A fault diagnosis method for centrifugal pumps of water injection stations based on identification rules

Abstract

The application discloses a kind of based on identification rule's water injection station centrifugal pump fault diagnosis method, first respectively collect the various operating parameters of centrifugal pump when normal operation and when fault occurs, find out the parameter of different change trend by comparison, then the boundary condition of centrifugal pump unit fault anomaly is calculated using the way of change rate and threshold, develop a kind of centrifugal pump fault identification rule, subsequent collection centrifugal pump real-time operating parameter, when the change trend of these parameters is matched with the identification rule designed before, system will pass the fault information of centrifugal pump to the guard personnel, and different fault finds the corresponding identification rule and can carry out real-time fault diagnosis.The application adopts the above-mentioned one based on identification rule's water injection station centrifugal pump fault diagnosis method, can detect the operating state of centrifugal pump in real time, avoid centrifugal pump to appear high load operation, pump stop situation.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and in particular to a fault diagnosis method for centrifugal pumps in water injection stations based on identification rules. Background Technology

[0002] Centrifugal water injection pumps have a complex structure and operate under high temperature and pressure for extended periods, making them prone to various malfunctions. A failure in a centrifugal water injection pump can hinder oilfield extraction. Therefore, as a core piece of equipment in an oilfield water injection station, the centrifugal water injection pump must be ensured to operate normally at all times. Real-time fault diagnosis can be performed, the pump's operating condition can be monitored continuously, and problems can be reported to staff immediately, thereby reducing losses and risks.

[0003] Currently, centrifugal pump water injection stations face reliability issues. Firstly, while the current oilfield information system has collected a large amount of data, it hasn't established a correlation or quantitative relationship with problems discovered through manual inspections, such as fluid leaks, abnormal noises, and excessive vibration. Manual analysis is still required, involving various methods such as observing the units on-site, smelling for unusual odors, listening to operating sounds, and feeling the vibration patterns of the units to comprehensively assess potential equipment problems. Secondly, with the implementation of information technology in water injection stations, they should have the capability for multi-dimensional problem analysis. When equipment malfunctions, they will be reflected in multiple production and operating parameters, and these data have logical relationships. However, the current system lacks a digitally supported fault diagnosis model, preventing online real-time diagnosis.

[0004] Currently, most centrifugal pump fault diagnosis methods rely on vibration and displacement signals. However, some oilfield water injection stations lack vibration and displacement sensors, making it impossible to collect these signals. Therefore, the aforementioned methods have limitations when applied to such water injection stations. In light of this situation, a fault diagnosis method for centrifugal pumps in water injection stations based on identification rules is proposed. This method can achieve fault diagnosis of centrifugal pump units without collecting abstract vibration and displacement signals. Summary of the Invention

[0005] The purpose of this invention is to provide a fault diagnosis method for centrifugal pumps in water injection stations based on identification rules, which can detect the operating status of centrifugal pumps in real time and avoid high-load operation or pump shutdown.

[0006] To achieve the above objectives, this invention provides a method for diagnosing centrifugal pump faults in water injection stations based on identification rules, comprising the following steps (taking a centrifugal pump shaft failure as an example):

[0007] Step S1: Collect various operating parameters of the centrifugal pump under normal operation and under conditions of shaft failure;

[0008] Step S2: Based on Step S1, identify the parameters with different trends under the two states of normal operation and fault operation of the centrifugal pump, namely: pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current and C-phase current;

[0009] Step S3: Record the data of the pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current and C-phase current from the start of operation to the pump stop when the centrifugal pump experiences a shaft failure.

[0010] Step S4: Based on the data obtained in Step S3, calculate the threshold values ​​for the rate of change of the pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current, and C-phase current using the following formulas:

[0011]

[0012] Where n is the total number of parameters obtained by the system, T i T represents the i-th data item. i-1 This represents the data from the previous time step (i-th time step).

[0013] Step S5: Customize the recognition rules based on the parameter change trend obtained in step S2 and the parameter threshold obtained in step S4;

[0014] Step S6: Match the changing trends of the pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current, and C-phase current with the identification rules customized in step S5.

[0015] Step S7: Determine whether the data of pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current and C-phase current during the operation of the centrifugal pump match the corresponding identification rules. If they match, the corresponding fault is diagnosed; otherwise, it means that the equipment is operating normally.

[0016] Step S8: Each fault has a different identification rule. Based on the different identification rules for each fault, the corresponding fault is diagnosed.

[0017] Preferably, for different parameters, the pump stop value is different, and the number of data entries n that the system can obtain is also different.

[0018] Preferably, the real-time front shaft temperature, rear shaft temperature, A-phase current, B-phase current, and C-phase current of the centrifugal pump are collected, and the rate of change and threshold are used to determine the trend of parameter changes.

[0019] Preferably, when the changing trends of each parameter match the designed identification rules, the system will transmit the information about the centrifugal pump experiencing a shaft failure to the on-duty personnel.

[0020] Preferably, the efficiency of the water injection pump unit and the three-phase voltage imbalance are also reference standards for diagnosing equipment faults. The efficiency of the water injection pump unit can reflect the energy conversion capability of the water injection pump unit, and at the same time, it can also reflect the operating status of the water injection pump unit.

[0021] Preferably, the motor input power is:

[0022]

[0023] Where N s denoted as , where I is the input power of the motor (kW); is the line current of the motor (A); is the line voltage of the motor (kV); and is the power factor of the motor.

[0024] Preferably, the input power of the water injection pump unit is:

[0025] N t =N s +P Pin ·G P / 3.6

[0026] Where N t The input power of the water injection pump unit is kW; PPin is the inlet pressure of the water injection pump, MPa; G P The flow rate of the water injection pump is in meters. 3 / h.

[0027] The preferred formula for calculating the output power of the water injection pump is as follows:

[0028] N u =P Pout ·δ P / 3.6

[0029] Where N u The output power of the water injection pump is expressed in kW; P. Pout The pressure is the outlet pressure of the water injection pump, in MPa.

[0030] The preferred formula for calculating the efficiency of the water injection pump unit is as follows:

[0031]

[0032] Where η mp The efficiency of the water injection pump unit is %.

[0033] Preferably, when a voltage imbalance occurs in a three-phase power supply system, the IEEE Std 112-2004 method is used to calculate the three-phase voltage imbalance using the phase voltages, as shown in the following formula:

[0034]

[0035] U A UB U C U is the three-phase voltage of the centrifugal pump motor. avg This represents the average value of the three-phase voltage.

[0036] Therefore, the present invention employs the above-mentioned rule-based fault diagnosis method for centrifugal pumps in water injection stations, which has the following advantages compared with the prior art:

[0037] (1) It is not necessary to collect complex and abstract vibration signals or displacement signals. Only common data such as temperature, pressure, liquid level, flow rate, and current are needed to realize real-time fault diagnosis of centrifugal pumps.

[0038] (2) It has very high scalability and can be applied to different mechanical equipment, different faults and different working conditions.

[0039] (3) Regarding the problem of fault diagnosis of centrifugal pumps, the accuracy reaches more than 99%, which meets the standard for normal production and operation of centrifugal pumps.

[0040] (4) It has very high sensitivity and takes very little time to diagnose faults. After multiple experiments, it can diagnose the corresponding faults in an average of 15 minutes, which can meet the needs of practical applications.

[0041] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating the overall process of a fault diagnosis method for centrifugal pumps in water injection stations based on identification rules, according to the present invention.

[0043] Figure 2 This is a graph showing the temperature change of the pump shaft.

[0044] Figure 3 This is an unmanned system for the water injection station;

[0045] Figure 4 This is a diagram from a centrifugal pump tandem shaft prediction experiment. Detailed Implementation

[0046] Example

[0047] When a centrifugal pump unit malfunctions, it often exhibits excessive vibration, overheating, and low efficiency. This can also affect connected sensors, causing abnormal data. Therefore, by analyzing the changes in relevant parameters under both normal and faulty operating conditions, and examining their characteristics, we can make inverse judgments. When these characteristics appear in the data, we can deduce the specific fault that has occurred.

[0048] Taking a centrifugal pump shaft misalignment fault as an example, firstly, parameters of the centrifugal pump are collected during normal operation and when a shaft misalignment fault occurs, and parameters with different trends of change are identified. From Figure 2 As can be seen, the pump inlet shaft temperature is stable at around 44℃ during normal operation of the centrifugal pump. However, when the centrifugal pump experiences shaft misalignment, the pump inlet shaft temperature shows an upward trend. Simultaneously, the trends of the pump outlet shaft temperature, A-phase current, B-phase current, and C-phase current are all observed to be significantly different from those under normal conditions.

[0049] However, observing only one trend is insufficient. In the algorithm of this study, the rate of change and a threshold are used as key parameters for judging the trend of data change. Since 80℃ is the shutdown value of the pump inlet shaft temperature in the water injection station management system, the centrifugal pump will automatically shut down after reaching this temperature. From the start of operation to the shutdown of the centrifugal pump that experienced a shaft failure, the system acquired a total of 54 data points. Therefore, these data were used to calculate the threshold of the pump inlet shaft temperature. The formula for calculating the threshold is as follows:

[0050]

[0051] In equation (1), n ​​is the total number of parameters obtained by the system. For different parameters, their pump stop values ​​are different, and the number of data points n that the system can obtain is also different. i T represents the temperature of the i-th pump shaft. i-1 This represents the pump shaft temperature at the previous moment before the i-th time.

[0052] Similarly, the changing trends and thresholds of the pump shaft temperature, A-phase current, B-phase current, and C-phase current are calculated, and their corresponding pump shutdown values ​​are recorded, as shown in Table 1.

[0053] Table 1 shows the identification rules for centrifugal pump shaft misalignment faults. Subsequently, real-time data on the pump's inlet and outlet shaft temperatures, A-phase current, B-phase current, and C-phase current can be collected. The rate of change and threshold values ​​are used to determine the trend of these parameters. When the trend of these parameters matches the previously designed identification rules, the system will transmit the centrifugal pump shaft misalignment fault information to the on-duty personnel. Each fault corresponds to a different identification rule; therefore, by using the same processing method to find the corresponding identification rule, real-time fault diagnosis of the centrifugal pump can be performed.

[0054]

[0055] Table 1

[0056] When a water injection station is running, certain parameters are inherently correlated. By analyzing a large amount of field data, the patterns involved in these correlations can be identified. For example, new characteristic values ​​can be calculated from existing parameters, and the fluctuation range of these characteristic values ​​during normal operation of the centrifugal pump can be observed. When these characteristic values ​​exceed the corresponding range, it can be considered that the parameters associated with these characteristic values ​​have become abnormal. Moreover, these abnormalities often reflect the current operating condition of the centrifugal pump equipment. Therefore, these new characteristic values ​​can be used as reference standards for equipment fault diagnosis.

[0057] (1) Efficiency of water injection pump unit

[0058] The efficiency of a water injection pump unit is calculated by dividing the output power by the input power. The efficiency reflects the unit's energy conversion capability and its operating status. Under normal operation, the efficiency of a water injection pump unit of the same model remains within a relatively fixed range. When a fault occurs, the efficiency typically drops significantly. Although the efficiency may occasionally drop to very low values, the pump unit can still operate normally. This is generally due to equipment aging, but in this case, the low energy conversion efficiency leads to significant energy waste. Therefore, when the efficiency drops to a certain level, the water injection pump unit can be considered to have malfunctioned and requires timely maintenance. For water injection pump units where the fault cannot be clearly identified but which can still operate, repair or replacement of certain components should be considered to avoid substantial energy waste. Therefore, the efficiency of a water injection pump unit should be used as an indicator for condition monitoring and fault diagnosis.

[0059] Table 2 shows the specific water injection pump units and their corresponding operating efficiency indicators, where Q represents the rated flow rate of the water injection pump, in cubic meters per hour (m³ / h). 3 / h).

[0060] Motor input power:

[0061]

[0062] In equation (2), N s denoted as , where I is the input power of the motor (kW); is the line current of the motor (A); is the line voltage of the motor (kV); and is the power factor of the motor.

[0063] Input power of water injection pump unit:

[0064] N t =N s +P Pin ·GP / 3.6 (3)

[0065] In equation (3), N tThe input power of the water injection pump unit is kW; PPin is the inlet pressure of the water injection pump, MPa; G P The flow rate of the water injection pump is in meters. 3 / h.

[0066] Formula for calculating the output power of a water injection pump:

[0067] N u =P Pout ·G P / 3.6 (4)

[0068] In equation (4), N u The output power of the water injection pump is expressed in kW; P. Pout The pressure is the outlet pressure of the water injection pump, in MPa.

[0069] Formula for calculating the efficiency of water injection pump unit:

[0070]

[0071] In equation (5), η mp The efficiency of the water injection pump unit is %.

[0072]

[0073]

[0074] Table 2

[0075] (2) Three-phase voltage imbalance

[0076] When voltage imbalance occurs in a three-phase power supply system, the voltage imbalance degree is typically used as a technical indicator for evaluation. Over the past few years, numerous methods for calculating the degree of voltage imbalance have been proposed in domestic and international research. NEMA, in its standard, first indicated that the three-phase line voltage can be used as the main parameter to obtain the voltage imbalance degree; IEEE pointed out that the phase voltage can be calculated to obtain this indicator value; IEC, in response to this issue, proposed using a precise calculation formula in a three-phase three-wire system to derive the voltage imbalance degree, a method that has since been widely adopted in Chinese national standards. Considering the characteristics, accuracy, and convenience of centrifugal pumps, the method of IEEE Std 112-2004 is ultimately adopted. This method primarily uses phase voltage to determine the three-phase voltage imbalance degree.

[0077]

[0078] In equation (6), U A U B U C U is the three-phase voltage of the centrifugal pump motor. avgThe average value of the three-phase voltage is given by equation (6). As can be seen from equation (6), this method only requires measuring the values ​​of the three-phase voltage to calculate the current voltage imbalance.

[0079] Numerous temperature, pressure, current, and flow sensors are installed at key locations on the centrifugal pump. These sensors transmit their current values ​​to the PLC cabinet every 50ms. The PLC then processes the data and forwards it to the server, where it is displayed in the water injection station's management system. This data transmission interval is one minute. Simultaneously, this data is also transmitted to the unattended water injection station system. This system stores the latest 15 data points in a queue, then uses thresholds and change rates to determine the trend of each parameter within these 15 data points. Finally, it matches these trends with previously designed identification rules to arrive at the final result.

[0080] Figure 3 The unmanned water injection station system displays various parameters in real-time as line graphs, and prints out data filtering information and centrifugal pump fault information below. Due to sensor aging, loose wiring, and data transmission fluctuations, a small portion of the data transmitted to the system may be inaccurate. The system first filters the data transmitted from the PLC, and then passes the filtered data to the fault diagnosis module for fault diagnosis. When the system diagnoses a fault, it prints the fault information in the operation log with a red background.

[0081] Figure 4 Centrifugal pump shaft failure prediction experiment. As shown in the figure, both the inlet and outlet shaft temperatures of the pump exhibit an increasing trend. The inlet shaft temperature increased from 44.43℃ to 60.51℃, and the outlet shaft temperature increased from 50.17℃ to 78.86℃. The A-phase, B-phase, and C-phase currents show an increasing and then decreasing trend. Specifically, the A-phase current increased from 124.8A to 139.1A and then decreased to 134.1A; the B-phase current increased from 125.2A to 139.7A and then decreased to 134.4A; and the C-phase current increased from 125.1A to 140.1A and then decreased to 133.9A. Simultaneously, the centrifugal pump shaft failure was successfully diagnosed by the centrifugal pump shaft failure model, and the fault information was printed in red on the operation log.

[0082] Therefore, the present invention employs the above-mentioned method for diagnosing centrifugal pump faults in water injection stations based on identification rules, which can detect the operating status of centrifugal pumps in real time and avoid high-load operation or pump shutdown.

[0083] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for fault diagnosis of centrifugal pumps in water injection stations based on recognition rules, characterized in that, Includes the following steps: Step S1: Collect various operating parameters of the centrifugal pump under normal operation and under conditions of shaft failure; Step S2: Based on Step S1, identify the parameters with different trends under the two states of normal operation and fault operation of the centrifugal pump, namely: pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current and C-phase current; Step S3: Record the data of the pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current and C-phase current from the start of operation to the pump stop when the centrifugal pump experiences a shaft failure. Step S4: Based on the data obtained in Step S3, calculate the threshold values ​​for the rate of change of the pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current, and C-phase current using the following formulas: Where n is the total number of parameters obtained by the system, T i T represents the i-th data item. i-1 This represents the data from the previous time step (i-th time step). Step S5: Customize the recognition rules based on the parameter change trend obtained in step S2 and the parameter threshold obtained in step S4; The identification rules are as follows: The changing trends and thresholds of the pump shaft temperature, A-phase current, B-phase current, and C-phase current are calculated, and the corresponding pump stop values ​​are also recorded. The changing trends of the parameters are judged by the rate of change and the threshold. Each fault corresponds to a different identification rule. Step S6: Match the changing trends of the pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current, and C-phase current with the identification rules customized in step S5. Step S7: Determine whether the data of pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current and C-phase current during the operation of the centrifugal pump match the corresponding identification rules. Step S8: Diagnose the corresponding fault according to the identification rules for each fault; The corresponding pump stop value and number of data entries are obtained by changing the parameters; The system collects real-time data on the pump front shaft temperature, pump rear shaft temperature, A-phase current, B-phase current, and C-phase current of the centrifugal pump, and uses the rate of change and threshold values ​​to determine the trend of parameter changes. When the changing trends of each parameter match the designed identification rules, the system will transmit the information about the centrifugal pump experiencing a shaft failure to the on-duty personnel.

2. The method for fault diagnosis of centrifugal pumps in water injection stations based on recognition rules according to claim 1, characterized in that: The characteristic values ​​for diagnosing equipment faults also include the efficiency of the water injection pump unit and the three-phase voltage imbalance.

3. The method for fault diagnosis of centrifugal pumps in water injection stations based on recognition rules according to claim 2, characterized in that, Input power of water injection pump motor: Where N s denoted as , where I is the input power of the motor (kW); is the line current of the motor (A); is the line voltage of the motor (kV); and is the power factor of the motor.

4. The method for fault diagnosis of centrifugal pumps in water injection stations based on recognition rules according to claim 2, characterized in that, Input power of water injection pump unit: N t =N s +P Pin ·G P / 3.6 Where N t The input power of the water injection pump unit is kW; P Pin The inlet pressure of the water injection pump is in MPa; G P The flow rate of the water injection pump is in meters. 3 / h.

5. The method for fault diagnosis of centrifugal pumps in water injection stations based on recognition rules according to claim 2, characterized in that, Formula for calculating the output power of a water injection pump: N u =P Pout ·G P / 3.6 Where N u The output power of the water injection pump is expressed in kW; P. Pout The pressure is the outlet pressure of the water injection pump, in MPa.

6. The method for fault diagnosis of centrifugal pumps in water injection stations based on recognition rules according to claim 2, characterized in that, Formula for calculating the efficiency of water injection pump unit: Where η mp The efficiency of the water injection pump unit is %.

7. The method for fault diagnosis of centrifugal pumps in water injection stations based on recognition rules according to claim 2, characterized in that: When a voltage imbalance occurs in a three-phase power supply system, the IEEE Std 112-2004 method is used to calculate the three-phase voltage imbalance using the phase voltages. The formula is as follows: UA, UB, and UC are the three-phase voltages of the centrifugal pump motor, and Uavg is the average value of the three-phase voltages.

Images