivariate linear regression method to find the correlation between each parameter and total acid value, and a mathematical model is established through curve fitting to form a total acid value detection model that can reflect the variation law of total acid value with multiple parameters.

[0084] During model training, the actual total acid value data of the oil sample obtained by the laboratory titration method can be used to calibrate and verify the model to improve the accuracy and applicability of the model prediction. The model can be further deployed to an online monitoring system to predict the current total acid value based on real-time collected oil parameters, realizing efficient evaluation and intelligent early warning of EH anti-flame oil condition.

[0085] In some embodiments, the pre-trained total acid number detection model comprises a plurality of temperature compensation sub-models corresponding to different temperature gradient ranges respectively; step S21: matching the corresponding temperature compensation sub-model through the temperature value, and correcting the moisture representation parameter and the pollution representation parameter in the total acid number detection model through the matched temperature compensation sub-model, and obtaining the real-time predicted total acid number based on the corrected moisture representation parameter and pollution representation parameter.

[0086] Specifically, the temperature compensation sub-model is set to eliminate the influence of temperature change on the moisture representation parameter and the pollution representation parameter, thereby improving the accuracy and stability of total acid value prediction. In the actual operation of the steam turbine equipment, the oil temperature may fluctuate in a large range, and these temperature changes will significantly affect the measurement results of parameters such as moisture content and resistivity. By matching the corresponding temperature compensation sub-model according to the temperature value of the current oil, these parameters can be dynamically corrected to be closer to the true value at the standard temperature. Each temperature compensation sub-model is optimized for a specific temperature gradient range and can effectively correct the deviation caused by temperature change. The corrected parameters are input into the pre-trained total acid number detection model, and the real-time predicted total acid number under the current conditions is obtained according to the temperature-compensated parameters.

[0087] For example: In order to accurately establish the total acid number detection model and consider the influence of temperature on the physicochemical properties of EH anti -burning oil, it is necessary to investigate the changes of total acid number and other related parameters of oil under different temperature conditions. Given that EH anti -burning oil will accelerate hydrolysis at high temperature, resulting in changes in total acid number, 25℃, 40℃ and 100℃ are selected as typical temperature points for experiments. For each temperature point, select oil samples with different total acid value levels, and directly measure the total acid value and moisture representation parameters (such as moisture content, resistivity, etc.) and pollution representation parameters (such as particle pollution degree, metal ion concentration, etc.) at 25℃ and 40℃. For 100℃ test, the oil sample is heated to 100℃ and kept at this temperature for a period of time to simulate the influence of high temperature operating environment on the oil, and then the oil sample is cooled to room temperature (such as 25℃) and the above parameters are measured. Through these experiments, the change trend of the physicochemical properties of the oil under different temperature conditions can be observed, the specific influence of these changes on the total acid number can be analyzed, and data support for establishing the mapping relationship of "original parameter-temperature-real parameter" for each temperature compensation sub-model can be provided.

[0088] For different temperature gradient ranges (such as 20-30℃, 30-40℃, etc.), the temperature compensation sub-model is trained based on the above experimental data. For example, temperature rise will reduce the viscosity of EH fire-resistant oil, causing the response signal of the moisture sensor (such as a capacitive sensor) to deviate from the moisture content, and the change in the solubility of the oil to moisture at high temperature may cause the original moisture content measurement value to be higher than the actual value; at the same time, temperature change will also affect the suspension state of the pollution particles, causing fluctuations in the detection of particle pollution by the laser particle counter, such as the aggregation of particles at low temperatures, which may misjudge as an increase in the proportion of large particles.

[0089] When correcting, the temperature compensation sub-model first receives the real-time temperature value, determines the corresponding temperature interval, and then calls the correction algorithm (such as a polynomial fitting formula, a neural network mapping function, etc.) of the interval; then input the original moisture content and pollution content, and eliminate the measurement deviation caused by temperature through the algorithm. For example, in the 30-40℃ sub-model, if the original moisture content measurement value is 0.08%, the sub-model calculates the deviation caused by temperature as +0.008% according to the correlation between viscosity and moisture sensor response at this temperature, combined with experimental data at temperature points such as 25℃ and 40℃, and then corrects it to 0.072%. Finally, the corrected parameters eliminate the temperature interference and more accurately reflect the actual state of moisture and pollution in EH fire-resistant oil, providing reliable input for accurate prediction of total acid number.

[0090] Optionally, in the embodiment, the temperature compensation sub-model is established by the following method: Figure 3 In the embodiment, the temperature compensation sub-model is established by the following method:

[0091] Step S211: Obtain the moisture content and pollution content of the EH fire-resistant oil sample under different preset temperature gradients.

[0092] Specifically, a sample set covering the actual working temperature range of EH fire-resistant oil (such as 20℃-120℃) is selected, and the test interval is divided according to the set temperature gradient - if equal interval division is used, 10℃ can be set as the interval (such as 20℃, 30℃, 40℃…120℃); if non-equal interval division is used based on working characteristics, the temperature segment where the physicochemical properties of the oil change dramatically (such as 30-50℃ as the hydrolysis acceleration interval, which can shorten the interval to 5℃, and 80-120℃ as the degradation stabilization interval, which can expand the interval to 20℃). For samples under each temperature gradient, the moisture content, resistivity, particle pollution, and metal ion concentration are measured synchronously, and the true total acid value is determined by the standard laboratory method, forming a sample database of "temperature-multiple parameters-total acid value".

[0093] Step S212: For each temperature gradient, train an independent sub-model to establish the mapping relationship between the moisture content, pollution content, and total acid value at that temperature.

[0094] Specifically, an independent sub-model is trained for each temperature gradient to build a dedicated mapping relationship. For sample data under a single temperature gradient, a machine learning algorithm (such as partial least squares regression, random forest, or neural network) is used for training, so that the sub-model learns the internal correlation between the moisture representation parameter, the pollution representation parameter, and the total acid number in the temperature interval (for example, in the 30-40°C sub-model, the nonlinear relationship between moisture content and total acid number is focused on fitting, because the hydrolysis reaction is more sensitive to moisture at this temperature range). Each sub-model independently stores the parameter weights and correction coefficients of its temperature interval to ensure that only the parameters of this interval are compensated.

[0095] Optionally, in Figure 2 In an embodiment, the detection method further comprises:

[0096] Step S3: inputting the real-time predicted total acid number and historical data into a pre-trained time series prediction model to predict a future predicted total acid number at a future preset time point or time period.

[0097] The pre-trained time series prediction model can be constructed based on algorithms such as long short-term memory (LSTM) and time series autoregressive integrated moving average (ARIMA), and its training data includes historical total acid number change curves, real-time parameter information (temperature, moisture, pollution, etc.) corresponding to the time period, and equipment operation condition data, which can capture the evolution law of the total acid number over time and the dynamic correlation with each parameter. For example, the model can predict the total acid number change trend every hour within the next 24 hours or the specific time point when the total acid number reaches the warning threshold based on the real-time predicted total acid number and synchronous parameters in the past 72 hours. Through this step, the prospective judgment of the EH anti-flame oil deterioration trend can be realized, providing a decision basis for early intervention in equipment maintenance and further making up for the lag of traditional offline detection.

[0098] Further, the real-time predicted total acid number is an instant total acid number calculated by the pre-trained total acid number detection model (including the temperature compensation sub-model) based on the current EH anti-flame oil real-time parameter information (temperature, moisture, pollution, etc.), which directly reflects the current deterioration state of the oil; while the future predicted total acid number is the total acid number at a future preset time point or time period derived by inputting the real-time predicted total acid number and historical data (such as total acid number change and parameter fluctuation records in the past several hours to several days) into the time series prediction model, and the core is to capture the evolution trend of the total acid number over time.

[0099] The core value of real-time prediction of total acid number lies in "immediate monitoring and rapid response": it breaks through the lag of traditional offline detection, can reflect the current accumulation state of acidic substances in the oil in real time, and can trigger intervention measures (such as oil purification, shutdown inspection) as soon as the warning threshold is exceeded, avoiding immediate failures such as equipment corrosion and seal aging caused by rapid increase of acidic substances, and ensuring the safety of current system operation.

[0100] The core value of future prediction of total acid number lies in "trend prediction and proactive planning": through the analysis of historical data and real-time values by the time series model, the change trajectory of total acid number (such as the increase amplitude in the next 24 hours, the specific time when the critical value is reached) can be predicted in advance. This feature enables it to support fine decision-making for equipment maintenance - for example, if the total acid number is predicted to approach the warning threshold in 3 days, the oil change time can be planned in advance to avoid unplanned downtime caused by sudden failure; at the same time, it can also optimize the maintenance cycle according to the equipment operating conditions, reduce the cost waste caused by excessive maintenance or insufficient maintenance, and improve the economy and stability of system operation.

[0101] The two together form a closed loop of "real-time monitoring-trend prediction-active intervention": real-time prediction solves the problem of "whether it is safe now", future prediction solves the problem of "how to deal with the future", and together they build a more comprehensive oil deterioration prevention and control system.

[0102] Step S4: generating an alarm information based on the future predicted total acid number.

[0103] When the future predicted total acid number reaches or exceeds the preset alarm threshold, the system will automatically generate an alarm information. The setting of the alarm threshold is usually based on the recommendations of equipment manufacturers, industry standards and historical failure data, for example, the total acid number reaching 0.3 mgKOH / g can be set as the warning threshold, and reaching 0.5 mgKOH / g can be set as the emergency alarm threshold. The alarm information will explicitly include the specific value of the future predicted total acid number, the time or time period when the value is expected to be reached, and the possible impact on the equipment (such as accelerated corrosion, affected lubrication performance, etc.), and will also be accompanied by specific recommended measures such as advance oil replacement and deep purification treatment. The alarm information can be pushed in real time through the monitoring platform, notified by SMS or reminded by email, etc., to timely convey to the equipment maintenance personnel, so that they can make preparations in advance and take proactive measures to avoid potential risks, further improving the forward-looking and effectiveness of equipment maintenance and reducing the probability of failure.

[0104] Optionally, the prediction of the future predicted total acid number at the future preset time point or time period also includes: updating the historical data of the real-time predicted total acid number at each time to update the future predicted total acid number at the preset time point or time period through the time series prediction model.

[0105] Specifically, when predicting the future predicted total acid value at a future preset time point or time period, the system will update the real-time predicted total acid value generated at each time point to the historical database in real time, forming a dynamically updated time series dataset. For example, every time a new real-time predicted total acid value is generated (e.g., every 5 minutes), the system will automatically store it in association with the corresponding timestamp and synchronously collected parameter information (temperature, moisture, contamination, etc.), while excluding historical records that exceed the data validity period (e.g., retaining the last 7 days of data), ensuring that the historical data always reflects the latest oil deterioration trend. On this basis, the time series prediction model will call the updated historical data at a preset period (e.g., every 30 minutes) or trigger (e.g., real-time value fluctuation exceeding 5%) to retrain or iteratively calculate the future predicted total acid value - for example, originally based on the past 24 hours of data to predict the trend 12 hours into the future, when 3 hours of real-time data are added, the model will include this 3 hours of data in the analysis, correcting any deviations that may exist in the original prediction (e.g., if the actual acid value increases faster than expected, the updated prediction will advance the time point at which the total acid value reaches the threshold).

[0106] For example, the system obtains a real-time predicted total acid value (e.g., 0.12 mgKOH / g) at 4:00 pm through the total acid value detection model, at which time the time series prediction model will combine this value with the historical data before 4:00 pm (e.g., real-time values at 1-3:00 pm, parameter changes) to predict the future total acid value at 8:00 pm (e.g., 0.15 mgKOH / g). When a new real-time predicted total acid value (e.g., 0.13 mgKOH / g) is generated at 5:00 pm, the system will automatically include this value in the historical database, replacing or supplementing the original blank data segment after 4:00 pm, so that the historical data sequence is updated to a continuous real-time value from 1-5:00 pm. Subsequently, the time series prediction model will be recalculated based on the updated historical data (including the latest real-time value at 5:00 pm), adjusting the model parameters to adapt to the latest changes in oil deterioration trends, and finally outputting the updated future predicted total acid value at 8:00 pm (e.g., corrected to 0.16 mgKOH / g), and automatically replacing the original prediction result.

[0107] Optionally, the generation of an alarm information according to the future predicted total acid value comprises:

[0108] Based on the consistency of the future predicted total acid value and the real-time predicted total acid value change trend within a specified time period before the future time point in indicating a total acid value warning event, a total acid value warning event is determined and an intervention action is triggered.

[0109] Specifically, the future predicted total acid value is analyzed in conjunction with the real-time predicted total acid value change trend within a specified time period before the future time point (e.g., 4-7:00 pm before 8:00 pm) to determine whether the two are consistent in indicating a total acid value warning event.

[0110] Optionally, the consistency indication satisfies at least one of the following conditions:

[0111] 1) the real-time predicted total acid value continuously increases within a preset time range, or the difference between the real-time predicted total acid value and the future predicted total acid value is less than a preset threshold.

[0112] Specifically, when the real-time predicted total acid value continuously increases within a preset time range (e.g., the past 2 hours), or the difference between the real-time predicted total acid value and the future predicted total acid value is less than a preset threshold (e.g., ±0.02 mgKOH / g), it indicates that the real-time deterioration trend is consistent with the future prediction direction, and the credibility of the warning event can be strengthened. For example, if the future predicted total acid value at 8 o'clock is 0.25 mgKOH / g, and the real-time values from 5 o'clock to 7 o'clock continuously increase from 0.20 mgKOH / g to 0.23 mgKOH / g (the difference is 0.02 mgKOH / g), it meets the consistency indication, and the warning logic is confirmed to be effective.

[0113] 2) the real-time predicted total acid value reaches a preset risk value at a preset time length before the future preset time point; wherein the preset risk value is a preset proportion of the future predicted total acid value.

[0114] At a preset time length (e.g., 2 hours in advance) before the future preset time point, if the real-time predicted total acid value reaches a preset risk value (e.g., 80% of the future predicted total acid value, i.e., 0.24 mgKOH / g), the consistency check is triggered. For example, the future predicted total acid value at 8 o'clock is 0.30 mgKOH / g, and the preset risk value is 80% of it (i.e., 0.24 mgKOH / g), if the real-time value at 6 o'clock reaches 0.24 mgKOH / g, it indicates that the deterioration rate matches the prediction, and the warning event is determined to be established.

[0115] 3) when the real-time predicted total acid value does not reach the future predicted total acid value at the future preset time point, it is determined that the total acid value warning event does not occur, and the future predicted total acid value at the next future preset time point is continued to be predicted.

[0116] If the real-time predicted total acid value does not reach the corresponding future predicted total acid value at the future preset time point (e.g., 8 o'clock), (e.g., the actual real-time value at 8 o'clock is 0.22 mgKOH / g, which is lower than the predicted value 0.25 mgKOH / g), it is determined that the total acid value warning event does not occur, the system automatically terminates the current warning process, and continues to predict the total acid value at the next future time point (e.g., 12 o'clock) based on the latest real-time data, avoiding invalid alarm interference.

[0117] In some embodiments, multi-level thresholds (e.g., early warning threshold, alarm threshold, emergency threshold) can also be set according to the safety level of the equipment, and when the future predicted total acid number reaches the corresponding threshold, different levels of alarms (e.g., early warning prompt, shutdown suggestion) are triggered. At the same time, combined with the distance between the current value of the real-time predicted total acid number and the threshold, the alarm intensity is dynamically adjusted (e.g., the closer to the threshold, the higher the alarm frequency). Through the hierarchical mechanism, the risk level is refined, avoiding the "black and white" judgment caused by a single threshold, and adapting to the maintenance needs in different scenarios (e.g., key equipment can increase the threshold sensitivity).

[0118] In yet another embodiment of the present disclosure, a steam turbine is provided, wherein the detection method of any of the above embodiments is applied.

[0119] As shown in FIG. 1, a structural schematic diagram of a computer device in an embodiment of the present disclosure is shown. Figure 4

[0120] The computer device 100 can be exemplified as a processing terminal in a cloud platform, such as a server, a desktop computer, a notebook computer, a tablet computer, a smart phone, or other terminals.

[0121] The computer device 100 includes a bus 101, a processor 102, and a memory 103. The processor 102 and the memory 103 can communicate through the bus 101. The memory 103 can store program instructions. The processor 102 implements the steps in the detection method in the previous embodiments by running the program instructions in the memory 103, for example Figure 1 .

[0122] The bus 101 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, although only one thick line is shown in the figure, it does not mean that there is only one bus or only one type of bus.

[0123] In some embodiments, the processor 102 can be implemented as a Central Processing Unit (CPU), a micro control unit (MCU), a System On Chip, or a Field Programmable Gate Array (FPGA), etc. The memory 103 can include a volatile memory (Volatile Memory) for data temporary storage during program running, such as a Random Access Memory (RAM).

[0124] ​The memory 103 can also include non-volatile memory for data storage, such as a Read-Only Memory (ROM), a flash memory, a Hard Disk Drive (HDD), or a Solid-State Disk (SSD).

[0125] In some embodiments, the computer device 100 can further include a communicator 104. The communicator 104 is configured to communicate with external devices. In specific examples, the communicator 104 can include one or a set of wired and / or wireless communication circuitry. For example, the communicator 104 can include one or more of, for example, a wired network card, a USB module, a serial interface module, and the like. Wireless communication protocols followed by the wireless communication module include one or more of, for example, Near Field Communication (NFC) technology, Infared (IR) technology, Global System for Mobile communications (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code division multiple access (WCDMA), Time-Division Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), Blue Tooth (BT), Global Navigation Satellite System (GNSS), and the like.

[0126] The present disclosure also provides a computer readable storage medium storing program instructions, which when executed implement the detection method of any of the preceding embodiments.

[0127] That is, the method steps in the above embodiments are implemented as software or computer code that can be stored in a recording medium such as a CD ROM, a RAM, a floppy disk, a hard disk or a magneto-optical disk, or by computer code originally stored in a remote recording medium or a non-transitory machine-readable medium and downloaded through a network and stored in a local recording medium, so that the method represented herein can be processed by such software on a recording medium using a general-purpose computer, a special-purpose processor or programmable or special-purpose hardware such as an ASIC or an FPGA.

[0128] The embodiments of the present disclosure can also provide a computer program product, comprising program instructions for executing the detection method described in any of the above embodiments.

[0129] The above embodiments are only illustrative of the principles and effects of the present disclosure, and are not intended to limit the present disclosure. Any person skilled in the art can modify or change the above embodiments without departing from the spirit and scope of the present disclosure. Therefore, all equivalent modifications or changes made by those skilled in the art without departing from the spirit and technical ideas disclosed by the present disclosure shall be covered by the protection scope of the present disclosure.

Claims

1. A method for detecting total acid value, characterized in that, Located in the turbine equipment oil tank, it is used to detect EH fire-resistant oil; the detection method includes: Real-time parameter information of EH fire-resistant oil in the turbine equipment oil tank is acquired through at least one detection sensor; wherein, the real-time parameter information includes temperature value, moisture characterization parameter and contamination characterization parameter; A pre-trained total acid value detection model is used to obtain a real-time predicted total acid value based on the real-time parameter information. The pre-trained total acid value detection model includes multiple temperature compensation sub-models, each corresponding to a different temperature gradient range. The temperature value is matched with the corresponding temperature compensation sub-model, and the moisture characterization parameters and pollution characterization parameters in the total acid value detection model are corrected by the matched temperature compensation sub-model. The real-time predicted total acid value is obtained based on the corrected moisture characterization parameters and pollution characterization parameters.

2. The detection method according to claim 1, characterized in that, Also includes: The real-time predicted total acid value and historical data are input into a pre-trained time series prediction model to predict the future predicted total acid value at a future preset time point or time period. An alarm message is generated based on the predicted total acid value.

3. The detection method according to claim 2, characterized in that, The method of predicting the future total acid value at a preset time point or within a time period also includes: updating historical data with the real-time predicted total acid value obtained at each moment, so as to update the future predicted total acid value at the preset time point or within a time period through the time-series prediction model.

4. The detection method according to claim 2, characterized in that, The step of generating an alarm message based on the predicted total acid value includes: Based on the consistency indication of the total acid value warning event based on the predicted total acid value and the real-time predicted total acid value change trend within a specified time period before the corresponding future time point, the total acid value warning event is determined and intervention action is triggered.

5. The detection method according to claim 4, characterized in that, The consistency indication satisfies at least one of the following conditions: 1) The real-time predicted total acid value continues to rise within a preset time range, or the difference between the real-time predicted total acid value and the future predicted total acid value is less than a preset threshold. 2) When the time is a preset time away from a future preset time point, predict in real time that the total acid value will reach a preset risk value; wherein, the preset risk value is a preset proportion of the future predicted total acid value; 3) If the real-time predicted total acid value does not reach the future predicted total acid value at the future preset time point, it is determined that the total acid value warning event has not occurred, and the future predicted total acid value at the next future preset time point continues to be predicted.

6. The detection method according to claim 1, characterized in that, The temperature compensation sub-model is established in the following way: Obtain the moisture characterization parameters, contamination characterization parameters, and measured total acid value of EH fire-resistant oil samples under different preset temperature gradients; For each temperature gradient, an independent sub-model is trained to establish the mapping relationship between water molecule parameters, pollution characterization parameters and total acid value at that temperature. The temperature gradient is either an equally spaced temperature range or a non-equally spaced temperature range based on the working temperature range of EH fire-resistant oil.

7. The detection method according to claim 1, characterized in that, A detection branch is also provided, the two ends of which are connected to the oil pipes of the turbine equipment oil tank; the sensor is installed in the detection branch.

8. A computer device, characterized in that, include: Processor and memory; The memory stores program instructions; The processor is configured to run the program instructions to perform the detection method as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The device stores program instructions that are executed to perform the detection method as described in any one of claims 1-7.

10. A computer program product, characterized in that, include: Program instructions for performing the detection method as described in any one of claims 1-7.

Images