Intelligent Operation Method Based on Hydraulic System Data Analysis of Hydropower Station Unit Governor

Abstract

This intelligent operation method, based on data analysis of the hydraulic system of the governor of a hydropower station unit, aims to solve the problems of limited alarm conditions and slow response speed in existing hydropower station monitoring systems through innovative multi-dimensional data feature extraction methods. By extracting feature values ​​across horizontal dimensions, time periods, and time intervals, and combining this with equipment operating patterns, the system achieves more accurate and timely automatic alarms for the operating status of the hydraulic system of the governor of the hydropower station unit.

Description

Technical Field

[0001] This invention relates to the field of hydro turbine operation technology, specifically to a smart operation method based on data analysis of the hydraulic system of the governor of a hydropower station unit. Background Technology

[0002] The speed control system of a hydropower station unit is used to control the opening of the unit's guide vanes and is an important component of the unit's control system. It mainly consists of the electrical part of the governor and the mechanical and hydraulic parts. The hydraulic system is controlled by a PLC in the governor control cabinet, which maintains the pressure, oil temperature, liquid level and other characteristic quantities of the hydraulic system within the normal operating range, providing stable and reliable operating oil pressure for the mechanical actuators of the governor.

[0003] Hydraulic system pressure tanks are classified into oil-air tanks and air tanks based on the internal medium. The oil-air tank contains turbine oil and compressed air; the air tank contains only compressed air. The upper parts of the two tanks are connected by a pipeline. The overall oil-air ratio is approximately 1:2. During adjustment, the consumed oil is replenished by the oil pump, and the consumed compressed air is replenished by an automatic air replenishment device.

[0004] The automatic air replenishment device replenishes the pressurized oil and gas tank automatically based on the oil level and pressure within the pressure tank. When the oil level in the pressure tank reaches the upper limit of the normal oil level range but the pressure is lower than the normal operating pressure, the automatic air replenishment device activates, supplying compressed air to the pressure tank and stopping after a fixed time. If the air replenishment conditions are met, air replenishment continues, with each replenishment lasting the same amount of time.

[0005] The current hydraulic system only triggers alarms when analog quantities such as hydraulic system level and pressure exceed thresholds or when switching quantities are activated.

[0006] The shortcomings of existing technologies include: a large margin between analog threshold values ​​and normal equipment operating values, which prevents timely detection of equipment anomalies. When an analog threshold exceeds the alarm limit, the equipment may have already been in an abnormal state for a considerable period, shortening the response time for maintenance personnel and increasing the likelihood of forced equipment shutdown. For defects such as air or oil leaks in hydraulic systems, simple analog alarms are insufficient for detection, requiring manual analysis using additional software to analyze data such as the frequency of automatic air replenishment actions, the time interval of oil pump loading, and the hydraulic system level. This approach suffers from slow fault response, incomplete monitoring, and difficulty in meeting real-time data requirements. Therefore, improving the operating efficiency and safety of hydropower station units urgently requires an automatic alarm method and system based on real-time data analysis. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide a smart operation method based on data analysis of the hydraulic system of the governor of a hydropower station unit. Through an innovative multi-dimensional data feature extraction method, it aims to solve the problems of single alarm conditions and slow response speed in existing hydropower station monitoring systems. The system utilizes feature value extraction in the horizontal, time period, and time period intervals, combined with equipment operation patterns, to achieve more accurate and timely automatic alarm of the operating status of the hydraulic system of the governor of the hydropower station unit.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0009] A smart operation method based on data analysis of the hydraulic system of the governor of a hydropower station includes the following steps:

[0010] Step 1: Construction of the feature database: Based on the monitoring data such as pressure and liquid level of the speed-regulating hydraulic system, select the equipment operating parameters such as the loading time and unloading time of the hydraulic system oil pump and the interval time of the automatic air replenishment valve to create a user-defined feature database.

[0011] Step 2, Construction of the early warning module: The early warning module is constructed using algorithms and logic. By setting logic and thresholds, the early warning module will generate an early warning signal immediately when the trend of the equipment operating parameters becomes abnormal or deviates from the set value.

[0012] Step 3: Establish an early warning mechanism: Establish an early warning mechanism for issues such as shortened oil pump loading interval, long oil pump loading operation time, excessive air replenishment frequency of unit governor, hydraulic system leakage, and water ingress into the hydraulic system.

[0013] The construction method in Step 1 above is as follows:

[0014] Horizontal feature value extraction: Horizontal feature values ​​are extracted from similar data, including empirical values ​​of oil pump loading interval time, oil pump loading time, and automatic air replenishment valve air replenishment interval time.

[0015] Empirical value of oil pump loading interval time: The oil pump loading interval time is calculated based on the difference between two loading times during continuous operation of the same oil pump. The average value of the oil pump loading interval time under automatic loading and unloading state within six months of the newly put into operation is taken as the empirical value of oil pump loading interval time.

[0016] Experience value for hydraulic pump loading time: The difference between the unloading time and the loading time of the same hydraulic pump during continuous operation is recorded as the hydraulic pump loading time. 80% of the average hydraulic pump loading time under automatic loading and unloading state within six months of a newly put into operation hydraulic pump is taken as the experience value for the pump loading interval time.

[0017] Experience value for automatic air replenishment valve air replenishment interval: 80% of the average air replenishment interval of the automatic air replenishment valve when the unit is not under maintenance in the past two years is taken as the experience value for automatic air replenishment valve air replenishment interval.

[0018] The construction method in Step 2 above is as follows:

[0019] Feature extraction for a given time period: extracting feature values ​​from data within a specified time period;

[0020] Data comparison across time periods: Compare data from different time periods to capture differences in device status;

[0021] Steady-state condition settings: Different data have different steady-state conditions, including grid-connected state, load greater than 60%, and delay time conditions.

[0022] The early warning method for the shortened loading interval of the internal pressure oil pump in Step 3 above is as follows:

[0023] Grid-connected status + load greater than 60% + delay An alarm will sound if the oil pressure pump loading interval is less than 0.8 times the average of the previous month; or if the load is greater than 60% and there is a delay in the grid-connected state. Hourly average loading interval of hydraulic pump over 3 hours < Then an alarm will be triggered;

[0024] in: This represents the empirical value of the oil pump loading interval in the database.

[0025] The early warning mechanism in Step 3 above, which involves a long internal pressure oil pump loading and running time warning method, is as follows:

[0026] Grid-connected status + load greater than 60% + delay Hours, hydraulic pump loading and running time > An alarm will be triggered if the active power fluctuation does not exceed 5% during the loading period;

[0027] in: This represents the empirical value for the loading time of the internal pressure oil pump in the database.

[0028] The early warning method for excessive air replenishment frequency of the unit governor in Step 3 above is as follows:

[0029] The interval between gas replenishment is less than That is, to sound an alarm;

[0030] in: This represents the empirical value for the automatic air replenishment valve air replenishment interval within the database.

[0031] The hydraulic system leakage early warning method within the early warning mechanism in Step 3 above is as follows:

[0032] Based on the oil collection tank level H and area And the liquid level h and area of ​​the pressurized oil tank Calculate the total oil volume of the hydraulic system *H+ *h, an alarm will be triggered when the total fuel quantity decreases by 3% within a certain time H;

[0033] Where: H represents time, taken as 12 hours.

[0034] The water ingress warning method for the hydraulic system within the warning mechanism in Step 3 above is as follows:

[0035] Based on the oil collection tank level H and area And the liquid level h and area of ​​the pressurized oil tank Calculate the total oil volume of the hydraulic system *H+ *h, an alarm will be triggered if the total fuel quantity increases by 3% within a certain time H;

[0036] Where: H represents time, taken as 12 hours.

[0037] The present invention provides a smart operation method based on data analysis of the hydraulic system of the governor of a hydropower station unit, which has the following beneficial effects:

[0038] 1. Multi-dimensional feature extraction: By extracting feature values ​​of hydraulic system oil level, oil pump loading time, loading interval, and automatic air replenishment valve air replenishment interval, the system can comprehensively and deeply analyze equipment status, improving data utilization efficiency.

[0039] 2. Multiple alarm types: The alarms include those for shortened oil pump loading intervals, long oil pump loading operation times, excessive air replenishment frequency of the unit governor, hydraulic system leaks, and water ingress into the hydraulic system, providing a more comprehensive coverage of potential problems in the speed-regulating hydraulic system.

[0040] 3. User-defined feature database: This invention allows users to create feature databases according to actual needs, making the system more flexible and applicable to different types of speed-regulating hydraulic systems. Attached Figure Description

[0041] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0042] Figure 1 Flowchart for alarm model of shortened loading interval of hydraulic oil pump Figure 1 ;

[0043] Figure 2 Flowchart for alarm model of shortened loading interval of hydraulic oil pump Figure 2 ;

[0044] Figure 3 Flowchart for a long-running alarm model for hydraulic oil pumps;

[0045] Figure 4 Flowchart for a model where the number of air replenishment cycles to the unit governor is too high;

[0046] Figure 5 Flowchart of a hydraulic system leakage alarm model;

[0047] Figure 6 This is a flowchart of a hydraulic system water ingress alarm model. Detailed Implementation

[0048] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0049] A smart operation method based on data analysis of the hydraulic system of the governor of a hydropower station includes the following steps:

[0050] Step 1: Construction of the feature database: Based on the monitoring data such as pressure and liquid level of the speed-regulating hydraulic system, select the equipment operating parameters such as the loading time and unloading time of the hydraulic system oil pump and the interval time of the automatic air replenishment valve to create a user-defined feature database.

[0051] Step 2, Construction of the early warning module: The early warning module is constructed using algorithms and logic. By setting logic and thresholds, the early warning module will generate an early warning signal immediately when the trend of the equipment operating parameters becomes abnormal or deviates from the set value.

[0052] Step 3: Establish an early warning mechanism: Establish an early warning mechanism for issues such as shortened oil pump loading interval, long oil pump loading operation time, excessive air replenishment frequency of unit governor, hydraulic system leakage, and water ingress into the hydraulic system.

[0053] The construction method in Step 1 above is as follows:

[0054] Horizontal feature value extraction: Horizontal feature values ​​are extracted from similar data, including empirical values ​​of oil pump loading interval time, oil pump loading time, and automatic air replenishment valve air replenishment interval time.

[0055] Empirical value of oil pump loading interval time: The oil pump loading interval time is calculated based on the difference between two loading times during continuous operation of the same oil pump. The average value of the oil pump loading interval time under automatic loading and unloading state within six months of the newly put into operation is taken as the empirical value of oil pump loading interval time.

[0056] Experience value for hydraulic pump loading time: The difference between the unloading time and the loading time of the same hydraulic pump during continuous operation is recorded as the hydraulic pump loading time. 80% of the average hydraulic pump loading time under automatic loading and unloading state within six months of a newly put into operation hydraulic pump is taken as the experience value for the pump loading interval time.

[0057] Experience value for automatic air replenishment valve air replenishment interval: 80% of the average air replenishment interval of the automatic air replenishment valve when the unit is not under maintenance in the past two years is taken as the experience value for automatic air replenishment valve air replenishment interval.

[0058] The construction method in Step 2 above is as follows:

[0059] Feature extraction over a given time period: Extracting feature values ​​from data within a specified time period to better analyze the equipment's operating status;

[0060] Data comparison across time periods: Compare data from different time periods to capture differences in device status;

[0061] Steady-state condition settings: Different data have different steady-state conditions, including grid-connected state, load greater than 60%, and delay time conditions.

[0062] The early warning method for the shortened loading interval of the internal pressure oil pump in Step 3 above is as follows:

[0063] Grid-connected status + load greater than 60% + delay An alarm will sound if the oil pressure pump loading interval is less than 0.8 times the average of the previous month; or if the load is greater than 60% and there is a delay in the grid-connected state. Hourly average loading interval of hydraulic pump over 3 hours < Then an alarm will be triggered;

[0064] in: This represents the empirical value of the oil pump loading interval in the database.

[0065] The early warning mechanism in Step 3 above, which involves a long internal pressure oil pump loading and running time warning method, is as follows:

[0066] Grid-connected status + load greater than 60% + delay Hours, hydraulic pump loading and running time > An alarm will be triggered if the active power fluctuation does not exceed 5% during the loading period;

[0067] in: This represents the empirical value for the loading time of the internal pressure oil pump in the database.

[0068] The early warning method for excessive air replenishment frequency of the unit governor in Step 3 above is as follows:

[0069] The interval between gas replenishment is less than That is, to sound an alarm;

[0070] in: This represents the empirical value for the automatic air replenishment valve air replenishment interval within the database.

[0071] The hydraulic system leakage early warning method within the early warning mechanism in Step 3 above is as follows:

[0072] Based on the oil collection tank level H and area And the liquid level h and area of ​​the pressurized oil tank Calculate the total oil volume of the hydraulic system *H+ *h, an alarm will be triggered when the total fuel quantity decreases by 3% within a certain time H;

[0073] Where: H represents time, taken as 12 hours.

[0074] The water ingress warning method for the hydraulic system within the warning mechanism in Step 3 above is as follows:

[0075] Based on the oil collection tank level H and area And the liquid level h and area of ​​the pressurized oil tank Calculate the total oil volume of the hydraulic system *H+ *h, an alarm will be triggered if the total fuel quantity increases by 3% within a certain time H;

[0076] Where: H represents time, taken as 12 hours.

[0077] Example:

[0078] This invention relates to an automatic alarm method and system for the hydraulic system of a hydropower station unit governor. Through an innovative multi-dimensional data feature extraction method, it aims to solve the problems of existing hydropower station monitoring systems, such as limited alarm conditions and slow response speeds. This system utilizes feature value extraction across horizontal dimensions, time periods, and time intervals, combined with equipment operating patterns, to achieve more accurate and timely automatic alarms for the operating status of the hydropower station unit governor hydraulic system.

[0079] This innovative system not only meets the real-time and comprehensive monitoring requirements of the hydraulic system of the governor of hydropower station units, but also has greater flexibility, adapting to the special operating conditions of different hydropower station units. Through this invention, hydropower station operation and maintenance personnel will be able to monitor equipment status more efficiently, identify potential problems in advance, thereby reducing maintenance costs and improving equipment operating efficiency. The system includes the following steps:

[0080] ① Construction of the feature database: Based on the monitoring data such as pressure and liquid level of the speed-regulating hydraulic system, this invention first selects equipment operating parameters such as the loading time and unloading time of the hydraulic system oil pump and the interval time of the automatic air replenishment valve to create a user-defined feature database.

[0081] ② Construction of the early warning module: Combining equipment operating patterns and the experience of operation and maintenance experts, this invention employs algorithms and logic to construct an early warning module. By setting appropriate logic and thresholds, the early warning module generates an early warning signal immediately when the trend of equipment operating parameters becomes abnormal or deviates from the normal operating experience value.

[0082] The construction method is as follows:

[0083] Horizontal feature value extraction: Extract horizontal feature values ​​from similar data, including but not limited to empirical values ​​of oil pump loading interval time, oil pump loading time, and automatic air replenishment valve air replenishment interval time.

[0084] Feature extraction over a specific time period: extracting feature values ​​from data within a specific time period to better analyze the operating status of the equipment.

[0085] Data comparison across time periods: Compare data from different time periods to capture differences in device status.

[0086] Steady-state condition settings: Different data have different steady-state conditions, including grid-connected state, load greater than 60%, delay time, etc.

[0087] ③ Early Warning Mechanism: The early warning mechanism of this invention can immediately generate warnings when the trend of key parameters becomes abnormal, and can also generate alarms in advance when equipment parameter values ​​deviate from normal operating experience values. Compared with traditional equipment threshold alarms, the early warning mechanism of this invention can eliminate equipment anomalies in their infancy, improving safety and reliability.

[0088] The warning functions are as follows: the oil pressure pump loading interval is too short; the oil pressure pump loading operation time is too long; the unit speed governor replenishes air too often; the hydraulic system leaks; the hydraulic system is flooded with water.

[0089] Alarm logic:

[0090] (1) Alarm for shortened loading interval of hydraulic oil pump

[0091] Grid-connected status + load greater than 60% + delay An alarm will sound if the oil pressure pump loading interval is less than 0.8 times the average of the previous month; or if the load is greater than 60% and there is a delay in the grid-connected state. Hourly average loading interval of hydraulic pump over 3 hours < Then an alarm will be triggered.

[0092] in: This represents the empirical value of the internal pressure oil pump loading interval in the database; the flowcharts are shown in 1 and 2.

[0093] (2) Alarm for long loading and running time of hydraulic oil pump

[0094] Grid-connected status + load greater than 60% + delay Hours, hydraulic pump loading and running time > An alarm will be triggered if the active power fluctuation during loading does not exceed 5%.

[0095] in: This represents the empirical value of the internal pressure oil pump loading time in the database; the flowchart is as follows: Figure 3 As shown in the image.

[0096] (3) The number of times the unit speed governor replenishes air is too high.

[0097] The interval between gas replenishment is less than This will trigger an alarm.

[0098] in: This represents the empirical value for the automatic air replenishment valve's air replenishment interval within the database; the flowchart is as follows: Figure 4 As shown in the image.

[0099] (4) Hydraulic system leakage alarm

[0100] Based on the oil collection tank level H and area And the liquid level h and area of ​​the pressurized oil tank Calculate the total oil volume of the hydraulic system *H+ *h, an alarm will be triggered when the total fuel quantity decreases by 3% within a certain time H.

[0101] Where: H represents time, taken as 12 hours; the flowchart is as follows: Figure 5 As shown in the image.

[0102] (5) Hydraulic system water ingress alarm

[0103] Based on the oil collection tank level H and area And the liquid level h and area of ​​the pressurized oil tank Calculate the total oil volume of the hydraulic system *H+ *h, an alarm will be triggered when the total fuel quantity increases by 3% within a certain time period H.

[0104] Where: H represents time, taken as 12 hours; the flowchart is as follows: Figure 6 As shown in the image.

Claims

1. A smart operation method based on data analysis of a hydroelectric power station generator set governor hydraulic system, characterized in that, The method comprises the following steps: Step 1, construction of a feature database: based on the monitoring data of the speed regulating hydraulic system, such as pressure and liquid level, the device operation parameters of the hydraulic system oil pump loading time, unloading time and automatic air supply valve interval time are selected to create a user-defined feature database; Step 2, construction of a warning module: an algorithm and logic are used to construct a warning module, and when the device operation parameter trend is abnormal or deviates from the set value, the warning module generates a warning signal in time; Step 3, establishment of an early warning mechanism: an early warning mechanism is established for the shortening of the oil pump loading interval time, the long running time of the oil pump, the excessive air supply frequency of the unit governor, the hydraulic system leakage and the water ingress of the hydraulic system; The construction method in Step 1 is as follows: Lateral characteristic value extraction: lateral characteristic values are extracted from similar data, including oil pump loading interval time empirical value, oil pump loading time empirical value and automatic air supply valve air supply interval time empirical value; Oil pump loading interval time empirical value: the difference between the two loading times when the same oil pump is continuously running is calculated to obtain the oil pump loading interval time, and the average value of the oil pump loading interval time under the automatic loading and unloading state of the newly put into operation oil pump within half a year is taken as 80% of the oil pump loading interval time empirical value; Oil pump loading time empirical value: the difference between the unloading time and the loading time when the same oil pump is continuously running is taken as the oil pump loading time, and the average value of the oil pump loading time under the automatic loading and unloading state of the newly put into operation oil pump within half a year is taken as 80% of the oil pump loading interval time empirical value; Automatic air supply valve air supply interval time empirical value: the average value of the automatic air supply valve air supply interval within the past two years under the non-maintenance state of the unit is taken as 80% of the automatic air supply valve air supply interval time empirical value; The early warning method for the shortening of the oil pump loading interval time in the early warning mechanism in Step 3 is as follows: Grid-connected state + load greater than 60% + delay Oil pump loading interval time < last month average * 0.8, alarm Grid + load greater than 60% + delay Hour, oil pump 3 hours average load interval time Then alarm; wherein: represents the empirical value of the loading interval time of the oil pump in the database; The early warning method for the water ingress of the hydraulic system in the early warning mechanism in Step 3 is as follows: According to the oil collecting tank liquid level H, area and the oil press tank liquid level h, area Calculate the total oil volume of the hydraulic system *H+ *h, when the total oil volume increases by 3% within a certain time H, an alarm is reported Wherein: H represents time, and 12 hours are taken.

2. The intelligent operation method based on data analysis of a hydraulic system of a hydroelectric generating set governor according to claim 1, characterized in that, The construction method in Step 2 is as follows: Time period characteristic value extraction: characteristic values are extracted from the data in the set time period; Comparison of data between time periods: different time period data are compared to capture the differences between device states; Steady state condition setting: different data have different steady state conditions, including grid connection state, load greater than 60% and time delay condition.

3. The intelligent operation method based on data analysis of a hydraulic system of a hydroelectric generating set governor according to claim 2, characterized in that, The early warning method for the long running time of the oil pump in the early warning mechanism in Step 3 is as follows: Grid-connected state + load greater than 60% + delay Oil pump loading operation time per hour Active power fluctuation during loading does not exceed 5% wherein: represents the empirical value of the loading time of the oil pump in the database.

4. The intelligent operation method based on data analysis of a hydraulic system of a hydroelectric generating set governor according to claim 3, characterized in that, The early warning method for the excessive air supply frequency of the unit governor in the early warning mechanism in Step 3 is as follows: The air supplement interval time is less than That is, an alarm; wherein: represents the empirical value of the air supply interval time of the air supply valve in the database.

5. The intelligent operation method based on data analysis of a hydraulic system of a hydroelectric generating set governor according to claim 4, characterized in that, The early warning method for the hydraulic system leakage in the early warning mechanism in Step 3 is as follows: According to the oil collecting tank liquid level H, area and the oil press tank liquid level h, area Calculate the total oil quantity of the hydraulic system *H+ *h, when the total oil quantity reduction ratio reaches 3% within a certain time H, an alarm is reported Wherein: H represents time, and 12 hours are taken.

Images