System and method for setting condition monitoring alarm for turbine in pumped storage power plant

Abstract

The present invention relates to a process for setting a condition monitoring alarm for a turbine in a pumped storage power plant, comprising: an operating condition determination unit that collects an operating data set and determines an operating condition on the basis of operating parameters including power output and flow rate; a signal characteristic extraction unit that analyzes vibration data measured for each operating condition to identify order components and extract signal characteristics of each operating condition; a data distribution analysis unit that performs probabilistic analysis of a distribution of vibration values on the basis of the extracted signal characteristic data, and sets an appropriate alarm threshold for each operating condition through statistical evaluation; and an alarm management unit that monitors abnormal conditions of the equipment in real time on the basis of the set alarm thresholds and the operating condition data, wherein by analyzing vibration data for each operating condition and measurement location to monitor abnormal conditions of the equipment in real time, and by setting alarm thresholds optimized for each condition, precision and reliability of equipment condition monitoring are enhanced.

Description

Condition monitoring alarm setting system for pumped storage power plant turbines and method thereof

[0001] The present invention relates to a process for setting a condition monitoring alarm for a pumped-storage power plant turbine, and more specifically, to a system and method for setting a condition monitoring alarm for a pumped-storage power plant turbine that analyzes operating conditions and vibration data by measurement location to monitor abnormal conditions of equipment in real time and set alarm thresholds.

[0002] Pumped storage power plants operate an online vibration measurement system that monitors changes in the amplitude of vibration responses to evaluate the soundness of turbine equipment.

[0003] Since water turbine equipment often experiences large output fluctuations and is operated under various operating conditions rather than at the optimal operating point, its vibration response characteristics vary depending on the operating state.

[0004] However, the vibration signal alarm levels currently applied in pumped-storage power plants are set to a constant value regardless of the operating state, making it ineffective for determining equipment abnormalities based on measured signals.

[0005] First, changes in vibration characteristics due to changes in driving are not being recognized.

[0006] Second, it fails to reflect vibration deviations depending on the measurement location.

[0007] Figure 1 illustrates an example of changes in vibration characteristics according to operating conditions, showing that the vibration response appears differently under specific operating conditions. In Figure 1, changes in vibration characteristics are divided into a normal section and a transient section under specific operating conditions, and it can be seen that changes in amplitude are prominent in the transient section. In Figure 1, the x-axis represents the time at which the equipment's vibration data was collected, showing how vibration characteristics change over time. The y-axis represents the frequency of the vibration data, visualizing the frequency or order component of the vibration signal. The z-axis represents the amplitude or vibration intensity, expressing the magnitude of the vibration measured at a specific frequency.

[0008] While the normal range exhibits a stable vibration pattern, the transient range is characterized by increased amplitude or variability due to rapid changes in output. This suggests that changes in operating conditions have a direct impact on vibration characteristics and highlights the problem that existing fixed alarm setting methods fail to reflect these changes.

[0009] [Prior Art Literature]

[0010] [Patent Literature]

[0011] (Patent Document 1) Republic of Korea Patent Publication 2014-0067429 (June 5, 2014)

[0012] To solve the aforementioned problem, the present invention aims to generate a customized alarm model based on data characteristics by operating state and measurement location, thereby monitoring the equipment status in real time and increasing accuracy and efficiency.

[0013] According to the present invention, the apparatus comprises: an operating state determination unit that collects an operating data set and determines the operating state based on operating parameters including output and flow rate; a signal characteristic extraction unit that analyzes vibration data measured for each operating state to identify order components and extracts signal characteristics for each operating state; a data distribution analysis unit that probabilistically analyzes the distribution of vibration values ​​based on the extracted signal characteristic data and sets an appropriate alarm value for each operating state through statistical evaluation; and an alarm management unit that monitors the abnormal state of the equipment in real time based on the set alarm value and operating state data.

[0014] The operating state determination unit determines the normal section and the transient section based on the output change rate, the flow rate change rate, and the guide vane opening, and uses the minimum value of the flow rate change and the threshold value of the output change rate as criteria for determining the transient section.

[0015] The above signal characteristic extraction unit extracts PP values ​​or RMS values ​​according to the Order component through frequency analysis of vibration data measured for each operating state, identifies specific vibration patterns based on the range of amplitude and RMS values ​​of the 1X component in the normal section and in the transient section, and defines characteristic indicators including frequency, amplitude, range and variability of RMS values ​​for each section based on this.

[0016] The above data distribution analysis unit assumes the extracted signal characteristic data as a probability distribution and sets the alarm value by multiplying the median by a statistically calculated coefficient.

[0017] A method using a system for setting a condition monitoring alarm for a pumped storage power plant turbine comprises: (a) a step in which the system collects a set of operating data including output and flow rate and determines the operating state into a normal section and a transient section based thereon; (b) a step in which the system analyzes vibration data measured for each determined operating state to identify an Order component and extracts signal characteristic data corresponding to each operating state; (c) a step in which the system probabilistically analyzes the distribution of vibration values ​​based on the extracted signal characteristic data and sets an appropriate alarm value for each operating state through statistical evaluation; and (d) a step in which the system monitors the abnormal condition of the equipment in real time based on the set alarm value and operating state data.

[0018] According to the present invention, abnormal conditions of equipment are monitored in real time by analyzing vibration data by operating status and measurement location, and the precision and reliability of equipment condition monitoring are improved by setting alarm thresholds optimized for each condition. This allows for a clear distinction between normal and abnormal conditions of the equipment, enables early detection of abnormal conditions for rapid response, reduces unnecessary alarms, and provides the effect of increasing maintenance efficiency.

[0019] Figure 1 shows an example of a change in vibration characteristics according to the operating condition.

[0020] FIG. 2 is a configuration diagram of a state monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0021] Figure 3 shows the turbine equipment alarm setting process of a state monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0022] Figure 4 shows the classification of operating sections of a condition monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0023] FIG. 5 visually illustrates how various driving sections (P Mode, G Mode, transient state) are distinguished through driving state determination according to one embodiment of the present invention.

[0024] FIG. 6 visually illustrates the process of extracting signal characteristics in the normal section of a state monitoring alarm setting system for a pumped-storage power plant turbine according to one embodiment of the present invention, and the structure of monitoring the state of the equipment based thereon.

[0025] FIG. 7 visually illustrates the process of extracting signal characteristics in a transient section of a state monitoring alarm setting system for a pumped-storage power plant turbine according to one embodiment of the present invention, and the structure of monitoring the state of the equipment based thereon.

[0026] FIG. 8 is intended to explain the data distribution fitting and alarm value setting of a state monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0027] FIG. 9 is a flowchart of a method using a state monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0028] The present invention is a process for determining the operating status of a water turbine by combining operating signals to distinguish similar operating conditions and performing comparative monitoring based on measurement signals for each operating status. In this case, the parameters for comparative monitoring are based on characteristics extracted by analyzing the characteristics of signals appearing for each operating status. Alarm settings are established through statistical and probabilistic analysis based on measurement data, and a model having different alarm thresholds depending on the operating status and measurement location can be generated.

[0029] Through this, the present invention effectively reflects changes in operating conditions and vibration deviations according to measurement locations, which were limitations of existing systems, thereby enabling more precise monitoring of equipment conditions.

[0030] FIG. 2 is a configuration diagram of a state monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0031] As illustrated in FIG. 2, the state monitoring alarm setting system (10) of a pumped storage power plant turbine includes an operating state determination unit (100), a signal characteristic extraction unit (200), a data distribution analysis unit (300), and an alarm management unit (400).

[0032] Figure 3 shows the turbine equipment alarm setting process of a state monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0033] The turbine equipment alarm setting process of the pumped-storage power plant turbine condition monitoring alarm setting system is as follows. First, an operating data set is collected, and based on this, the operating status is determined and classified into normal and transient sections. Subsequently, vibration data measured according to each operating state is analyzed to extract PP or RMS values ​​based on the order component, thereby deriving signal characteristics for each operating state. Based on the derived signal characteristic data, the data distribution is probabilistically analyzed, and an appropriate alarm value is set through statistical evaluation. Finally, the process consists of a step in which the set alarm value is utilized to detect abnormal conditions of the equipment and output an alarm.

[0034] The operating state determination unit (100) collects an operating data set of a pumped storage power plant turbine and determines the operating state based on operating parameters including output and flow rate.

[0035] This operating state determination unit determines the normal and transient sections based on rotational speed, output change rate, flow rate change rate, and guide vane opening, and uses the minimum value of the flow rate change and the threshold value of the output change rate as criteria for determining the transient section.

[0036] For example, as a criterion for determining the transient section, a state where the pump / turbine is not at its rated speed (e.g., 0 to 400 rpm) and a threshold value is set when the rate of change in flow rate exceeds 2% or the rate of change in output exceeds 5%, and the threshold value can be adjusted based on the statistical analysis of the operating data. Here, although the threshold values ​​for the rate of change in flow rate and the rate of change in output are specified as specific numerical values ​​(e.g., 2%, 5%), the threshold value can be adjusted based on the statistical analysis of the operating data.

[0037] Figure 4 shows the classification of operating sections of a condition monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0038] The operating state determination according to the present embodiment determines the operating state by combining output and flow (flow rate or guide vane opening) operating parameters. The distinction between pumping mode and power generation mode is made using an output factor, and the distinction between the transient section and the normal section is made using rotational speed and flow rate factors. The transient section includes the point in time when the rotational speed changes according to the start / stop and output fluctuation mechanism, and is determined by selecting a threshold for flow rate change. At this time, the threshold is selected by considering the minimum value of flow rate change in the transient section and the maximum value of flow rate change in the normal section.

[0039] As illustrated in Fig. 4, the process of classifying the operating state of a pumped-storage power plant turbine into a steady-state and a transient state is diagrammed. The operating state is classified into pumped-storage mode or power generation mode (P-Mode / G-Mode), and each state is subdivided into a steady-state and a transient state.

[0040] The steady-state is a period in which the equipment operates in a stable state, including the normal operating range. On the other hand, the transient is a period in which rapid changes in the state of the equipment occur, including start-up / shutdown and partial load / overload states. Figure 4 illustrates a structure that systematically classifies the operating state of a pumped-storage power plant turbine to effectively monitor and analyze the stability and abnormal conditions of the equipment.

[0041] FIG. 5 visually illustrates how various driving sections (P Mode, G Mode, transient state) are distinguished through driving state determination according to one embodiment of the present invention.

[0042] Based on the time axis, it clearly shows what operating state the equipment was in during a specific period, and each section is separated according to the result of determining the operating state.

[0043] The normal range (P Mode, G Mode) represents a state with a relatively stable output pattern and little fluctuation in flow rate, while the transient state shows an unstable pattern due to sudden changes in output or flow rate. The transient state is a range that exceeds the threshold, suggesting the possibility of equipment malfunction or a change in state.

[0044] In addition, key operating data such as output and flow rate are visualized to enable clear analysis of changes occurring in specific sections (e.g., output fluctuations, sudden or decreased flow rates). This demonstrates that determining the operating status goes beyond mere data processing, allowing for intuitive verification and real-time analysis of the equipment status.

[0045] The black alarm indicator displayed at a specific point suggests that operating section identification and alarm setting are linked, enabling real-time monitoring and response to abnormal equipment conditions. In conclusion, this image demonstrates that the operating sections of the equipment can be clearly distinguished through operating status identification, and based on this, equipment status monitoring and abnormality alarm setting can be effectively performed.

[0046] The signal characteristic extraction unit (200) analyzes vibration data measured for each operating state to identify order components and extracts signal characteristics for each operating state. This signal characteristic extraction unit extracts PP values ​​or RMS values ​​according to the order components through frequency analysis of the vibration data measured for each operating state.

[0047] For reference, the PP (Peak-to-Peak Value) refers to the difference between the maximum and minimum values ​​of a signal.

[0048] The further operating conditions deviate from rated conditions, the more disturbed the flow within the equipment becomes, potentially causing vortices or water separation and resulting in fluid-induced vibration. In other words, in these sections, vibration characteristics caused by non-uniform fluid flow appear as the dominant frequency, rather than vibration components arising from mechanical causes (frequencies corresponding to synchronous velocity and harmonic frequencies). Therefore, specific vibration patterns are identified for each section, and based on this, characteristic indicators are defined that include the range and variability of frequency, amplitude, and RMS values.

[0049] For reference, a specific vibration pattern refers to the dominant form of vibration that includes characteristics such as the inherent frequency, amplitude, RMS value, and variability of the vibration signal appearing under given operating conditions. This is used to analyze the causes of vibrations occurring in machinery or equipment or to diagnose their condition.

[0050] As operating conditions deviate from rated conditions, fluid flow disturbances become more severe, potentially leading to abnormal fluid flow phenomena such as vortices or flow separation. The resulting vibrations manifest as fluid-induced vibrations; in such cases, the dominant frequency generated by the fluid flow becomes more prominent in the vibration signal than the mechanical causes (vibrations corresponding to synchronous velocity and harmonic frequencies).

[0051] A specific vibration pattern is defined as the dominant signal characteristics that appear according to the operating state, and includes frequency characteristics (dominant frequency and the harmonic components of the corresponding signal), amplitude characteristics (PP values ​​and RMS values), variability (change or irregularity of the signal over time), and pattern periodicity (signal characteristics that appear repeatedly under specific conditions).

[0052] For example, under normal operating conditions, oscillations of synchronous speed (operating speed) and harmonic frequency due to mechanical causes appear as the main patterns. On the other hand, under disturbed operating conditions, low-frequency components caused by abnormal flow (vortex, separation, etc.) appear as the main patterns, and natural frequencies and nonlinear components (asynchronous frequencies) become prominent.

[0053] In conclusion, specific vibration patterns refer to the unique characteristics of vibration signals that vary depending on operating conditions and the environment, and changes in dominant frequency, amplitude, and variability are key factors determining these patterns. This can be utilized to diagnose the condition of equipment or detect abnormalities.

[0054] In this embodiment, the reason for defining characteristic indicators is to clearly distinguish and standardize the characteristics of vibration data measured by operating state, thereby accurately diagnosing the normal and abnormal operating states of the equipment. Generally, the power transmitted from the drive unit possesses the greatest physical force, and since the physical force decreases as the frequency band moves away from this operating component, changes in the region of minute defects (frequency components with smaller amplitudes) have almost no effect on the total vibration value. Therefore, in order to monitor the progression of defects or risks in advance, it is necessary to monitor changes in vibration values ​​by frequency band based on order components extracted from the vibration data and to manage trends.

[0055] By systematically defining the vibration characteristics occurring in normal and transient sections in terms of frequency, amplitude, and the range and variability of RMS values, differences according to each operating state can be clearly distinguished. This increases the precision of equipment condition monitoring and alarm setting, and enables the provision of alarms appropriate for actual abnormal conditions.

[0056] Defining these characteristic indicators increases efficiency by enabling the concise comparison and judgment of complex data when monitoring and analyzing equipment status in real time. Furthermore, these indicators enhance equipment stability and reliability by detecting abnormal conditions early, prevent unexpected accidents, and reduce maintenance costs. In the long term, these characteristic indicators can support data-driven decision-making, such as analyzing equipment performance, optimizing operating conditions, and setting maintenance cycles.

[0057] In the signal characteristic extraction of the present embodiment, since signal characteristics differ depending on the operating state, a process of identifying components based on the order components of vibration data by section is performed, and the identified components are defined as section characteristics and extracted.

[0058] For reference, the Order component refers to specific frequency components related to the rotational speed of the equipment; 1X represents the fundamental rotational frequency, while 2X, 9X, etc., represent its harmonics. Identifying this is useful for monitoring the progression of defects or potential risks in advance. In normal operation, fundamental frequency components such as 1X and multiple components are predominant, whereas in transient operation, fractional harmonic components below 1X and fluid-induced vibration components may appear. Therefore, by analyzing the Order component, the condition of the equipment can be precisely monitored, and abnormal conditions can be detected early.

[0059] FIG. 6 visually illustrates the process of extracting signal characteristics in the normal section of a state monitoring alarm setting system for a pumped-storage power plant turbine according to one embodiment of the present invention, and the structure of monitoring the state of the equipment based thereon.

[0060] Figure 6 illustrates the extraction of signal characteristics in the steady-state. In the steady-state, order components (1X, 2X, 9X, 20X, etc.) are identified based on relative shaft vibration (S_pp) and bearing housing vibration (V_rms), and vibration characteristics such as RSI, vortex, and cavitation, which indicate specific abnormal conditions, are analyzed and connected to alarm C. This is intended to accurately detect whether there is an abnormality in the equipment by analyzing detailed vibration patterns even in the steady-state.

[0061] FIG. 7 visually illustrates the process of extracting signal characteristics in a transient section of a state monitoring alarm setting system for a pumped-storage power plant turbine according to one embodiment of the present invention, and the structure of monitoring the state of the equipment based thereon.

[0062] This section explains the extraction of signal characteristics during the transient phase. During the transient phase, abnormal vibration characteristics such as asynchronous vibration energy of less than 1X, surge vortex rope, and vortex turbulence are primarily analyzed. Additionally, the state is subdivided into FL, HPL, PL, and DPL and connected to Alarm D. This is intended to effectively detect unstable vibration characteristics occurring during the transient phase and to precisely analyze changes in the equipment's state.

[0063] For reference, FL, HPL, PL, and DPL are abbreviations used to classify transient intervals based on load levels or output, each representing vibration characteristics under different load conditions. FL signifies when the equipment operates under a full load condition, while HPL indicates a high partial load corresponding to 0.6 to 1 times the rated load. PL indicates a partial load condition between 0.3 and 0.6 times the rated load, and DPL signifies a deep partial load condition of less than 0.3 times the rated load, which is a load state much lower than the load for which the turbine was designed.

[0064] These abbreviations are used to analyze various vibration characteristics according to load conditions in detail during transient periods. This allows for the detection of complex vibration patterns in equipment, effective analysis of the causes of abnormal vibrations, and early diagnosis of abnormal conditions in equipment.

[0065] Figures 6 and 7 highlight that signal characteristics differ between normal and transient sections, suggesting that the equipment condition can be precisely monitored by analyzing vibration components suitable for each section and defining characteristics based on this analysis. This enables customized alarm settings for each operating state and early detection of equipment abnormalities.

[0066] The data distribution analysis unit (300) probabilistically analyzes the distribution of vibration values ​​based on the extracted signal characteristic data and sets an appropriate alarm value for each operating state through statistical evaluation.

[0067] This data distribution analysis unit assumes the extracted signal characteristic data as a probability distribution and sets the alarm value by multiplying the median by statistically calculated coefficients (1.6 and 2.5).

[0068] The reason for setting alarm values ​​by assuming extracted signal characteristic data as a probability distribution and multiplying the median by a statistically calculated coefficient is to clearly distinguish between normal and abnormal states by reflecting the variability of the signal data. This allows for the establishment of flexible alarm criteria that consider data distribution characteristics for each equipment operating state. Furthermore, setting based on the median and coefficients enables the early detection of equipment anomalies by effectively detecting statistical outliers.

[0069] The resulting effect is to reduce unnecessary false alarms and enhance the reliability of the alarm system. Furthermore, by detecting signal changes occurring under abnormal conditions at an early stage, it is possible to improve equipment stability and operational efficiency. By setting alarm values ​​based on data distribution, equipment status can be precisely monitored under various operating conditions, thereby reducing maintenance costs and enabling flexible responses to changes in equipment status.

[0070] FIG. 8 is intended to explain the data distribution fitting and alarm value setting of a state monitoring alarm setting system for a pumped storage power plant turbine according to one embodiment of the present invention.

[0071] In Figure 8, the x-axis represents a value related to the measured vibration value (amplitude or RMS value) and signifies the magnitude of the vibration. The y-axis represents the frequency or probability density of the corresponding vibration value and shows how often a specific vibration value occurs. In other words, this graph visualizes the distribution of vibration values ​​and is used to set criteria for the median and alarm values ​​based on the frequency of occurrence corresponding to a specific vibration value.

[0072] In analyzing data distribution and setting alarm values, if the measured vibration values ​​are assumed to be random variables, they can form a continuously distributed sample. Normal data is accumulated over a certain period, and a suitable probability distribution is estimated for the statistical evaluation of the measured vibration values. The alarm values ​​are set based on the statistics obtained from the probability distribution (median*1.6, median*2.5).

[0073] The graph in Fig. 8 visually illustrates the process of analyzing data distribution and setting alarm values ​​based on vibration data measured by the condition monitoring alarm setting system of a pumped-storage power plant turbine. This graph in Fig. 8 shows a continuous data distribution formed by assuming the measured vibration values ​​as random variables. A suitable probability distribution is estimated based on normal data accumulated over a certain period, and the alarm value is set based on the median of this distribution. In the graph, the values ​​obtained by multiplying the median by 1.6 and 2.5 are set as boundaries, representing the initial warning stage and an abnormal state, respectively.

[0074] This graph emphasizes the objectivity of statistical alarm settings and provides alarm values ​​that can clearly distinguish between normal and abnormal states. This helps prevent excessive alarms or missed alarms and contributes to improving the accuracy of equipment status monitoring. Furthermore, it suggests that the variability of operating conditions can be effectively reflected by probabilistically analyzing the data.

[0075] The alarm management unit (400) monitors the abnormal condition of the equipment in real time based on the set alarm value and operating status data.

[0076] This alarm management unit can generate a model with different alarm thresholds depending on the operating status and measurement location, and can effectively monitor the turbine status by setting alarms that reflect changes in the turbine's operating characteristics.

[0077] The alarm management unit (400) is configured to monitor the abnormal condition of the equipment in real time and output an alarm in the pumped-storage power plant turbine condition monitoring and alarm setting system. This alarm management unit precisely analyzes the condition of the equipment using set alarm values ​​and operating status data, and immediately generates an alarm when an abnormal condition is detected. In addition, by applying modularized standards according to the operating status and measurement location, it enables optimized condition monitoring and alarm setting even under various operating conditions.

[0078] For reference, the alarm management unit can generate models with different alarm thresholds depending on the operating status and measurement location. These models are utilized to precisely monitor the status of equipment and detect abnormal conditions early by reflecting various operating conditions and location-specific characteristics.

[0079] For example, since generator motor bearings and water turbine bearings are exposed to different dynamic load environments, alarm models have different characteristics depending on the measurement location. At the generator motor bearing location, the focus is on monitoring vibrations caused by electromagnetic loads induced from the generator side, rotational imbalance, and shaft misalignment; in the normal range, the RMS value remains in the 1.0 to 1.5 range, and the 1X component appears as a major characteristic. On the other hand, water turbine bearings are characterized by vibrations caused by hydraulic load fluctuations, unbalanced forces transmitted from runners, cavitation, and vortices. These alarm criteria are designed based on data observed at each measurement location.

[0080] The alarm models also differ depending on the operating conditions. In the normal range, vibration remains stable, and the 1X component is a key feature. In this case, alarm criteria are set at 1.6 and 2.5 times the median value to detect unexpected abnormal conditions early. In the transient range, there is a high probability that fluid-induced vibration components, such as vortex turbulence or surge vortex ropes, will occur due to changes in output and flow rate fluctuations. In this case, the system is designed to output an alarm if the RMS value exceeds 3.0 or if the intensity of the surge vortex rope exceeds a specific threshold.

[0081] In conclusion, the alarm management unit can precisely monitor the status of the equipment and rapidly detect and respond to abnormal conditions by analyzing data by location and operating status to generate a model with optimized alarm thresholds.

[0082] FIG. 9 is a flowchart of a method using a state monitoring alarm setting system (hereinafter referred to as the system) of a pumped storage power plant turbine according to one embodiment of the present invention.

[0083] As illustrated in FIG. 9, in a method for monitoring the status of a pumped storage power plant turbine and setting an alarm, (a) the system collects a set of operating data including output and flow rate, and based on this, determines the operating status into a normal range and a transient range.

[0084] Next, (b) the system analyzes vibration data measured for each determined operating state to identify order components and extracts signal characteristic data corresponding to each operating state.

[0085] Here, signal characteristic data corresponding to each operating state refers to key vibration characteristics analyzed based on vibration data measured in the normal and transient sections. The normal section includes stable vibration characteristics such as the amplitude and RMS values ​​of major order components such as 1X and 2X, while the transient section includes RMS values, amplitude, and frequency characteristics related to fluid-induced vibration patterns such as surge vortex rope, vortex turbulence, and cavitation.

[0086] Next, (c) the system probabilistically analyzes the distribution of vibration values ​​based on the extracted signal characteristic data and sets appropriate alarm values ​​for each operating state through statistical evaluation.

[0087] Next, (d) the system monitors the abnormal condition of the equipment in real time based on the set alarm values ​​and operating status data.

[0088] The operation of this system can be briefly summarized as follows.

[0089] First, data such as the equipment's output, flow rate, and guide vane opening are collected in real time to determine the operating status into normal and transient ranges. For example, if the output is stable at 80% and the flow rate fluctuation is 2% or less, it is determined to be in the normal range, and if the output rapidly decreases to 50% and the flow rate fluctuation exceeds 5%, it is determined to be in the transient range.

[0090] Vibration data measured according to operating conditions is analyzed to identify Order components (1X, 2X, 9X, etc.), and the RMS values ​​and amplitude ranges of major Order components are extracted during normal operation. For example, if the RMS value remains constant between 1.2 and 1.5 during normal operation and the 1X component is prominent, it is determined to be in a normal state. During transient operation, specific vibration patterns such as surge vortex ropes or vortex turbulence occur, which may indicate changes in the equipment condition.

[0091] Alarm values ​​are set by analyzing the probability distribution of the data based on the extracted signal characteristic data. For example, when the median RMS value in the normal range is found to be 1.3, 1.6 times the median (approx. 2.08) is set as the first alarm value, and 2.5 times the median (approx. 3.25) is set as the second alarm value. These alarm values ​​are designed to detect abnormal conditions of the equipment at an early stage.

[0092] The alarm management unit monitors abnormal conditions by comparing set alarm values ​​with real-time data. For example, if the RMS value rises to 2.1 in the normal range, it outputs the first alarm, and if it rises to 3.3, it outputs the second alarm to indicate the dangerous condition of the equipment. In the transient range, it extracts the band value where the surge vortex rope is observed and warns of changes in the equipment condition.

[0093] In addition, the alarm management unit applies modularized standards based on measurement locations. For example, it focuses on monitoring shaft vibration patterns at the upper bearing location and analyzes cavitation vibration patterns at the lower bearing location to set alarm values. This enables effective monitoring of specific vibrations that may occur at various locations within the equipment.

[0094] In summary, the method for setting alarm thresholds by analyzing vibration data based on operating conditions and measurement locations is as follows. First, the operating conditions of the equipment are classified into steady-state and transient states. In the steady-state, stable vibration patterns are analyzed focusing on major order components (e.g., 1X, 2X) and RMS values, while in the transient state, fluid-induced vibration components such as surge vortex ropes and vortex turbulence are additionally analyzed. The distribution of vibration data collected in each operating state is statistically evaluated.

[0095] Regarding measurement locations, data measured at the equipment's upper and lower bearings are analyzed individually. For instance, since shaft vibration primarily occurs in the upper bearing and cavitation vibration in the lower bearing, the data is analyzed by location to account for these factors.

[0096] Based on the analyzed data, a probability distribution is estimated, and an alarm threshold is set by multiplying the statistically derived median by a coefficient (1.6 times, 2.5 times). The first alarm value indicates an initial warning, and the second alarm value is used to warn of an abnormal condition.

[0097] The configured alarm value is compared with real-time measured data, and if vibrations exceeding the alarm value are detected, it warns of an abnormal condition in the equipment. This enables the early detection of equipment abnormalities and a rapid response. This method enhances the precision of equipment status monitoring and allows for customized alarm settings that reflect operating conditions and location-specific characteristics.

[0098] In conclusion, this system sets customized alarm values ​​by reflecting the operating status of the equipment and characteristics of each measurement location, and monitors abnormal conditions of the equipment in real time based on these values. This enables a clear distinction between normal and abnormal states, enhances equipment stability, and improves maintenance efficiency.

Claims

1. An operating state determination unit that collects an operating data set and determines the operating state based on operating parameters including output and flow rate; A signal characteristic extraction unit that analyzes vibration data measured for each operating state to identify order components and extracts signal characteristics for each operating state; A data distribution analysis unit that probabilistically analyzes the distribution of vibration values ​​based on extracted signal characteristic data and sets appropriate alarm values ​​for each operating state through statistical evaluation; A condition monitoring alarm setting system for a pumped storage power plant turbine, comprising: an alarm management unit that monitors the abnormal condition of the equipment in real time based on the alarm value and operating status data set above.

2. In Paragraph 1, The above-mentioned operating state determination unit determines the normal section and the transient section based on the output change rate, the flow rate change rate, and the guide vane opening, and A condition monitoring alarm setting system for a pumped-storage power plant turbine, characterized by using a minimum value of flow rate change and a threshold value of output change rate as criteria for determining the transient section.

3. In Paragraph 1, The above signal characteristic extraction unit extracts PP values ​​or RMS values ​​according to the Order component through frequency analysis of vibration data measured for each operating state, and In the normal range, it identifies specific vibration patterns based on the range of amplitude and RMS values ​​of the 1X component, and in the transient range, it identifies specific vibration patterns, A condition monitoring alarm setting system for a pumped storage power plant turbine, characterized by defining characteristic indicators including frequency, amplitude, range and variability of RMS values ​​for each section based on this.

4. In Paragraph 1, The above data distribution analysis unit assumes the extracted signal characteristic data as a probability distribution, and A condition monitoring alarm setting system for a pumped storage power plant turbine characterized by setting an alarm value by multiplying a median value by a statistically calculated coefficient.

5. A method using a condition monitoring alarm setting system for a pumped-storage power plant turbine, (a) A step in which the system collects an operation data set including output and flow rate, and determines the operation state into a normal section and a transient section based thereon, (b) A step of analyzing vibration data measured for each operating state determined by the above system to identify order components and extracting signal characteristic data corresponding to each operating state, (c) A step in which the above system probabilistically analyzes the distribution of vibration values ​​based on the extracted signal characteristic data and sets appropriate alarm values ​​for each operating state through statistical evaluation, and (d) A method using a pumped storage power plant turbine condition monitoring alarm setting system comprising the step of monitoring the abnormal condition of the equipment in real time based on alarm values ​​and operating status data set by the above system.

Images