Large transformer cooling system leakage fault monitoring method

Abstract

The invention provides a large transformer cooling system leakage fault monitoring method comprising the following steps: building a transformer cooler system model, and monitoring pressure and temperature on two sides of a cooler pipeline to obtain reference data; establishing a pressure drop model of oil flow and water flow of the cooler pipeline; the method comprises the following steps: enabling lost oil and water caused by a leakage fault to be equivalent to virtual loads, and establishing an oil flow and water flow system temperature change dynamic model considering the leakage fault by considering oil flow and water flow pipeline energy flow and topological change caused by the virtual loads; the pressure drop model and the temperature change model jointly form a cooler dynamic flow model to describe the oil and water leakage process in pipeline operation; the difference between the minimized reference data and the operation data obtained by the fault detection method is used as a target function to optimize the dynamic coefficient of the model, so that the corrected cooling system leakage fault monitoring model is obtained, and the oil flow and water flow leakage early warning of the main transformer cooler is more accurately carried out.

Description

Technical Field

[0001] The present invention relates to the field of transformer cooling system monitoring, and in particular to a leakage fault monitoring method for a large transformer cooling system. Background Art

[0002] The main transformer cooler, a device that dissipates heat from the transformer's core and windings, keeps the transformer's operating temperature within a normal range. Traditionally, large hydropower plants use forced oil circulation and water cooling to cool the main transformer. An oil pump transports hot transformer oil to the cooler, where it exchanges heat with the cooling water, lowering the insulating oil temperature and, in turn, the main transformer winding and core temperatures. This keeps the main transformer operating at an appropriate temperature, ensuring safe and stable operation and ensuring the delivery of high-quality electricity to all parts of the country.

[0003] The main transformer cooler is a critical safety feature of the main transformer in a hydroelectric power plant. However, due to long-term operation in high-voltage, high-temperature, strong magnetic fields, and vibration environments, it is also susceptible to corrosion from the current-carrying medium, air medium, and environmental microorganisms. Furthermore, oil and water pipelines have a certain transmission distance. As a result, leakage failures are common, causing oil and water loss and reducing resource supply. Whether leaking in the oil or water line, it will inevitably disrupt the balance of the existing transformer cooling system, ultimately affecting the transformer's cooling efficiency and even causing the cooler to completely shut down, leading to unit shutdown. Therefore, the control logic of the main transformer cooler urgently needs to account for pipeline leakage failures to prevent the cooler from completely shutting down.

[0004] At present, large hydropower plants generally install leakage sensors at one part of the cooler pipe (the cooler pipes mentioned in the present invention refer to oil pipes and water pipes, and the transmission forms of oil pipes and water pipes are the same, so the present invention uses pipes to replace oil pipes and water pipes) to monitor the leakage of oil and water flows. However, for pipeline leakage failures, their impact on pipeline pressure is small and difficult to detect, and usually can only be detected through human patrols. Compared with traditional power systems, changes in some physical quantities of oil and water systems will last for a long time, that is, they have significant dynamic characteristics. In oil and water flow systems, due to their slow flow speed, there is a time delay in the transmission process. If it is not discovered in time, it may cause the main transformer cooler to stop completely, and then cause the unit to shut down. Summary of the Invention

[0005] To solve the current technical problems, the main purpose of the present invention is to provide a leakage fault monitoring method for the cooling system of a large transformer, which aims to monitor the leakage faults of the oil flow and water flow of the main transformer cooler. It can solve the generator shutdown problems caused by two aspects: ① The complete shutdown of the cooler directly leads to the shutdown of the unit; ② The partial shutdown of the cooler, but the existing cooler cannot reduce the transformer temperature to a safe range, thereby triggering the gas protection tripping and causing the unit to shut down.

[0006] In order to achieve the above technical features, the purpose of the present invention is achieved as follows: A method for monitoring leakage faults in a large transformer cooling system comprises the following steps:

[0007] Step 1: Build a transformer cooling system model:

[0008] Based on the on-site cooler environment of a large hydropower plant transformer, a set of cooler models identical to the on-site environment was built to construct a test system.

[0009] Step 2: Get temperature change and pressure drop data:

[0010] Based on the transformer cooling system model built in step 1, pressure sensors and temperature sensors are installed at both ends of the oil and water flow pipes to monitor the pressure and temperature on both sides of the cooler pipes to obtain baseline data;

[0011] Step 3: Establish a pressure drop model:

[0012] Establish a pressure drop model based on the oil and water flows in the cooler pipes;

[0013] Step 4: Establish a temperature change model:

[0014] The oil and water lost due to leakage failure are equivalent to virtual loads. By considering the energy flow and topology changes of the oil and water flow pipelines caused by virtual loads, a temperature change model of the oil and water flow system considering leakage failure is established.

[0015] Step 5: Establish the dynamic flow model of the cooler pipeline:

[0016] A dynamic flow model of the cooler pipeline is constructed based on the pressure drop model in step 3 and the temperature change model in step 4 to describe the oil and water leakage process during pipeline operation.

[0017] Step 6: Establish the dynamic flow correction model of the cooler pipeline:

[0018] Based on the cooler pipe dynamic flow model in step 5 and the temperature change and pressure drop data obtained in step 2, the dynamic coefficient of the model is optimized with the objective function of minimizing the difference between the baseline data and the operating data obtained by the fault detection method, thereby obtaining a revised cooler pipe dynamic flow correction model and ultimately forming a cooling system leakage fault monitoring model, thereby more accurately providing early warning of oil and water flow leakage in the main transformer cooler;

[0019] Step 7, control of the cooler of the transformer cooling system:

[0020] During the normal operation of the transformer cooling system, the monitoring results are obtained according to the cooling system leakage fault monitoring model, and the cooler is controlled when a fault occurs to ensure the normal operation of the transformer cooling system.

[0021] The specific construction process of the transformer cooling system model in step 1 is as follows:

[0022] Step 1.1: Build a transformer model based on the structural characteristics of the transformer. The upper end of the transformer is composed of a heater and container 1. Container 1 is filled with oil, and the heater is used to heat the oil to simulate the increase in oil temperature during transformer operation. The lower end of transformer 1 is replaced by container 2, which is filled with oil cooled by a cooler.

[0023] Step 1.2, partially filling the container 1 with oil, setting different oil temperatures, and heating the oil with a heater;

[0024] Step 1.3: Transfer the hot oil to the cooler through the oil pump and cool it with water. At the same time, read the oil temperature, water temperature, oil pressure, and water pressure at the beginning and end of the oil and water pipes, and record the oil and water quantities.

[0025] In step 1.4, the changes in oil temperature, water temperature, oil pressure, and water pressure are calculated respectively. The characteristic set of temperature or pressure changes at both ends of the oil pipe and water pipe at different outlet oil temperatures at different times is obtained and used to modify the dynamic flow model of the cooler pipe.

[0026] The same oil temperature and water temperature in step 1 need to be tested multiple times, and the final result is the average value of each test to ensure accuracy.

[0027] The specific process of establishing the dynamic flow model of the cooler pipeline in step 5 is as follows:

[0028] Step 5.1, calculation of transformer cooling system pressure change:

[0029] The oil and water in the cooling system of a large transformer will produce a voltage drop during the transmission process, and the calculation method for the two is the same. The voltage drop is calculated according to Darcy's formula:

[0030]

[0031] The friction coefficient is calculated using the Kleinbrock formula:

[0032]

[0033] According to the flow formula The final calculation formula for pressure drop is derived from formula (1):

[0034]

[0035] Where, P m 、P n is the oil and water pressure at the start and end nodes of the cooler pipeline; mn is the friction resistance coefficient of the cooler pipe; w mn is the flow rate of hot oil or hot water in the pipeline; D mn is the cooler pipe diameter; ρ represents the density of water; L mn represents the length of the pipeline; Δ is the equivalent roughness of the pipeline, which is a constant; Re is the Reynolds number; υ is the viscosity coefficient, which is a constant; m mn Indicates the working fluid quality of the water supply pipeline mn at different times; A mn Indicates the cross-sectional area of ​​the pipe; is the pipeline flow rate; l is the pipeline number; t is the time;

[0036] Step 5.2, calculation of transformer cooling system temperature change:

[0037] When a leak occurs in the cooler pipe, virtual oil and water loads are used to replace the oil and water lost in the leak. The mass of oil and water flowing out of the pipe is m mn Δt is equal to the mass composition of the oil and water injected into the pipeline from t-t1 to t-t2. The time it takes for oil and water to be transported in the pipeline is called the transmission delay time. After a leakage fault occurs, the oil and water flow rates in the pipeline change, causing the transmission delay time to also change. Based on this, the original pipeline is split into two sections for calculation, and the transmission delay time is redefined as t1Δt = κ1Δt + κ2Δt and t2Δt = κ1Δt + κ3Δt.

[0038]

[0039] In the formula, t, t1, t2 represent time; t-t1, represents time interval; Δt represents time variation; κ1, κ2, κ3 represent intermediate variation, which is the value to be optimized; st represents constraint condition; L in represents the length of the pipeline between node i and node n; L nj represents the length of the pipeline between node j and node n; represents the working fluid quality of the pipeline between node i and node n at time k; represents the working fluid quality of the pipeline between node j and node n at time k; κ, κ', κ" are intermediate variables, and N is a set;

[0040] The oil or water leakage caused by pipeline leakage The calculation method is expressed as:

[0041]

[0042] In the formula, C2 is the flow coefficient, g is a constant, and A lis the area of ​​the leakage hole; is the pipeline pressure; virtual oil or water load Oil or water flow between two sections of pipeline The relationship is expressed as:

[0043]

[0044] The oil or water temperature is calculated based on the changed oil or water flow rate:

[0045]

[0046]

[0047] in:

[0048]

[0049] Where, Indicates the mass of oil or water injected into the pipeline during the period from t-t2+1 to t; Indicates the mass of oil or water injected into the pipeline during the period from t-t1 to t; Respectively represent the head end temperature and the end temperature of the oil return or water return pipe; denote the temperature of the pipeline between node i and node n at time t-t2, k, and t-t1 respectively; denote the temperature of the pipeline between node j and node n at time t-t2, k, and t-t1 respectively;

[0050] The dynamic flow model of the cooler pipe taking leakage into account is constructed as follows:

[0051]

[0052] Where ΔP and ΔT represent the pressure and temperature changes at the beginning and end of the pipeline, respectively.

[0053] The specific process of establishing the dynamic flow correction model of the cooler pipeline in step 6 is as follows:

[0054] Use the historical data obtained from the transformer cooling system model test in step 1 to As the objective function, the dynamic coefficients α and β are optimized, and a three-dimensional Sigmoid dynamic flow correction model for the cooler pipeline is proposed:

[0055]

[0056]

[0057] Where ΔP i st , ΔT ist They represent the pressure and temperature variation data obtained through the transformer cooling system model test; ΔP i ', ΔT i 'represent the pressure and temperature variation data obtained by calculation; ΔP' and ΔT'represent the correction values ​​of the pressure and temperature variation at the beginning and end of the pipeline respectively; α and β are the correction coefficients of the pressure and temperature variation respectively, which are obtained by formula (13). α and β are dynamic coefficients, and their values ​​are calculated based on the pressure and temperature parameters obtained from the transformer cooler system model to ensure the calculation accuracy; b is a constant; e is a natural constant;

[0058] w1, w2, w3, w4, w5, and w6 are variables to be optimized.

[0059] In the model optimization process in step 6, there may be overfitting data. Based on this, the ElasticNet regularization technology is introduced to penalize the parameter size while estimating the parameters, so that some parameters tend to zero, thereby achieving feature selection. At this time, the objective function becomes:

[0060] Where, represents the regularization term;

[0061] Based on this, a gradient descent-based optimization method is used to iteratively adjust the parameters so that the objective function: The value of gradually decreases until the optimal parameter estimate is found, specifically, the initialization parameter χ i After that, calculate Gradient with respect to each parameter:

[0062] Update χ according to the gradient i 、 Specifically:

[0063]

[0064] Where φ and ν are learning rates, which control the magnitude of each update. Repeat the above steps until the objective function stops decreasing and output the parameters to be optimized w1, w2, w3, w4, w5, and w6.

[0065] The specific calculation process of obtaining the monitoring result according to the cooling system leakage fault monitoring model in step 7 includes:

[0066] Step 7.1, collect test data: Based on the transformer cooler system model, conduct multiple tests to obtain the temperature and pressure of the oil or water at both ends of the pipeline at different initial temperatures, and calculate the changes in temperature and pressure as baseline data. The pressure change is recorded as: The temperature change is recorded as:

[0067] Step 7.2: Calculate the pressure change according to formula (14):

[0068]

[0069] Where, is the pressure change of group 1, group 2, group 3, ..., group K; is the temperature change of group 1, group 2, group 3, ..., group K; is the water flow;

[0070] Step 7.3, calculate the temperature change according to formula (15):

[0071]

[0072] Step 7.4, using the data from step 7.1, As the objective function, we can optimize the dynamic coefficients α and β:

[0073]

[0074] Step 7.5: Output the final leakage fault monitoring result ΔP'=α·ΔP, ΔT'=β·ΔT.

[0075] Controlling the cooler when a fault occurs in step 7 specifically includes:

[0076] In step 7.6, the cooler monitoring system server performs dynamic flow analysis on each main transformer cooler based on the collected analog information of oil flow and water flow, according to ΔP' = α·ΔP and ΔT' = β·ΔT, to obtain the changes in oil and water pressure and temperature in the cooler at adjacent moments.

[0077] Step 7.7: Based on the preset pressure and temperature change thresholds in the cooler monitoring system server, it is determined whether an alarm signal should be issued or a cooler shutdown command should be issued. The result is transmitted to the cooler on-site monitoring system via optical fiber.

[0078] In step 7.8, the cooler on-site monitoring system transmits the information to the PLC digital input module in the main transformer cooler control cabinet, and the PLC executes the alarm signal or the cooler shutdown command.

[0079] The present invention has the following beneficial effects:

[0080] 1. The present invention provides a method for monitoring leakage faults in the cooling system of a large transformer. Compared with existing monitoring methods, this patented method controls the cooler by considering the dynamic changes in the pressure and temperature of the oil and water flows. It can sense the operating conditions of each cooler in advance, detect cooler leakage faults in a timely manner, and arrange maintenance personnel for inspections more quickly to prevent larger accidents: ① All coolers are shut down, directly causing the unit to shut down; ② Partial cooler shutdown, but the existing cooler cannot reduce the transformer temperature to a safe range, thereby triggering a gas protection trip and causing the unit to shut down. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] The present invention will be further described below with reference to the accompanying drawings and examples.

[0082] Figure 1 It is the cooler system model of the present invention.

[0083] Figure 2 This is the transformer equivalent system of the present invention.

[0084] Figure 3 Schematic diagram of pipeline leakage in the present invention.

[0085] Figure 4 This is a structural diagram of the present invention. DETAILED DESCRIPTION

[0086] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0087] Example 1:

[0088] The present invention is a method for monitoring leakage faults in a large transformer cooling system, which is described in detail as follows:

[0089] 1. Basic introduction of main transformer cooler system:

[0090] 1) Each transformer is equipped with 6 coolers;

[0091] 2) Each cooler is equipped with a local monitoring system (MS), a monitoring system server (MSS), a programmable logic controller (PLC), a main transformer cooler power cabinet, a main transformer cooler control cabinet, etc.

[0092] 3) The local monitoring system MS is connected to the PLC of the cooler control cabinet through optical fiber to collect analog signals such as oil flow and water flow, and digital signals such as switch status in the PLC;

[0093] 4) The local monitoring system MS is connected to the monitoring system server MSS through optical fiber, and the collected analog and digital signals are transmitted to the MSS.

[0094] 5) MSS performs dynamic flow analysis on each main transformer cooler and transmits the analysis results in digital form to the monitoring system MS via optical fiber. MS then transmits the information to the PLC digital input module in the main transformer cooler control cabinet, and the PLC outputs the execution command.

[0095] 2. Correction model of the cooler pipeline dynamic flow model taking leakage into account:

[0096] In large hydroelectric power plants, certain cooler pipes experience high damage rates. The high-temperature fluids of oil, water, and air corrode the pipe surfaces, potentially causing small pores to rupture and leak oil and water. This is one of the main causes of this high damage rate. Therefore, a detailed model that can describe the cooler leakage process is necessary.

[0097] The present invention provides a method for monitoring leakage faults in a large transformer cooling system, as detailed below:

[0098] Step 1: Build a transformer cooling system model:

[0099] According to the on-site environment of the cooler in a large hydropower plant, a set of cooler models identical to the on-site environment is built to construct a test system, such as Figure 1 It should be noted that Figure 1 The upper end of the transformer in the figure is composed of a heater and a container 1. Container 1 is filled with oil. The heater is used to heat the oil to simulate the increase in oil temperature during transformer operation. The lower end of the transformer is replaced by a container 2. Container 2 is filled with oil cooled by a cooler. Figure 2 shown.

[0100] The test simulation process is as follows:

[0101] Step 1.1: Build a transformer model based on the structural characteristics of the transformer. The upper end of the transformer is composed of a heater and container 1. Container 1 is filled with oil, and the heater is used to heat the oil to simulate the increase in oil temperature during transformer operation. The lower end of transformer 1 is replaced by container 2, which is filled with oil cooled by a cooler.

[0102] Step 1.2, partially filling the container 1 with oil, setting different oil temperatures, and heating the oil with a heater;

[0103] Step 1.3, transfer the hot oil to the cooler through the oil pump and cool it with water. At the same time, read Figure 1 The oil temperature, water temperature, oil pressure and water pressure at the beginning and end of the oil pipe and water pipe, and record the oil and water volume;

[0104] In step 1.4, the changes in oil temperature, water temperature, oil pressure, and water pressure are calculated. A characteristic set of temperature or pressure changes at the ends of the oil and water pipes at different outlet oil temperatures is obtained at different times. This is used to modify the dynamic flow model of the cooler pipes. It is important to note that multiple tests are required for the same oil and water temperatures, and the final results are averaged to ensure accuracy.

[0105] Step 2: Get temperature change and pressure drop data:

[0106] Based on the transformer cooling system model built in step 1, pressure sensors and temperature sensors are installed at both ends of the oil and water flow pipes to monitor the pressure and temperature on both sides of the cooler pipes to obtain baseline data;

[0107] Step 3: Establish a pressure drop model:

[0108] Establish a pressure drop model based on the oil and water flows in the cooler pipes;

[0109] Step 4: Establish a temperature change model:

[0110] The oil and water lost due to leakage failure are equivalent to virtual loads. By considering the energy flow and topology changes of the oil and water flow pipelines caused by virtual loads, a temperature change model of the oil and water flow system considering leakage failure is established.

[0111] Step 5: Establish the dynamic flow model of the cooler pipeline:

[0112] Step 5.1, calculation of transformer cooling system pressure change:

[0113] The oil and water in the cooling system of a large transformer will produce a voltage drop during the transmission process, and the calculation method for the two is the same. The voltage drop is calculated according to Darcy's formula:

[0114]

[0115] The friction coefficient is calculated using the Kleinbrock formula:

[0116]

[0117]

[0118] According to the flow formula The final calculation formula for pressure drop is derived from formula (1):

[0119]

[0120] Where, P m 、P n is the oil and water pressure at the start and end nodes of the cooler pipeline;mn is the friction resistance coefficient of the cooler pipe; w mn is the flow rate of hot oil or hot water in the pipeline; D mn is the cooler pipe diameter; ρ represents the density of water; L mn represents the length of the pipeline; Δ is the equivalent roughness of the pipeline, which is a constant; Re is the Reynolds number; υ is the viscosity coefficient, which is a constant; m mn Indicates the working fluid quality of the water supply pipeline mn at different times; A mn Indicates the cross-sectional area of ​​the pipe; is the pipeline flow rate; l is the pipeline number; t is the time;

[0121] Step 5.2, calculation of transformer cooling system temperature change:

[0122] When a leak occurs in the cooler pipe, virtual oil and water loads are used to replace the oil and water lost in the leak. The mass of oil and water flowing out of the pipe is m mn Δt is equal to the mass composition of the oil and water injected into the pipeline from t-t1 to t-t2. The time it takes for oil and water to be transported in the pipeline is called the transmission delay time. After a leakage fault occurs, the oil and water flow rates in the pipeline change, causing the transmission delay time to also change. Based on this, the original pipeline is split into two sections for calculation, and the transmission delay time is redefined as t1Δt = κ1Δt + κ2Δt and t2Δt = κ1Δt + κ3Δt.

[0123]

[0124] In the formula, t, t1, t2 represent time; t-t1, represents time interval; Δt represents time variation; κ1, κ2, κ3 represent intermediate variation, which is the value to be optimized; st represents constraint condition; L in represents the length of the pipeline between node i and node n; L nj represents the length of the pipeline between node j and node n; represents the working fluid quality of the pipeline between node i and node n at time k; represents the working fluid quality of the pipeline between node j and node n at time k; κ, κ', κ" are intermediate variables, and N is a set;

[0125] The oil or water leakage caused by pipeline leakage The calculation method is expressed as:

[0126]

[0127] In the formula, C2 is the flow coefficient, g is a constant, and A l is the area of ​​the leakage hole; is the pipeline pressure; virtual oil or water load Oil or water flow between two sections of pipeline The relationship is expressed as:

[0128]

[0129] The oil or water temperature is calculated based on the changed oil or water flow rate:

[0130]

[0131]

[0132] in:

[0133]

[0134] Where, Indicates the mass of oil or water injected into the pipeline during the period from t-t2+1 to t; Indicates the mass of oil or water injected into the pipeline during the period from t-t1 to t; Respectively represent the head end temperature and the end temperature of the oil return or water return pipe; denote the temperature of the pipeline between node i and node n at time t-t2, k, and t-t1 respectively; denote the temperature of the pipeline between node j and node n at time t-t2, k, and t-t1 respectively;

[0135] The dynamic flow model of the cooler pipe taking leakage into account is constructed as follows:

[0136]

[0137] Where ΔP and ΔT represent the pressure and temperature changes at the beginning and end of the pipeline, respectively.

[0138] Step 6: Establish the dynamic flow correction model of the cooler pipeline:

[0139] Based on the cooler pipe dynamic flow model in step 5 and the temperature change and pressure drop data obtained in step 2, the dynamic coefficient of the model is optimized with the objective function of minimizing the difference between the baseline data and the operating data obtained by the fault detection method, thereby obtaining a revised cooler pipe dynamic flow correction model and ultimately forming a cooling system leakage fault monitoring model, thereby more accurately providing early warning of oil and water flow leakage in the main transformer cooler;

[0140] Specifically include:

[0141] In order to ensure the accuracy of the calculation results of the model of the present invention, the historical data obtained from the transformer cooling system model test in step 1 is used to As the objective function, the dynamic coefficients α and β are optimized, and a three-dimensional Sigmoid dynamic flow correction model for the cooler pipeline is proposed:

[0142]

[0143]

[0144] Where ΔP i st , ΔT i st They represent the pressure and temperature variation data obtained through the transformer cooling system model test; ΔP i ', ΔT i 'represent the pressure and temperature variation data obtained by calculation; ΔP' and ΔT'represent the correction values ​​of the pressure and temperature variation at the beginning and end of the pipeline respectively; α and β are the correction coefficients of the pressure and temperature variation respectively, which are obtained by formula (13). α and β are dynamic coefficients, and their values ​​are calculated based on the pressure and temperature parameters obtained from the transformer cooler system model to ensure the calculation accuracy; b is a constant; e is a natural constant;

[0145] w1, w2, w3, w4, w5, and w6 are variables to be optimized.

[0146] It should be noted that during the model optimization process, there may be overfitting data. Therefore, this embodiment introduces the Elastic Net regularization technology to penalize the parameter size while estimating the parameters, so that some parameters tend to zero, thereby achieving feature selection. Therefore, the objective function becomes:

[0147]

[0148] In the above formula, represents the regularization term.

[0149] Therefore, this application adopts the optimization method based on gradient descent to iteratively adjust the parameters so that the objective function The value of gradually decreases until the optimal parameter estimate is found, specifically, the initialization parameter χ i After that, calculate Gradient with respect to each parameter:

[0150]

[0151] Update based on gradient Specifically:

[0152]

[0153] Where φ and ν are learning rates, which control the magnitude of each update. Repeat the above steps until the objective function stops decreasing, and output the parameters to be optimized, w1, w2, w3, w4, w5, and w6.

[0154] Step 7, control of the cooler of the transformer cooling system:

[0155] The present invention provides a method for monitoring leakage faults in a large transformer cooling system, and the calculation steps are shown in Algorithm 1.

[0156] Input: Historical data of pressure variation of group K main transformer cooler

[0157] K group main transformer cooler temperature change historical data

[0158] Output: Calculation model of temperature and pressure changes on both sides of the pipeline of large transformer cooling system;

[0159] The specific calculation process includes:

[0160] Step 7.1, collect test data: Based on the transformer cooler system model, conduct multiple tests to obtain the temperature and pressure of the oil or water at both ends of the pipeline at different initial temperatures, and calculate the changes in temperature and pressure as baseline data. The pressure change is recorded as: The temperature change is recorded as:

[0161] Step 7.2: Calculate the pressure change according to formula (14):

[0162]

[0163] Where, is the pressure change of group 1, group 2, group 3, ..., group K; is the temperature change of group 1, group 2, group 3, ..., group K; is the water flow;

[0164] Step 7.3, calculate the temperature change according to formula (15):

[0165]

[0166] Step 7.4, using the data from step 7.1, As the objective function, we can optimize the dynamic coefficients α and β:

[0167]

[0168] Step 7.5: Output the final leakage fault monitoring result ΔP'=α·ΔP, ΔT'=β·ΔT.

[0169] Therefore, after obtaining the modified model through Algorithm 1, the cooler is controlled according to the following steps:

[0170] In step 7.6, the cooler monitoring system server (MSS) performs dynamic flow analysis on each main transformer cooler based on the collected analog information of oil flow and water flow, according to ΔP' = α·ΔP and ΔT' = β·ΔT, to obtain the changes in oil and water pressure and temperature in the cooler at adjacent moments.

[0171] Step 7.7: Based on the preset pressure and temperature change thresholds in the cooler monitoring system server (MSS), it is determined whether an alarm signal should be issued or a cooler shutdown command should be issued. The result is transmitted to the cooler on-site monitoring system (MS) via optical fiber.

[0172] In step 7.8, the cooler local monitoring system (MS) transmits the information to the PLC digital input module in the main transformer cooler control cabinet, and the PLC executes the alarm signal or the cooler shutdown command.

Claims

1. A method for monitoring leakage faults in a large transformer cooling system, characterized in that: The following steps are involved: Step 1: Build a transformer cooling system model: Based on the on-site cooler environment of a large hydropower plant transformer, a set of cooler models identical to the on-site environment was built to construct a test system. Step 2: Get temperature change and pressure drop data: Based on the transformer cooling system model built in step 1, pressure sensors and temperature sensors are installed at both ends of the oil and water flow pipes to monitor the pressure and temperature on both sides of the cooler pipes to obtain baseline data; Step 3: Establish a pressure drop model: Establish a pressure drop model based on the oil and water flows in the cooler pipes; Step 4: Establish a temperature change model: The oil and water lost due to leakage failure are equivalent to virtual loads. By considering the energy flow and topology changes of the oil and water flow pipelines caused by virtual loads, a temperature change model of the oil and water flow system considering leakage failure is established. Step 5: Establish the dynamic flow model of the cooler pipeline: A dynamic flow model of the cooler pipeline is constructed based on the pressure drop model in step 3 and the temperature change model in step 4 to describe the oil and water leakage process during pipeline operation. Step 6: Establish the dynamic flow correction model of the cooler pipeline: Based on the cooler pipe dynamic flow model in step 5 and the temperature change and pressure drop data obtained in step 2, the dynamic coefficient of the model is optimized with the objective function of minimizing the difference between the baseline data and the operating data obtained by the fault detection method, thereby obtaining a revised cooler pipe dynamic flow correction model and ultimately forming a cooling system leakage fault monitoring model, thereby more accurately providing early warning of oil and water flow leakage in the main transformer cooler; Step 7, control of the cooler of the transformer cooling system: During the normal operation of the transformer cooling system, the monitoring results are obtained according to the cooling system leakage fault monitoring model, and the cooler is controlled when a fault occurs to ensure the normal operation of the transformer cooling system.

2. A method for monitoring leakage faults in a large transformer cooling system according to claim 1, characterized in that: The specific construction process of the transformer cooling system model in step 1 is as follows: Step 1.1: Build a transformer model based on the structural characteristics of the transformer. The upper end of the transformer is composed of a heater and container 1. Container 1 is filled with oil, and the heater is used to heat the oil to simulate the increase in oil temperature during transformer operation. The lower end of transformer 1 is replaced by container 2, which is filled with oil cooled by a cooler. Step 1.2, partially filling the container 1 with oil, setting different oil temperatures, and heating the oil with a heater; Step 1.3: Transfer the hot oil to the cooler through the oil pump and cool it with water. At the same time, read the oil temperature, water temperature, oil pressure, and water pressure at the beginning and end of the oil and water pipes, and record the oil and water quantities. In step 1.4, the changes in oil temperature, water temperature, oil pressure, and water pressure are calculated respectively. The characteristic set of temperature or pressure changes at both ends of the oil pipe and water pipe at different outlet oil temperatures at different times is obtained and used to modify the dynamic flow model of the cooler pipe.

3. A method for monitoring leakage faults in a large transformer cooling system according to claim 2, characterized in that: The same oil temperature and water temperature in step 1 need to be tested multiple times, and the final result is the average value of each test to ensure accuracy.

4. A method for monitoring leakage faults in a large transformer cooling system according to claim 3, characterized in that: The specific process of establishing the dynamic flow model of the cooler pipeline in step 5 is as follows: Step 5.1, calculation of transformer cooling system pressure change: The oil and water in the cooling system of a large transformer will produce a voltage drop during the transmission process, and the calculation method for the two is the same. The voltage drop is calculated according to Darcy's formula: The friction coefficient is calculated using the Kleinbrock formula: According to the flow formula The final calculation formula for pressure drop is derived from formula (1): Where, P m 、P n is the oil and water pressure at the start and end nodes of the cooler pipeline; mn is the friction resistance coefficient of the cooler pipe; w mn is the flow rate of hot oil or hot water in the pipeline; D mn is the cooler pipe diameter; ρ represents the density of water; L mn represents the length of the pipeline; Δ is the equivalent roughness of the pipeline, which is a constant; Re is the Reynolds number; υ is the viscosity coefficient, which is a constant; m mn Indicates the working fluid quality of the water supply pipeline mn at different times; A mn Indicates the cross-sectional area of ​​the pipe; is the pipeline flow rate; l is the pipeline number; t is the time; Step 5.2, calculation of transformer cooling system temperature change: When a leak occurs in the cooler pipe, virtual oil and water loads are used to replace the oil and water lost in the leak. The mass of oil and water flowing out of the pipe is m mn Δt is equal to the mass composition of the oil and water injected into the pipeline from t-t1 to t-t2. The time it takes for oil and water to be transported in the pipeline is called the transmission delay time. After a leakage fault occurs, the oil and water flow rates in the pipeline change, causing the transmission delay time to also change. Based on this, the original pipeline is split into two sections for calculation, and the transmission delay time is redefined as t1Δt = κ1Δt + κ2Δt and t2Δt = κ1Δt + κ3Δt. In the formula, t, t1, t2 represent time; t-t1, represents time interval; Δt represents time variation; κ1, κ2, κ3 represent intermediate variation, which is the value to be optimized; st represents constraint condition; L in represents the length of the pipeline between node i and node n; L nj represents the length of the pipeline between node j and node n; represents the working fluid quality of the pipeline between node i and node n at time k; represents the working fluid quality of the pipeline between node j and node n at time k; κ, κ', κ" are intermediate variables, and N is a set; The oil or water leakage caused by pipeline leakage The calculation method is expressed as: In the formula, C2 is the flow coefficient, g is a constant, and A l is the area of ​​the leakage hole; is the pipeline pressure; virtual oil or water load Oil or water flow between two sections of pipeline The relationship is expressed as: The oil or water temperature is calculated based on the changed oil or water flow rate: in: Where, Indicates the mass of oil or water injected into the pipeline during the period from t-t2+1 to t; Indicates the mass of oil or water injected into the pipeline during the period from t-t1 to t; Respectively represent the head end temperature and the end temperature of the oil return or water return pipe; denote the temperature of the pipeline between node i and node n at time t-t2, k, and t-t1 respectively; denote the temperature of the pipeline between node j and node n at time t-t2, k, and t-t1 respectively; The dynamic flow model of the cooler pipe that takes leakage into account is constructed as follows: Where ΔP and ΔT represent the pressure and temperature changes at the beginning and end of the pipeline, respectively.

5. A method for monitoring leakage faults in a large transformer cooling system according to claim 4, characterized in that: The specific process of establishing the dynamic flow correction model of the cooler pipeline in step 6 is as follows: Use the historical data obtained from the transformer cooling system model test in step 1 to As the objective function, the dynamic coefficients α and β are optimized, and a three-dimensional Sigmoid dynamic flow correction model for the cooler pipe is proposed: Where, ΔP i st , ΔT i st They represent the pressure and temperature variation data obtained through the transformer cooling system model test; ΔP i ', ΔT i 'represent the pressure and temperature variation data obtained by calculation; ΔP' and ΔT'represent the correction values ​​of the pressure and temperature variation at the beginning and end of the pipeline respectively; α and β are the correction coefficients of the pressure and temperature variation respectively, which are obtained by formula (13). α and β are dynamic coefficients, and their values ​​are calculated based on the pressure and temperature parameters obtained from the transformer cooler system model to ensure the calculation accuracy; b is a constant; e is a natural constant; w1, w2, w3, w4, w5, and w6 are variables to be optimized.

6. A method for monitoring leakage faults in a large transformer cooling system according to claim 5, characterized in that: In the model optimization process in step 6, there may be overfitting data. Based on this, the Elastic Net regularization technology is introduced to penalize the parameter size while estimating the parameters, so that some parameters tend to zero, thereby achieving feature selection. At this time, the objective function becomes: Where, represents the regularization term; Based on this, a gradient descent-based optimization method is used to iteratively adjust the parameters so that the objective function: The value of gradually decreases until the optimal parameter estimate is found, specifically, the initialization parameter χ i After that, calculate Gradient with respect to each parameter: Update χ according to the gradient i 、 Specifically: Where φ and ν are learning rates, which control the magnitude of each update. Repeat the above steps until the objective function stops decreasing and output the parameters to be optimized w1, w2, w3, w4, w5, and w6.

7. A method for monitoring leakage faults in a large transformer cooling system according to claim 6, characterized in that: The specific calculation process of obtaining the monitoring result according to the cooling system leakage fault monitoring model in step 7 includes: Step 7.1, collect test data: Based on the transformer cooler system model, conduct multiple tests to obtain the temperature and pressure of the oil or water at both ends of the pipeline at different initial temperatures, and calculate the changes in temperature and pressure as baseline data. The pressure change is recorded as: The temperature change is recorded as: Step 7.2: Calculate the pressure change according to formula (14): Where, is the pressure change of group 1, group 2, group 3, ..., group K; is the temperature change of group 1, group 2, group 3, ..., group K; is the water flow; Step 7.3, calculate the temperature change according to formula (15): Step 7.4, using the data from step 7.1, As the objective function, we can optimize the dynamic coefficients α and β: Step 7.5: Output the final leakage fault monitoring result ΔP'=α·ΔP, ΔT'=β·ΔT.

8. A method for monitoring leakage faults in a large transformer cooling system according to claim 7, characterized in that: Controlling the cooler when a fault occurs in step 7 specifically includes: In step 7.6, the cooler monitoring system server performs dynamic flow analysis on each main transformer cooler based on the collected analog information of oil flow and water flow, according to ΔP' = α·ΔP and ΔT' = β·ΔT, to obtain the changes in oil and water pressure and temperature in the cooler at adjacent moments. Step 7.7: Based on the preset pressure and temperature change thresholds in the cooler monitoring system server, it is determined whether an alarm signal should be issued or a cooler shutdown command should be issued. The result is transmitted to the cooler on-site monitoring system via optical fiber. In step 7.8, the cooler on-site monitoring system transmits the information to the PLC digital input module in the main transformer cooler control cabinet, and the PLC executes the alarm signal or the cooler shutdown command.

Images