Hydraulic turbine generator set speed regulation system oil leakage alarm method

Abstract

A method for alarming oil leakage in a hydro-generator speed control system includes the following steps: S1, reading historical time-series data of oil level in the pressure tank and oil level in the sump; S2, performing steady-state screening on the historical time-series data of oil level in the pressure tank and oil level in the sump to obtain steady-state historical time-series data of oil level in the pressure tank and oil level in the sump, H1 and H2; S3, performing linear fitting on H1 and H2 in S2 using the least squares method to obtain the fitting result, where k is the linear coefficient obtained by the least squares method fitting, b is the constant term obtained by the least squares method fitting, and is the estimated value of H2, and the fitting deviation y is calculated. The technical problem to be solved by this invention is to provide an alarm method for oil leakage in a hydro-generator speed control system. This method comprehensively considers the oil level values ​​of the pressure tank and oil sump in the hydro-generator speed control system, and can issue an alarm in the early stage of oil leakage fault.

Description

Technical Field

[0001] This invention relates to the field of operation and maintenance of speed control systems for hydro-generator sets, and in particular to a method for alarming oil leakage in the speed control system of a hydro-generator set. Background Technology

[0002] The speed control system of a hydro-generator unit is a core component. It controls the turbine guide vane opening and adjusts the unit's operating force by driving the water guide mechanism with hydraulic oil. Oil leakage in the speed control system can affect the safe and stable operation of the unit and may also cause environmental pollution. Therefore, early detection and alarming of oil leakage in the speed control system are crucial.

[0003] The hydraulic oil in the speed control system is distributed in pipelines, pressure tanks, and sump tanks. During normal operation, the pipelines are always full of oil, and the oil level in the pressure tank changes dynamically with system operation. Since the total oil volume remains constant, the oil level in the sump tank also changes dynamically with the pressure tank level. Traditional methods for alarming oil leaks in speed control systems involve monitoring the oil level in the sump tank and setting a low-level alarm threshold. However, because the oil level in the sump tank fluctuates within a wide range during normal operation due to changes in system oil pressure, and the total oil volume may differ after each addition or discharge, the sump tank level alarm threshold is often set too low to avoid false alarms. This often results in an alarm being triggered only when the leak has become quite severe, failing to detect faults in their early stages and hindering the safe and stable operation of the equipment. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide a method for oil leakage alarm in the speed regulation system of a hydro-generator unit. This method comprehensively considers the oil level values ​​of the pressure oil tank and the oil collection tank in the speed regulation system of the hydro-generator unit, and can issue an alarm in the early stage of oil leakage fault. The technical principle of this method is as follows:

[0005] The cross-sectional area of ​​the pressure oil tank is S1, and the cross-sectional area of ​​the oil collection tank is S2. After the most recent oil filling and draining, the initial oil level in the pressure oil tank is h1, the initial oil level in the oil collection tank is h2, and the total oil volume of the speed control system is Q. At time t, the oil level in the pressure oil tank is H1, the oil level in the oil collection tank is H2, and the total oil volume of the speed control system is Q + ΔQ, where ΔQ = S1(H1 - h1) * (1 + αP) + S2(H2 - h2), where P is the rated pressure of the speed control system, and α is the compressibility coefficient of the hydraulic oil. When the system is leak-free, ΔQ is approximately 0, i.e.:

[0006] S1(H1-h1)*(1+αP)+S2(H2-h2)≈0

[0007] Right now:

[0008]

[0009] make

[0010] but

[0011] S1, S2, α, and P are system characteristics that do not change with the random oil filling / drainage or the system's operating state; therefore, k is a constant. h1 and h2 reflect the initial state of the system after each oil filling / drainage, and are constants within the time interval when the unit has not been filled / drained. Therefore, H1 and H2 can be collected within a healthy time interval when the speed control system has not been filled / drained for a period of time. After steady-state screening, k is obtained by linear fitting using the least squares method. The value of k does not change with oil filling / drainage. After each oil filling / drainage, only b needs to be updated, which can be done using the following formula.

[0012] b = h2 - kh1

[0013] Where h1 and h2 are the initial oil levels in the pressure tank and the oil collection trough after the most recent oil injection / discharge of the speed control system. During real-time operation, the deviation between the predicted value obtained through the least squares fitting method and the real-time value can be calculated as follows:

[0014] y′=kH1′+b-H2′

[0015] Where H1' is the real-time oil level data of the pressure oil tank and H2' is the real-time oil level data of the oil collection tank.

[0016] By comparing the real-time deviation with the historical health deviation, it can be determined whether an alarm should be triggered.

[0017] The specific steps are as follows:

[0018] A method for alarming oil leakage in the speed control system of a hydro-generator set includes the following steps:

[0019] S1. Read the historical time-series data of the oil level in the pressure oil tank and the oil level in the oil collection tank;

[0020] S2. Perform steady-state screening on the historical time series data of oil level in pressure tank and oil sump to obtain steady-state historical time series data of oil level in pressure tank and oil sump, H1 and H2.

[0021] S3. Perform linear fitting on H1 and H2 in S2 using the least squares method to obtain the fitting results. Where k is the linear coefficient obtained by the least squares fitting method, and b is the constant term obtained by the least squares fitting method. To estimate H2, calculate the fit deviation y.

[0022] S4. Calculate the deviation threshold σ based on the distribution of the fitting deviation y;

[0023] S5. Read the initial oil level h1 of the pressure oil tank and the initial oil level h2 of the oil collection tank after the most recent oil filling and emptying of the speed control system;

[0024] S6. Read the real-time oil level data H1' of the pressure oil tank and the real-time oil level data H2' of the oil collection tank, and determine whether they are steady-state data;

[0025] S7. If the real-time data after S6 is non-steady-state data, then the real-time data will not trigger an alarm. If it is steady-state data, then calculate the real-time deviation y', y'=k*H1'-k*h1+h2-H2', and determine whether y'>σ is true. If it is true, then the real-time data will trigger an oil leak alarm. Otherwise, the real-time data will not trigger an alarm.

[0026] S8. Read the next real-time data and repeat steps S6-S7 above to achieve real-time monitoring and alarm of oil leakage in the speed control system.

[0027] Preferably, in S1, the time interval corresponding to the historical time series data of the pressure oil tank level and the historical time series data of the oil collection tank level is greater than one month, and the speed regulation system has been in normal operation during this period without any depressurization or oil addition / discharge.

[0028] Preferably, the filtering method in S2 is: tracing back time t from the historical time series data of each time point. s If the pump start / stop signal remains stopped and the main pressure regulating valve displacement remains less than the main pressure regulating valve displacement dead zone throughout the traceback time interval, then the pressure tank oil level data and the oil collection tank oil level data at that point in time are steady-state data; otherwise, they are unsteady-state data. The time t... s More than 5 seconds.

[0029] Preferably, the method for calculating the deviation threshold σ in S4 is as follows: take the absolute value of y to obtain |y|, arrange all the data in |y| from largest to smallest, remove the data points in the first 0.1%, and then read the data point that is ranked first after removal, denoted as ymax, σ = 2*ymax.

[0030] Preferably, the determination method in step S6 is: tracing back time t from the time corresponding to the real-time data. s If the pump start / stop signal remains stopped and the main pressure valve displacement remains less than the main pressure valve displacement dead zone throughout the traceback time interval, then H1' and H2' at the current moment are steady-state data; otherwise, they are unsteady-state data. The time t... s More than 5 seconds.

[0031] This invention provides a method for alarming oil leakage in the speed control system of a hydro-generator set, which has the following beneficial effects:

[0032] 1. Taking into account both the oil level in the pressure tank and the oil level in the collection tank, the alarm is more accurate and can issue an alarm in the early stage of oil leakage, giving time to deal with the fault.

[0033] 2. The oil pump start / stop signal and the main pressure regulating valve displacement signal eliminate the value of oil level fluctuation during the dynamic adjustment of the speed regulation system, thus avoiding the influence of the dynamic adjustment process on the early warning result;

[0034] 3. The alarm deviation value is calculated based on historical deviations, avoiding the problem of inaccurate deviations due to human subjective setting;

[0035] 4. k is obtained by automatically fitting historical data, eliminating the need to calculate parameters such as S1, S2, and α, which simplifies the calculation process and improves alarm accuracy.

[0036] 5. Each time the unit is filled or drained, only h1 and h2 need to be updated, making maintenance relatively simple and effectively solving the problem of inconsistent alarm reference values ​​after each filling or draining of the unit. Attached Figure Description

[0037] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0038] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0039] like Figure 1 As shown, S1, read the historical time series data of the oil level in the pressure oil tank and the historical time series data of the oil collection tank. The time interval corresponding to the historical time series data should be greater than 1 month, and the speed regulation system should be in normal operation during this period without depressurization or oil addition / discharge.

[0040] S2. Perform steady-state filtering on the historical time-series data of pressure oil tank levels and oil collection tank levels to obtain steady-state historical time-series data of pressure oil tank levels and oil collection tank levels, H1 and H2. The filtering method is as follows: trace back time t from the historical time-series data of each time point. s If the pump start / stop signal remains stopped and the main pressure regulating valve displacement remains less than the main pressure regulating valve displacement dead zone throughout the traceback time interval, then the oil level data in the pressure tank and the oil level data in the sump at that point in time are steady-state data; otherwise, they are unsteady-state data. The time t... s More than 5 seconds;

[0041] S3. Use the least squares method to perform linear fitting on the historical values ​​of H1 and H2 after screening in S2, and obtain the fitting formula. Where k is the linear coefficient obtained by the least squares fitting method, and b is the constant term obtained by the least squares fitting method. Calculate the estimated value of H2 and the fitting bias.

[0042] S4. Calculate the deviation threshold σ based on the distribution of the fitting deviation y. The calculation method is as follows: take the absolute value of y to obtain |y|, arrange all the data in |y| from largest to smallest, remove the data points in the first 0.1%, and then read the first data point after removal, denoted as ymax, σ = 2*ymax;

[0043] S5. Read the initial oil level h1 of the pressure oil tank and the initial oil level h2 of the oil collection tank after the most recent oil filling and emptying of the speed control system;

[0044] S6. Read the real-time oil level data H1' of the pressure oil tank and the real-time oil level data H2' of the oil collection tank, and determine whether they are steady-state data. The determination method is to trace back time t from the time corresponding to the real-time data. s If the pump start / stop signal remains stopped and the main pressure valve displacement remains less than the main pressure valve displacement dead zone throughout the traceback time interval, then H1' and H2' at the current moment are steady-state data; otherwise, they are unsteady-state data. The time t... s More than 5 seconds;

[0045] S7. If the real-time data after S6 is non-steady-state data, then the real-time data will not trigger an alarm. If it is steady-state data, then calculate the real-time deviation y', y'=k*H1'-k*h1+h2-H2', and determine whether y'>σ is true. If it is true, then the real-time data will trigger an oil leak alarm. Otherwise, the real-time data will not trigger an alarm.

[0046] S8. Read the next real-time data and repeat steps S6-S7 above to achieve real-time monitoring and alarm of oil leakage in the speed control system.

[0047] The following example uses data exported from the online monitoring system of Unit 2 of a power plant as an illustration:

[0048] Step 1: Read the historical time-series data of the pressure oil tank level and the oil collection tank level for the unit for the most recent 3 months (February 1, 2023 to May 1, 2023). During this period, the speed control system has been in normal operation and has not been depressurized or added / discharged.

[0049] Step 2: Perform steady-state screening on the historical time-series data of the pressure tank oil level and the oil collection tank oil level to obtain steady-state historical time-series data of the pressure tank oil level and the oil collection tank oil level, H1 and H2. The screening method is as follows: trace back 10 seconds from the historical time-series data at each time point. If the oil pump start / stop signal is always stopped and the displacement of the main pressure regulating valve is always less than the dead zone of the main pressure regulating valve displacement during the traced time interval, then the pressure tank oil level data and the oil collection tank oil level data at that time point are steady-state data; otherwise, they are non-steady-state data.

[0050] Step 3: Use the least squares method to perform linear fitting on the steady-state data H1 and H2 from Step 2 to obtain the fitting formula. Calculate the fit bias

[0051] Step 4: Take the absolute value of y to obtain |y|, arrange all data in |y| from largest to smallest, remove the first 0.1% of data points, and then read the first data point after removal, denoted as ymax = 14.2. Calculate the deviation threshold σ, σ = 2 * ymax = 28.4;

[0052] Step 5: Read the initial oil level h1 of the pressure oil tank after the most recent oil filling and emptying of the speed control system, which is 2095, and the initial oil level h2 of the oil collection tank, which is 1384.

[0053] Step 6: Read the real-time data of the pressure oil tank level H1' = 1925 and the real-time data of the oil collection tank level H2' = 1430 (18:30:45 on May 7, 2023), and determine whether they are steady-state data. The determination method is as follows: trace back 10 seconds from the time corresponding to the real-time data. If the oil pump start / stop signal is always stopped and the displacement of the main pressure regulating valve is always less than the dead zone of the main pressure regulating valve displacement during the traced time interval, then H1' and H2' at the current time are steady-state data. Otherwise, they are non-steady-state data. After judgment, this data is steady-state data.

[0054] Step 7: Calculate the real-time deviation y', y'=k*H1'-k*h1+h2-H2'=-0.3045*H1'+0.3045*2095+1384-H2'=5.765, determine y'<σ, this real-time data will not trigger an alarm.

[0055] At 15:26:30 on July 4, 2023, the monitoring system reported a low oil level in the oil collection tank. On-site inspection revealed an oil leak in the speed control system pipeline, causing the low oil level in the oil collection tank. Data from June 25 to July 5 was exported, and the real-time deviation of the steady-state data after steady-state screening in step 6 was calculated as y' = -0.3045*H1' + 0.3045*2095 + 1384 - H2'. According to the calculation results, y' > σ was triggered starting at 14:20 on July 2, 2023, which was about 2 days earlier than the on-site monitoring system.

[0056] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A method for alarming oil leakage in the speed regulation system of a hydro-generator set, characterized in that, Includes the following steps: S1. Read the historical time-series data of the oil level in the pressure oil tank and the oil level in the oil collection tank; S2. Steady-state filtering is performed on the historical time-series data of pressure oil tank levels and oil collection tank levels to obtain steady-state historical time-series data of pressure oil tank levels and oil collection tank levels. and ; S3, Use the least squares method to analyze S2. , Perform linear fitting to obtain the fitting result. Where k is the linear coefficient obtained by least squares fitting, and b is the constant term obtained by least squares fitting. for The estimated value is obtained, and the fitting deviation y is calculated, y= ; S4. Calculate the deviation threshold σ based on the distribution of the fitting deviation y; the method for calculating the deviation threshold σ in S4 is: take the absolute value of y to obtain... ,Will All data points are sorted from largest to smallest. The top 0.1% of data points are removed. Then, the first data point in the remaining sorted list is read and denoted as ymax. ; S5. Read the initial oil level of the pressure oil tank after the most recent oil filling and emptying of the speed control system. relative to the initial oil level in the oil collection tank ; S6. Read real-time data of oil level in pressure tank. Real-time data of oil level in oil collection tank And determine whether it is steady-state data; S7. If the real-time data after judgment in S6 is non-steady-state data, then no alarm will be triggered for that real-time data. If it is steady-state data, then the real-time deviation y' will be calculated, where y' = And determine y'> If the condition is met, the real-time data will trigger an oil leak alarm; otherwise, the real-time data will not trigger an alarm. S8. Read the next real-time data and repeat steps S6-S7 above to achieve real-time monitoring and alarm of oil leakage in the speed control system.

2. The method for oil leakage alarm in the speed regulation system of a hydro-generator unit according to claim 1, characterized in that, The time interval between the historical time series data of the pressure oil tank level and the historical time series data of the oil collection tank level in S1 is greater than one month, and the speed regulation system has been in normal operation during this period without depressurization or oil addition / discharge.

3. The method for oil leakage alarm in the speed regulation system of a hydro-generator unit according to claim 1, characterized in that, The filtering method in S2 is as follows: trace back time t from the historical time series data of each time point. s If the pump start / stop signal remains stopped and the main pressure regulating valve displacement remains less than the main pressure regulating valve displacement dead zone throughout the traceback time interval, then the pressure tank oil level data and the oil collection tank oil level data at that point in time are steady-state data; otherwise, they are unsteady-state data. The time t... s More than 5 seconds.

4. The method for oil leakage alarm in the speed regulation system of a hydro-generator unit according to claim 1, characterized in that, The determination method in step S6 is as follows: trace back time t from the time corresponding to the real-time data. s If the pump start / stop signal remains stopped and the main pressure valve displacement remains less than the main pressure valve displacement dead zone throughout the traceback time interval, then the current moment's... and If the data is in steady state, it is considered steady-state data; otherwise, it is considered non-steady-state data. The time t... s More than 5 seconds.

Images