Starting-up process auxiliary decision-making method for giant hydroelectric generating set

Abstract

The invention relates to the technical field of hydropower station automation systems, in particular to an auxiliary decision-making method for a starting process of a giant hydroelectric generating set, which comprises the following steps of: respectively constructing three pieces of sequence data for motors of a water supply system, a high-pressure oil system, a thrust bearing outer circulation system and a water guide bearing outer circulation system; calculating equivalent operation time and single average equivalent operation time of the motor, and constructing an equivalent operation time sequence; predicting the next starting time of the motor; distinguishing an abnormal region and a normal region in the starting time; after starting time abnormity analysis of all motors of the auxiliary system is completed, the motor needing to be started next time is determined; determining limited starting motors of all auxiliary systems; and starting the corresponding motor when the startup process is executed until the startup is needed. Reasonable work of the auxiliary machine motor can be intelligently controlled, the service time of the motor is automatically balanced, and the safety risk is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of hydropower station automation systems, in particular to a startup process auxiliary decision method for a giant hydropower unit. Background Art

[0002] The automatic start-up and shutdown process guarantees the safe and efficient operation of hydropower units. In the operation of the power grid, hydropower units use the automated start-up and shutdown process to achieve the functions of peak load regulation, frequency regulation and emergency standby. The start-up and shutdown process of giant hydropower units will automate the technical water supply system and oil system. The core components of these auxiliary systems are motors. The start-up and shutdown process will automatically control the start and stop of the motor in a specific step and sequence to ensure the safety of the unit. In order to ensure that the unit can work reliably, the motor of the auxiliary system adopts a dual redundant configuration, that is, one main control and one standby. Under normal circumstances, the start-up and shutdown process will only use fixed equipment and cannot be intelligently selected according to the actual use of the motor. After the unit has been running for a long time, the operating time and health status of the main control motor and the standby motor are quite different, and the failure risk of the main control motor will increase. If the motor cannot be started and stopped smoothly, the start-up and shutdown will also be affected, which in turn affects the safety of the unit.

[0003] In actual work, the current startup process cannot automatically select the optimized motor for startup, and it is generally switched manually during operation. The present invention provides an auxiliary decision-making technology for giant hydropower units. When starting up and executing the startup process, the appropriate motor is automatically selected for startup according to the usage of the auxiliary motor, thereby reducing the risk of power operation. Summary of the invention

[0004] The present invention provides a method for assisting decision-making in the startup process of a giant hydropower unit, which automatically selects motors that are preferentially involved in startup and shutdown by predicting, analyzing and comparing the states of the motors, thereby reducing the operational risks. Specifically, the method comprises the following steps:

[0005] Step 1: Construct three sequence data for the motors of the technical water supply system, high-pressure oil system, thrust bearing external circulation system, and water-guided bearing external circulation system: the motor start time sequence T1, the motor stop time sequence T2, and the motor running time sequence T3, and count and clean the sequence data:

[0006] The starting time sequence of the motor is T1, T1 = (t 11 ,t 21 ,t 31 ,···,t 1m1 ), the number of data in the sequence m1 = the number of times the motor is started, and the data value t1 = the starting time. The starting time is the time from when the motor receives the starting command to when the equipment is fully put into operation.

[0007] The motor stop time sequence T2, T2 = (t 21 ,t 22 ,t 23 ,···,t 2m2 ), the number of data in the sequence m2 = the number of times the motor stops, the data value t2 = the stop time, the stop time is the time from the motor receiving the stop command to the complete shutdown of the equipment.

[0008] The running time sequence of the motor is T3, T3 = (t 31 ,t 32 ,t 33 ,···,t 3m3 ), the number of data in the sequence m3 = the number of times the auxiliary motor runs, the data value t3 = the running time, the running time is the time from the equipment being "fully put into operation" to "receiving a stop command".

[0009] Only the data of normal start and stop are counted, and the data of fault shutdown are not counted. Each time the motor is started, if the motor does not shut down due to a fault, T1 will add a new data; each time the motor stops normally, T2 and T3 will add a new data. Yes, m1 = m2 = m3 = m;

[0010] Step 2: Based on step 1, calculate the equivalent running time T of the motor eq and single average equivalent running time T eq / , construct the equivalent running time series T4;

[0011] Specifically, it includes the following sub-steps.

[0012] (1) Equivalent running time calculation T eq Calculation. First, convert each start and stop time of the motor into running time, and then add up all the running times, as follows:

[0013]

[0014] t 1i ,t 2i and t 3i are the starting time when the motor starts for the i-th time, the stopping time when it stops for the i-th time, and the running time when it runs for the i-th time; m1, m2, and m3 are the total number of starts, stops, and runs, respectively, and K is the starting multiple;

[0015] (2) Construct the motor equivalent operation time series T4. The number of data in the series m = the number of times the auxiliary motor operates, and the data value T eq =The sum of the historical equivalent running time of each run.

[0016] T4=(T eq1 ,Teq2 ,T eq3 ,···,T eqm )

[0017] (3) Every time T1, T2, and T3 are updated, a new T is calculated. eq and T eq / , T4 adds a new data.

[0018] Step 3: Based on step 1 and step 2, a machine learning method including but not limited to a support vector machine is used to predict the next start time of the motor. When predicting the start time, T1, T2 and T4 are used as input quantities, and the next start time is used as output quantity.

[0019] Step 4: Distinguish the abnormal area and normal area in the startup time. First, calculate the maximum, minimum and median of the startup time, and use the 80 / 20 rule to determine the normal area and abnormal area. Input the result of step 3 to determine whether the predicted value is in the abnormal area.

[0020] Step 5. After completing the abnormal analysis of the start-up time of all motors in the auxiliary system, determine the motor that should be started next time. Let the two motors of the auxiliary system be represented as motor 1 and motor 2 respectively, and determine the motor to be controlled by the next automatic start-up process according to the following process.

[0021] (1) If the predicted start-up time of motor 1 and motor 2 is both in the abnormal range or the normal range.

[0022] a. If the equivalent running time of motor 1 - the equivalent running time of motor 2 ≥ the average single equivalent running time of motor 2, select motor 2 as the motor to be used for the next start and stop;

[0023] b. If the equivalent running time of motor 2 - the equivalent running time of motor 1 ≥ the single average equivalent running time of motor 1, select motor 1 as the motor to be used for the next start and stop;

[0024] c. If |Equivalent running time of motor 1 - equivalent running time of motor 2| < single average equivalent running time of motor 2, and equivalent running time of motor 1 - equivalent running time of motor 2| < single average equivalent running time of motor 1, select the motor with less running times as the motor to be used for the next start and stop;

[0025] (2) If one of the predicted values ​​is within the abnormal range:

[0026] a. The predicted value of motor 1 is in the abnormal range. For example, if the equivalent running time of motor 2 minus the equivalent running time of motor 1 is ≥ 2 times the single average equivalent running time of motor 1, motor 1 is determined as the motor to be used for the next start and stop; otherwise, motor 2 is selected as the motor to be used for the next start and stop.

[0027] b. The predicted value of Motor 2 is in the abnormal range. For example, if the equivalent running time of Motor 1 minus the equivalent running time of Motor 2 is ≥ 2 times the single average equivalent running time of Motor 2, Motor 2 is determined as the motor to be used for the next start and stop; otherwise, Motor 1 is selected as the motor to be used for the next start and stop.

[0028] Step 6: Determine the motors for limited start of all auxiliary systems according to steps 2 to 5.

[0029] Step 7: When the startup process reaches the stage where the technical water supply pump, high-pressure oil pump, upper guide bearing, lower guide bearing, thrust bearing and water guide bearing external circulation oil pump need to be started, if the auxiliary system has only one motor, start the motor directly. If there are two motors, start the corresponding motor according to the result of step 5.

[0030] Compared with the prior art, the present invention has the following technical effects:

[0031] 1) The start-up and shutdown process can intelligently control the reasonable operation of auxiliary motors. When automatically starting and stopping, it can automatically balance the use time of the motors and reduce safety risks.

[0032] 2) No additional detection devices are required. Decisions can be made automatically by simply counting and calculating the start time, stop time and running time of the auxiliary motor. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flow chart of the present invention. Specific implementation methods

[0035] In the following, we will present example embodiments of the present invention in the form of diagrams. However, it should be emphasized that the invention has many different feasible implementation forms and is not limited to the examples described here. The purpose of providing these examples is to fully and thoroughly disclose the present invention and to fully convey the wide range of applicability of the invention to technicians in the relevant professional fields. The terms used in the embodiments shown in the diagrams do not limit the present invention. Figure 1 The flowchart shown further describes the present invention, a startup process auxiliary decision method for a giant hydropower unit, comprising the following steps:

[0036] Step 1: construct three sequence data for the motors of the technical water supply system, the high-pressure oil system, the thrust bearing external circulation system, and the water-guided bearing external circulation system: the motor start time sequence T1, the motor stop time sequence T2, and the motor running time sequence T3, respectively expressed as T1=(t 11 ,t 21 ,t 31 ,···,t 1m1 ), T2=(t 21 ,t 22 ,t 23 ,···,t 2m2 ), T3=(t 31 ,t 32 ,t 33 ,···,t 3m3 ), count and clean up the data of normal startup, and do not count the operation records of shutdown due to faults. It can be seen that: m1=m2=m3=m.

[0038] Step 2: Calculate the equivalent running time T of the motor eq and single average equivalent running time T eq / , and the equivalent running time series T4.

[0039] T1, T2, T 23 Each time an update is made, T is calculated according to the following formula: eq , T eq / , and update T4 at the same time.

[0040]

[0041] T4=(T eq1 ,T eq2 ,T eq3 ,···,T eqm )

[0042] Step 3: Use T1, T2 and T4 to make sample X. Assume that the mth data in T1, T2 and T4 are t 1m , t 2m and t 3m , let X m =[t 1m , t 2m , t 3m ] is used as the input variable, and the support vector machine SVM is used to predict the next start time t of the motor. 1(m+1) , in this example, the RBF kernel function is used for calculation.

[0043] Step 4: Find the maximum value T of T1 1max and minimum value T 1min, median T 1med =(T 1max +T 1min ) / 2, based on the median, according to the 80 / 20 principle, the interval 1(T 1med -4(T 1med -T 1min ) / 5,T 1med +4(T 1max -T 1med ) / 5) as the normal interval, interval 2 (T 1min ,T 1med -4(T 1med -T 1min ) / 5) and the interval 3(4(T 1max -T 1med ) / 5, T 1max ) is the abnormal interval. As predicted in step 3, t 1(m+1) If it is in interval 1, it is judged as a normal value; if it is in interval 2 and interval 3, it is judged as an abnormal value.

[0044] Step 5: Assuming that the motors of the auxiliary system are represented by motor 1 and motor 2, determine the motor to be controlled by the next automatic startup process according to the following process:

[0045] (1) If the predicted start-up time of motor 1 and motor 2 are in the same range.

[0046] a. Condition 1: If the equivalent running time of motor 1 - the equivalent running time of motor 2 ≥ the single average equivalent running time of motor 2, motor 2 is selected as the motor to be used for the next start and stop;

[0047] b. Condition 2: If the equivalent running time of motor 2 - the equivalent running time of motor 1 ≥ the single average equivalent running time of motor 1, select motor 1 as the motor to be used for the next start and stop;

[0048] c. Condition 3: If |Equivalent running time of motor 1 - equivalent running time of motor 2| < single average equivalent running time of motor 2, and equivalent running time of motor 1 - equivalent running time of motor 2| < single average equivalent running time of motor 1, the motor with fewer running times is selected as the motor to be used for the next start and stop;

[0049] (2) If one of the predicted values ​​is within the abnormal range:

[0050] a. The predicted value of motor 1 is in the abnormal range; Condition 4: If the equivalent running time of motor 2 - the equivalent running time of motor 1 ≥ 2 times the single average equivalent running time of motor 1, motor 1 is determined as the motor to be used for the next start and stop; otherwise, motor 2 is selected as the motor to be used for the next start and stop.

[0051] b. The predicted value of Motor 2 is in the abnormal range; Condition 5: If the equivalent running time of Motor 1 - the equivalent running time of Motor 2 ≥ 2 times the single average equivalent running time of Motor 2, Motor 2 is determined as the motor to be used for the next start and stop; otherwise, Motor 1 is selected as the motor to be used for the next start and stop.

[0052] Step 6: When the start / stop process reaches the stage of starting the technical water supply pump, high-pressure oil pump, thrust external circulation oil pump, and water-guide bearing external circulation oil pump, start the corresponding motor according to the result of step 5.

Claims

1. A startup process decision-making method for a giant hydropower unit, characterized in that: By predicting, analyzing and comparing the status of the motors, the motors that are prioritized for startup and shutdown are automatically selected to reduce operational risks. The following steps are included: Step 1: Construct three sequence data for the motors of the technical water supply system, high-pressure oil system, thrust bearing external circulation system, and water-guided bearing external circulation system: the motor start time sequence T1, the motor stop time sequence T2, and the motor running time sequence T3, and count and clean the sequence data: The starting time sequence of the motor is T1, T1 = (t 11 ,t 21 ,t 31 ,···,t 1m1 ), the number of data in the sequence m1 = the number of times the motor is started, the data value t1 = the start time, the start time is the time from when the motor receives the start command to when the equipment is fully put into operation; The motor stop time sequence T2, T2 = (t 21 ,t 22 ,t 23 ,···,t 2m2 ), the number of data in the sequence m2 = the number of times the motor stops, the data value t2 = the stop time, the stop time is the time from when the motor receives the stop command to when the equipment stops completely; The running time sequence of the motor is T3, T3 = (t 31 ,t 32 ,t 33 ,···,t 3m3 ), the number of data in the sequence m3 = the number of times the auxiliary motor runs, the data value t3 = the running time, the running time is the time from "full operation" to "receiving the stop command" of the equipment; Only the data of normal start and stop are counted, and the data of fault shutdown are not counted; each time the motor is started, if the motor does not shut down due to fault during this start, T1 will add a new data; each time the motor stops normally, T2 and T3 will add a new data; yes, m1 = m2 = m3 = m; Step 2: Based on step 1, calculate the equivalent running time T of the motor eq and single average equivalent running time T eq / , construct the equivalent running time series T4; Step 3: Based on step 1 and step 2, predict the next start time of the motor; When predicting the start time, T1, T2, and T4 are used as input quantities, and the next start time is used as the output quantity; Step 4: Distinguish the abnormal area and normal area in the startup time; first calculate the maximum value, minimum value and median of the startup time, and use the 80 / 20 rule to determine the normal area and abnormal area; input the result of step 3 to determine whether the predicted value is in the abnormal area; Step 5: After completing the abnormal analysis of the start-up time of all motors in the auxiliary system, determine the motor that should be started next time; let the two motors of the auxiliary system be represented as motor 1 and motor 2 respectively, and determine the motor to be controlled by the next automatic start-up process; Step 6, determine the motors for limited start of all auxiliary systems according to steps 2 to 5; Step 7. When the startup process reaches the stage where the technical water supply pump, high-pressure oil pump, upper guide bearing, lower guide bearing, thrust bearing and water guide bearing external circulation oil pump need to be started, if the auxiliary system has only one motor, start the motor directly; if there are two motors, start the corresponding motor according to the result of step 5.

2. The startup process auxiliary decision method of a giant hydropower unit according to claim 1 is characterized in that: Step 2 specifically includes the following sub-steps: (1) Equivalent running time calculation T eq Calculation: First, convert each start and stop time of the motor into running time, and then add up all the running times, as follows: t 1i ,t 2i and t 3i They are the starting time when the motor starts for the i-th time, the stopping time when it stops for the i-th time, and the running time when it runs for the i-th time; m1, m2 and m3 are the total number of starts, stops and runs respectively, and K is the start multiple; (2) Construct the motor equivalent operation time series T4. The number of data in the series m = the number of times the auxiliary motor operates, and the data value T eq =The sum of the historical equivalent running time of each run; T4=(The q1 ,The q2 ,The q3 ,···,The qm ) (3) Every time T1, T2, and T3 are updated, a new T is calculated. eq and T eq / , T4 adds a new data.

3. The startup process auxiliary decision method of a giant hydropower unit according to claim 1 is characterized in that: Step 3 uses the support vector machine machine learning method to predict the next start-up time of the motor.

4. The startup process auxiliary decision method of a giant hydropower unit according to claim 1 is characterized in that: In step 5, the motor to be controlled by the next automatic startup process is determined according to the following process: (1) If the predicted start-up time of motor 1 and motor 2 is within the abnormal range or the normal range; a. If the equivalent running time of motor 1 - the equivalent running time of motor 2 ≥ the average single equivalent running time of motor 2, select motor 2 as the motor to be used for the next start and stop; b. If the equivalent running time of motor 2 - the equivalent running time of motor 1 ≥ the single average equivalent running time of motor 1, select motor 1 as the motor to be used for the next start and stop; c. If |Equivalent running time of motor 1 - equivalent running time of motor 2| < single average equivalent running time of motor 2, and equivalent running time of motor 1 - equivalent running time of motor 2| < single average equivalent running time of motor 1, select the motor with less running times as the motor to be used for the next start and stop; (2) If one of the predicted values ​​is within the abnormal range: a. The predicted value of motor 1 is in the abnormal range. For example, if the equivalent running time of motor 2 minus the equivalent running time of motor 1 is ≥ 2 times the average equivalent running time of motor 1, motor 1 is determined as the motor to be used for the next start and stop; otherwise, motor 2 is selected as the motor to be used for the next start and stop; b. The predicted value of Motor 2 is in the abnormal range. For example, if the equivalent running time of Motor 1 minus the equivalent running time of Motor 2 is ≥ 2 times the single average equivalent running time of Motor 2, Motor 2 is determined as the motor to be used for the next start and stop; otherwise, Motor 1 is selected as the motor to be used for the next start and stop.

Images