A centrifugal pump health assessment method

Abstract

This invention provides a method for assessing the health of a centrifugal pump, comprising: determining the static weights of each indicator based on a standardized set of health assessment indicators; determining a first health score for each indicator based on its historical mean baseline, and a second health score based on its corresponding safety threshold boundary; assigning dynamic weights to the first and second health scores by defining risk factors, and weighting and fusing the first and second health scores according to the dynamic weights to obtain a single health index for each indicator; weighting and fusing the single health indices of each indicator according to the static weights corresponding to the standardized indicator set to obtain a comprehensive health index for the centrifugal pump; and determining the health status level of the centrifugal pump based on the comprehensive health index. This invention effectively improves the sensitivity of identifying early progressive degradation of centrifugal pumps and achieves the goal of accurately capturing degradation trends.

Description

Technical Field

[0001] This invention relates to the field of centrifugal pump health assessment technology, and in particular to a method for centrifugal pump health assessment. Background Technology

[0002] Centrifugal pumps operating in environments with high sediment content and variable operating conditions are often affected by various factors such as fluctuations in the sediment content of the medium, cumulative mechanical wear, and on-site noise interference. Their health status exhibits a complex evolution process or weak early deterioration characteristics. Most existing assessment methods directly compare a single collected parameter with a preset threshold. This single-benchmark assessment method cannot effectively analyze the interrelationships between multi-source monitoring information, thus failing to provide reliable quantitative data support for predictive maintenance of the equipment.

[0003] To address the technical challenge of analyzing the correlation of multi-source information using a single benchmark, existing technologies propose a comprehensive health status assessment method for water pumps that incorporates weighting factors. This method primarily assigns fixed static weights to various indicators by setting a fixed operating cycle or static threshold, combining subjective Analytic Hierarchy Process (AHP) and objective entropy weighting. Subsequently, the collected equipment operating data is compared with the preset thresholds to assess the status and trigger early warnings.

[0004] However, the aforementioned existing technologies rely on fixed alarm thresholds for evaluation, and the indicator weights are static weights lacking adaptive adjustment capabilities, failing to reflect the slow performance degradation trend of the equipment. This fixed evaluation model, due to insufficient sensitivity, has a weak ability to detect early, gradual degradation, easily generating false alarms under normal operating condition fluctuations or missing optimal maintenance opportunities, leading to a distorted overall health score. Therefore, how to improve the sensitivity of centrifugal pump health assessment under complex multi-source operating conditions to accurately capture early degradation trends of centrifugal pumps has become an urgent problem to be solved in this field. Summary of the Invention

[0005] This invention provides a centrifugal pump health assessment method to solve the technical problem of how to improve the sensitivity of centrifugal pump health assessment under complex multi-source operating conditions in order to accurately capture the early deterioration trend of centrifugal pumps.

[0006] This invention provides a method for assessing the health of a centrifugal pump, characterized by comprising: Multiple health assessment index sets for centrifugal pumps are standardized and dimensionality reduced to obtain multiple standardized index sets. The static weights of each indicator are determined based on the various standardized indicator sets. The first health score of the indicator is determined based on the historical mean baseline corresponding to the indicator, and the second health score of the indicator is determined based on the safety threshold boundary corresponding to the indicator. By defining risk factors to assign dynamic weights to the first health score and the second health score, and by weighting and fusing the first health score and the second health score according to the dynamic weights, a single health index of the indicator is obtained. The single health index of each indicator is weighted and fused according to the static weights corresponding to the standardized indicator set to obtain the comprehensive health index of the centrifugal pump. The health status level of the centrifugal pump is determined based on the comprehensive health index.

[0007] According to the present invention, a centrifugal pump health assessment method is provided, wherein the health assessment index set includes, but is not limited to, rotational speed, head, flow rate, pump outlet pipe pressure, pump inlet pipe pressure, peak-to-peak value of spindle swing, effective value of vibration velocity at the drive end, effective value of vibration velocity at the non-drive end, effective value of vibration displacement at the drive end, and effective value of vibration displacement at the non-drive end.

[0008] According to a centrifugal pump health assessment method provided by the present invention, the health assessment index set includes flow rate and rotational speed, and the standardized index set includes flow coefficient; The dimensionality reduction of the health assessment indicator set includes: The flow coefficient is determined based on the flow rate and the rotational speed.

[0009] According to a centrifugal pump health assessment method provided by the present invention, the health assessment index set includes head and speed, and the standardized index set includes head coefficient; The dimensionality reduction of the health assessment indicator set includes: The head coefficient is determined based on the head and the rotational speed.

[0010] According to a centrifugal pump health assessment method provided by the present invention, a standardization process is performed on multiple sets of health assessment indicators, including: Determine the deviations between the first type of indicators and their corresponding theoretical values ​​in each of the aforementioned health assessment indicator sets; The deviations of each indicator are normalized based on the maximum and minimum values ​​of the indicator deviations, thereby completing the standardization of the first type of indicators.

[0011] According to a centrifugal pump health assessment method provided by the present invention, a standardization process is performed on multiple sets of health assessment indicators, including: Determine the maximum and minimum values ​​of each second-category indicator in each set of health assessment indicators; The second category of indicators is normalized based on the maximum and minimum values ​​of the second category of indicators, thus completing the standardization of the second category of indicators.

[0012] According to a centrifugal pump health assessment method provided by the present invention, the static weight of each indicator is determined based on the various standardized indicator sets, including: The static weights of each standardized index set are obtained by calculating the weights using the entropy weight method or the Critic method.

[0013] According to the centrifugal pump health assessment method provided by the present invention, the historical mean baseline includes a first positive ideal solution and a first negative ideal solution, and the safety threshold boundary includes a second positive ideal solution and a second negative ideal solution; Under the historical average baseline, the average value of each indicator of the equipment during the historical healthy operation phase is selected as the first positive ideal solution, and the boundary value that deviates from the first positive ideal solution to a certain extent is set as the first negative ideal solution; Under the safety threshold boundary, for hydraulic performance indicators, a very narrow allowable interval near the optimal curve is selected as the second positive ideal solution, and a limit line of ±20% deviation from the optimal curve is selected as the second negative ideal solution. For mechanical stability indicators, the second positive ideal solution is set to 0, and the value specified in relevant national standards or engineering design is used as the second negative ideal solution. The first health score of the indicator is determined based on the historical mean baseline corresponding to the indicator, and the second health score is determined based on the safety threshold boundary corresponding to the indicator, including: The first Euclidean distance between the target index in the standardized index set and the corresponding first positive ideal solution, the second Euclidean distance between the target index and the corresponding first negative ideal solution, the third Euclidean distance between the target index and the corresponding second positive ideal solution, and the fourth Euclidean distance between the target index and the corresponding second negative ideal solution are determined respectively. The first health score of the target indicator is determined based on the first Euclidean distance and the second Euclidean distance, and the second health score of the target indicator is determined based on the third Euclidean distance and the fourth Euclidean distance.

[0014] According to a centrifugal pump health assessment method provided by the present invention, a risk factor is defined as a dynamic weight assigned to a first health score and a second health score, and the first health score and the second health score are weighted and fused according to the dynamic weight to obtain a single health index, including: The first risk factor corresponding to the first health score and the second risk factor corresponding to the second health score are determined according to a preset exponential function relationship; wherein, the risk factor is used to characterize the significance of the abnormal risk of the corresponding health score, and the closer the health score is to the abnormal limit, the larger the corresponding risk factor is. A risk game model is constructed based on the first risk factor and the second risk factor. The first risk factor is divided by the sum of the first risk factor and the second risk factor to obtain the first dynamic weight. Divide the second risk factor by the sum of the first risk factor and the second risk factor to obtain the second dynamic weight; The first health score and the second health score are weighted and summed using the first dynamic weight and the second dynamic weight to obtain the single health index.

[0015] According to a centrifugal pump health assessment method provided by the present invention, the health status level of the centrifugal pump is determined based on the comprehensive health index, including: The health status level of the centrifugal pump is determined based on the technical indicators of the comprehensive health index, risk assessment requirements, and long-term operating experience.

[0016] The centrifugal pump health assessment method provided by this invention introduces a dual-benchmark assessment of historical average baseline and safety threshold boundary, and combines static weights for weighted fusion, which effectively improves the sensitivity of identifying early progressive deterioration of centrifugal pumps. It solves the problem of distortion of comprehensive health score caused by the weak ability of assessment model to capture deterioration under complex working conditions with a lot of sediment, and achieves the goal of accurately capturing the deterioration trend. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the centrifugal pump health assessment method provided by the present invention.

[0019] Figure 2 This is a schematic diagram of the centrifugal pump health assessment device provided by the present invention.

[0020] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] Centrifugal pumps operating in environments with high sediment content and variable operating conditions are often simultaneously affected by multiple factors, including fluctuations in the sediment content of the medium, accumulated mechanical wear, and on-site noise interference. Their health status exhibits a complex evolution process or weak early deterioration characteristics. Most existing health assessment methods compare collected parameters such as vibration, temperature, and oscillation with preset thresholds. While these methods are relatively direct, they have the following shortcomings under conditions of high sediment content and strong disturbance: (1) The evaluation method based on a single benchmark cannot analyze the correlation between multiple sources of information. For example, during the operation of the unit, the bearing wear may simultaneously cause changes in the bearing base temperature and an increase in vibration energy in a specific frequency band, resulting in a decrease in the unit's working efficiency. The reason may be related to parameters such as inlet and outlet pressure and flow rate. However, a single evaluation benchmark cannot capture such comprehensive fault characteristics, which makes it impossible to provide reliable data support for predictive maintenance.

[0023] (2) Insufficient ability to identify deterioration and thus unable to meet the requirements of predictive maintenance for sensitivity and accuracy. For example, as the unit's operating time increases, under the influence of external conditions such as mud and sand abrasion, the various performance parameters of the pump station gradually deteriorate, and the resulting performance baseline shows a trend of decline, causing the original reserved margin to gradually shrink. This can easily lead to the unit being significantly deteriorated in the later stages but not exceeding the set threshold, resulting in missed alarms, or the unit's parameters briefly exceeding the set threshold under normal conditions, resulting in false alarms.

[0024] Furthermore, for newly built or short-term operational pumping stations, the following methods and their limitations address the issue of an imbalance between healthy and faulty samples due to the difficulty in accumulating sufficient effective data during the initial stage of equipment operation, where the equipment is generally healthy. (1) Machine learning: Machine learning methods such as random forests and neural networks require a large amount of raw data to train the model. If there are insufficient effective samples, the trained model may not be able to learn effective judgment conditions, resulting in poor fault identification ability and thus false negatives.

[0025] (2) Artificial reproduction: that is, to obtain the required data by artificially creating or simulating specific faults. Although the required data can be obtained in a targeted manner, the artificial reproduction cycle is long, which increases the cost. It is often difficult to achieve in actual industry. At the same time, artificial reproduction cannot fully simulate the natural deterioration caused by complex working conditions in actual operation, and the data obtained is not authentic enough.

[0026] A certain existing technology adopts a traditional method for assessing the health status of pumping stations. The core of this method is to assess the status and trigger early warnings by setting a fixed operating cycle or static threshold and combining subjective AHP and objective entropy weighting methods to allocate weights. Rather than objectively and comprehensively assessing the operational health status of the pumping station units in real time.

[0027] The existing technology performs health status checks on pump station units based on a fixed operating cycle, relying on the comparison between a single static threshold to trigger early warnings. The detailed steps are as follows: Step 1: Determine the monitoring cycle and collect raw data. Based on historical experience or manufacturer recommendations, set fixed operating time periods and collect raw data on basic parameters such as vibration and temperature. Note that data collection should be done during maintenance, not continuously.

[0028] Step Two: Establish and normalize the assessment system. Based on the collected data, determine the key assessment indicators (such as vibration, sway, temperature, etc.), preprocess the determined indicator parameters to convert them into quantitative values ​​for health status assessment, and assign assessment levels.

[0029] Step 3: Threshold comparison triggers early warning. The collected data is compared with preset thresholds. If the data exceeds the set threshold, an alarm is triggered and maintenance is initiated; if the data does not exceed the threshold, the system is considered to be in normal operation and continues to operate until the next maintenance period.

[0030] Step 4: Construct a fuzzy relation matrix to complete the transformation from numerical values ​​to evaluation states. A fuzzy relation matrix is ​​constructed using the transformed quantified values ​​and the mapped evaluation levels. The weights assigned by AHP and entropy weighting methods are combined and normalized to obtain the weight vectors for each indicator.

[0031] Step 5: Construct a fuzzy comprehensive evaluation system. Using fuzzy operators, the weights and fuzzy relation matrix mentioned above are synthesized to obtain a comprehensive evaluation vector, which is then transformed to obtain a percentage-based health assessment result.

[0032] The sensor dataset in this prior art only includes a subset of sensors, such as vibration sensors (measuring peak-to-peak values), temperature sensors (monitoring the temperature of electrical and mechanical components), and noise sensors used for auxiliary analysis (such as cavitation). Data acquisition is primarily based on a single point in time and does not consider continuous variations. Specifically, the vibration peak-to-peak value is calculated using a simple partitioning method: the difference between the maximum and minimum values ​​in each partition of the time-domain waveform is taken as the peak-to-peak value, and the average of all partition peak-to-peak values ​​is then taken as the overall peak-to-peak value. For temperature, the average of the maximum values ​​within the current operating period is used as the overall mean.

[0033] The existing technology sets static thresholds based on relevant expert experience and industry standards, such as vibration alarm threshold A0 (e.g., 200μm), emergency shutdown threshold A1 (e.g., 300μm), temperature alarm threshold C0 (e.g., 80℃), and emergency shutdown threshold C1 (e.g., 100℃). The setting of these thresholds may be unreasonable and inaccurate.

[0034] The existing alarm mechanism is triggered based solely on whether a single evaluation criterion is exceeded, without incorporating multiple parameter weights, thus lacking objectivity. For example, if the vibration value exceeds A0, an alarm is triggered; if the vibration value exceeds A1, a shutdown state is directly triggered.

[0035] The existing technology uses static weights for its indicators, which do not consider the overall operating status of the pumping station and lack adaptive adjustment. Static weights cannot adjust the pumping station according to different health conditions. For example, in the later stages of unit operation, bearing wear may become the dominant cause, thus increasing the importance of vibration and runout parameters. However, static weights cannot achieve adaptive adjustment, which may reduce sensitivity and cause the optimal maintenance time to be missed.

[0036] The existing technology's evaluation benchmark relies on preset alarm thresholds A0 and emergency shutdown thresholds A1, which cannot reflect continuous deterioration trends, resulting in weak early warning capabilities and limited stability. For example, if the spindle runout value continuously and slowly decreases from an excellent level to near but not exceeding the alarm value, the scoring calculation formula may not change much, making it impossible to reflect the slow deterioration trend.

[0037] The maintenance time of this existing technology depends on a fixed operating cycle, which can easily lead to repeated maintenance of units that are in good operating condition, increasing unnecessary maintenance costs. At the same time, repeated maintenance results in a large amount of maintenance work and low maintenance efficiency; or the unit may fail to deal with faults in time, causing the best maintenance stage to be missed, increasing operating costs.

[0038] The following is combined Figures 1 to 3 The present invention describes a centrifugal pump health assessment method.

[0039] Figure 1 This is a flowchart illustrating the centrifugal pump health assessment method provided by the present invention, as shown below. Figure 1 As shown, the method includes, but is not limited to, steps S1, S2, S3, S4, S5, and S6.

[0040] Step S1: Standardize and reduce the dimensionality of multiple health assessment index sets for centrifugal pumps to obtain multiple standardized index sets.

[0041] The health assessment index set for centrifugal pumps is a data set characterizing the operating status of centrifugal pumps under different dimensions. These indicators can be collected in real time by a multi-sensor condition monitoring system installed on the centrifugal pump unit.

[0042] To eliminate errors caused by different units of measurement among different indicators and highlight the essential changes in performance, it is necessary to normalize the data of each indicator so that the data is mapped to the interval [0, 1], thereby obtaining a standardized indicator set that eliminates the influence of units of measurement. This provides a unified and standardized data foundation for subsequent weight calculation and health status scoring, ensuring the accuracy of the assessment.

[0043] Step S2: Determine the static weights of each indicator based on each standardized indicator set.

[0044] Static weights can be assigned to standardized indicators using objective weighting methods (such as entropy weighting or the CRITIC method). Information entropy can be used to measure the importance of each indicator in the comprehensive evaluation. By calculating the information entropy provided by each standardized indicator, the more volatile the data, the more information it provides, and the greater its corresponding static weight. Step S2 ensures the objectivity of indicator weight allocation and avoids biases caused by subjective experience.

[0045] Step S3: Determine the first health score of the indicator based on the historical average baseline corresponding to the indicator, and determine the second health score of the indicator based on the safety threshold boundary corresponding to the indicator.

[0046] To quantify equipment health status, this invention proposes using both historical average baselines (mean method) and safety threshold boundaries (threshold method) as dual benchmarks for preliminary assessment. The Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) can be used to calculate the relative closeness index of standardized indicators under these two different benchmarks. The calculated relative closeness under the mean method is used as the first health score, reflecting the deviation trend of the equipment's current state from its historical health baseline; the calculated relative closeness under the threshold method is used as the second health score, reflecting the distance of the equipment's current state from the absolute safety boundary.

[0047] Step S3 uses a dual-benchmark independent evaluation, which can remain sensitive to slow performance degradation and respond quickly to sudden anomalies or instantaneous over-limits.

[0048] Step S4: By defining risk factors and assigning dynamic weights to the first and second health scores, and by weighting and fusing the first and second health scores according to the dynamic weights, a single health index is obtained.

[0049] By introducing game theory models (such as the Shapley value model), the mean method and threshold method are treated as participants in a cooperative game. The contribution of these two assessment methods to the overall risk warning at each time step is quantified, and dynamic fusion weights are assigned accordingly. These dynamic fusion weights are then used to weight and sum the first and second health scores to obtain a fused single health index.

[0050] Step S5: Based on the static weights corresponding to the standardized indicator set, the individual health indices of each indicator are weighted and fused to obtain the comprehensive health index of the centrifugal pump.

[0051] Steps S4 and S5 can achieve dynamic and static weighted fusion of multiple benchmarks and indicators to form a comprehensive health index that can fully quantify the overall health level of the centrifugal pump.

[0052] Using the objective static weights obtained in step S2, the individual health indices of each indicator are weighted and summed to obtain the comprehensive health index of the centrifugal pump.

[0053] Step S6: Determine the health status level of the centrifugal pump based on the comprehensive health index.

[0054] The obtained comprehensive health index can be classified into corresponding health status levels through pre-defined numerical range mapping rules. For example, the comprehensive health index is between [0, 100], and the closer the value is to 100, the closer the overall operating status of the water pump is to the ideal health status. The health status level of the centrifugal pump can be determined based on the technical indicators of the comprehensive health index, risk assessment requirements, and long-term operating experience: If the index is in the range of 80-100, it indicates that the equipment is in a healthy level and is judged as "excellent"; if the index is in the range of 60-80, it indicates that the equipment is in a normal state with no obvious abnormalities and is in an acceptable state, but there may be slight fluctuations, and routine monitoring is recommended, and it is judged as "normal"; if the index is in the range of 40-60, it indicates that the equipment is in a stage that requires attention, and the overall operation has slightly deteriorated, requiring increased attention and timely maintenance, and it is judged as "requires attention"; if the index is below 40, it indicates that the equipment is in an abnormal stage, and the overall operation has obviously deteriorated, requiring immediate repair to prevent the risk of failure from escalating, and it is judged as "abnormal". This allows complex, multi-dimensional data analysis results to be transformed into levels with clear physical meaning, providing intuitive guidance for subsequent operations.

[0055] As can be seen from the above, the centrifugal pump health assessment method of the present invention effectively improves the sensitivity of identifying early progressive deterioration of centrifugal pumps by introducing a dual benchmark assessment of historical mean baseline and safety threshold boundary, and combining static weights for weighted fusion. It solves the problem of distortion of comprehensive health score caused by the weak ability of assessment model to capture deterioration under complex working conditions with multiple sediments, and achieves the goal of accurately capturing the deterioration trend.

[0056] In one embodiment, the health assessment index set of the present invention may include, but is not limited to, rotational speed, head, flow rate, pump outlet pipe pressure, pump inlet pipe pressure, peak-to-peak value of spindle swing, effective value of vibration velocity at the drive end, effective value of vibration velocity at the non-drive end, effective value of vibration displacement at the drive end, and effective value of vibration displacement at the non-drive end.

[0057] The peak-to-peak value of the spindle runout can be obtained from the spindle runout sensor that monitors the dynamic eccentricity of the impeller or spindle. The effective values ​​of vibration velocity and vibration displacement at the drive and non-drive ends are used to detect common mechanical faults such as rotor imbalance and bearing wear, and can be monitored separately by vibration sensors installed on the bearing housings at the drive and non-drive ends of the water pump. These five indicators—peak-to-peak value of spindle runout, effective value of vibration velocity at the drive end, effective value of vibration velocity at the non-drive end, effective value of vibration displacement at the drive end, and effective value of vibration displacement at the non-drive end—directly reflect the dynamic balance accuracy of the water pump rotor, the bearing operating condition, and the overall mechanical vibration level.

[0058] This invention, by selecting multi-dimensional key indicators such as flow rate, head, and vibration swing at multiple locations, can comprehensively and accurately reflect the hydraulic efficiency and mechanical operating status of centrifugal pumps, providing high-information data support for accurately capturing early deterioration trends.

[0059] In one embodiment, the health assessment index set of the present invention may include flow rate and rotation speed, and the standardized index set may include flow rate coefficient; In step S1, the dimensionality reduction of the health assessment indicator set may further include: The flow coefficient is determined based on the flow rate and rotational speed.

[0060] To eliminate interference caused by fluctuations in operating speed, the flow coefficient can be calculated using the original flow rate and impeller speed through a dimensionless hydraulic formula, thus unifying the operating conditions. The formula for calculating the flow coefficient is as follows: ; in, denoted as the flow coefficient, Q as the initial flow rate, D as the nominal diameter of the centrifugal pump impeller, and u as the impeller speed.

[0061] This invention calculates the dimensionless flow coefficient by obtaining the original flow rate and impeller speed, which can effectively eliminate data interference caused by fluctuations in operating conditions (such as changes in speed) and improve the stability of indicator data and the accuracy of health status assessment.

[0062] In one embodiment, the health assessment index set of the present invention may include head and speed, and the standardized index set may include head coefficient. In step S1, the dimensionality reduction of the health assessment indicator set may further include: The head coefficient is determined based on the head and rotational speed.

[0063] This invention utilizes hydraulic formulas to calculate the head coefficient, thereby eliminating the interference of rotational speed fluctuations. The formula for calculating the head coefficient is as follows: ; in, denoted as the head coefficient, H as the original head, u as the impeller speed, and g as the gravitational acceleration constant.

[0064] This invention calculates a dimensionless head coefficient using the original head and rotational speed, which can further eliminate the influence of rotational speed fluctuations under varying operating conditions on the evaluation of pump energy conversion efficiency and improve the robustness and accuracy of the health assessment model.

[0065] In one embodiment, the standardization process for multiple health assessment indicator sets in step S1 may further include: Determine the deviations between the first type of indicators and their corresponding theoretical values ​​in each set of health assessment indicators; The deviations of each indicator are normalized based on the maximum and minimum values ​​of the indicator deviations, thus completing the standardization of the first type of indicators.

[0066] The first type of indicator is the stable indicator, such as the flow coefficient and head coefficient. For this type of indicator, it is best to be close to its ideal theoretical value.

[0067] Specifically, for the i-th health assessment indicator set, first calculate the absolute difference between the j-th indicator in the health assessment indicator set and the theoretical value to obtain the indicator deviation. Then, the deviations of each indicator are normalized according to the following formula: ; in, For standardized indicators, Let $\frac{j}{j}$ be the maximum value among the index deviations corresponding to the $j$-th index in each set of health assessment indicators. It represents the minimum value among the index deviations corresponding to the j-th index in each set of health assessment indicators.

[0068] This invention employs a normalization method based on index deviation for the first type of stable index, which accurately quantifies the degree of deviation between actual operating parameters and the theoretical optimal state, providing a reliable standardized processing mechanism for subsequent dimensionless fusion of multiple parameters.

[0069] In one embodiment, the standardization process for multiple health assessment indicator sets in step S1 may further include: Determine the maximum and minimum values ​​of each Category II indicator in each set of health assessment indicators; The second category of indicators is normalized based on the maximum and minimum values ​​of the second category of indicators, thus completing the standardization of the second category of indicators.

[0070] The second type of indicator is a cost-related indicator or a negative indicator, such as the peak-to-peak value of spindle swing, the effective value of vibration velocity at the drive end, the effective value of vibration velocity at the non-drive end, the effective value of vibration displacement at the drive end, and the effective value of vibration displacement at the non-drive end. The smaller the value of this type of indicator (i.e., the more negative), the better the equipment condition.

[0071] Specifically, first, determine the maximum and minimum values ​​of the j-th indicator in each set of health assessment indicators, and then normalize the j-th indicator according to the following formula: ; in, For the j-th indicator in the i-th health assessment indicator set, For each set of health assessment indicators, the maximum value of the j-th indicator is... It is the minimum value among the j-th indicators in each set of health assessment indicators.

[0072] This invention performs a special maximum-minimum normalization process for cost-type (negative) indicators, which realizes the effective mapping of multi-source heterogeneous data to a positive unified interval, and ensures the consistency of the calculation caliber of the comprehensive evaluation model.

[0073] In one embodiment, step S2 may further include: The static weights of each indicator are obtained by calculating the static weights of each standardized indicator set using the entropy weight method or the Critic method.

[0074] Entropy weighting is an objective weighting method that determines weights based on the dispersion of indicator data; the Critic method is an objective weighting method that simultaneously considers the comparative strength of indicators and the conflict between indicators. This invention uses entropy weighting as an example for illustration. The calculation process of entropy weighting is as follows: Determine the proportion of the target indicator in each standardized indicator set within the same category of indicators. The information entropy of the target indicator is determined based on the weight of each indicator. Determine the difference coefficient of the target indicator based on information entropy; Based on the difference coefficients corresponding to the standardized indicator set, the difference coefficients of the target indicator are normalized to obtain the static weight of the target indicator.

[0075] Specifically, the proportion of the j-th indicator in the i-th standardized indicator set among all similar indicators. It can be represented as: ; Where m is the total number of standardized index sets.

[0076] The weight of an indicator reflects the relative contribution of a single standardized indicator set to the overall indicator distribution and is the basis for calculating information entropy.

[0077] Information entropy of the j-th indicator It can be represented as: ; Information entropy is in the range [0, 1]. If the information entropy is large, it indicates that the indicator provides less information and has a low degree of dispersion; conversely, the more drastic the data fluctuation, the smaller the information entropy and the more information it provides.

[0078] Coefficient of difference The larger the difference coefficient, the higher the status of the indicator.

[0079] The static weight of the j-th indicator It can be represented as: ; When the unit is in a healthy and stable period, the static weight distribution of each indicator is relatively balanced; if a certain indicator fluctuates abnormally due to the risk of failure, its difference coefficient increases instantly, and the corresponding static weight is automatically increased. In the subsequent comprehensive evaluation, the contribution of the abnormal indicator will be amplified, and the abnormal stage of equipment operation will be quickly captured.

[0080] This invention utilizes an objective weighting method based on information entropy (entropy weighting method) to allocate static weights. This method can effectively utilize the fluctuation characteristics and dispersion of the data itself to measure the importance of each indicator, eliminate the interference of subjective experience, and improve the objectivity and scientific nature of the weights fused from multi-source information.

[0081] In one embodiment, the historical mean baseline of the present invention includes a first positive ideal solution and a first negative ideal solution, and the safety threshold boundary includes a second positive ideal solution and a second negative ideal solution; Under the historical average baseline, the average value of each indicator of the equipment during the historical healthy operation phase is selected as the first positive ideal solution, and the boundary value that deviates from the first positive ideal solution to a certain extent is set as the first negative ideal solution; Under the safety threshold boundary, for hydraulic performance indicators, a very narrow allowable interval near the optimal curve is selected as the second positive ideal solution, and a limit line of ±20% deviation from the optimal curve is selected as the second negative ideal solution. For mechanical stability indicators, the second positive ideal solution is set to 0, and the value specified in relevant national standards or engineering design is used as the second negative ideal solution. The first health score of the indicator is determined based on the historical mean baseline corresponding to the indicator, and the second health score is determined based on the safety threshold boundary corresponding to the indicator, including: Determine the first Euclidean distance between the target index and the corresponding first positive ideal solution, the second Euclidean distance between the target index and the corresponding first negative ideal solution, the third Euclidean distance between the target index and the corresponding second positive ideal solution, and the fourth Euclidean distance between the target index and the corresponding second negative ideal solution in the standardized index set, respectively. The first health score of the target indicator is determined based on the first and second Euclidean distances, and the second health score of the target indicator is determined based on the third and fourth Euclidean distances.

[0082] The first positive ideal solution under the historical mean baseline (i.e. the mean method) can be the average value of each index data of the centrifugal pump during the historical healthy operation phase, and the first negative ideal solution can be the boundary value that deviates from the first positive ideal solution to a certain extent (such as ±20% deviation from the first positive ideal solution for hydraulic performance indicators).

[0083] Under the safety threshold boundary (i.e., the threshold method), for hydraulic performance indicators, the second positive ideal solution can be set as an extremely narrow allowable range near the optimal performance curve; for mechanical stability indicators, the positive ideal solution is usually set to 0, i.e., the ideal state of no vibration and no sway. For hydraulic performance indicators, the second negative ideal solution is usually set as a limit line deviating from the optimal performance curve by ±20%; exceeding this range is generally considered as substandard performance. For mechanical stability indicators, the second negative ideal solution directly adopts the value specified in relevant national standards or engineering designs.

[0084] The first and second Euclidean distances can be calculated using the Euclidean distance formula. The first health score for the j-th indicator... It can be represented as = Second Euclidean distance / (First Euclidean distance + Second Euclidean distance). The health score, also known as the relative proximity score, is located in the interval [0, 1]. The closer the value is to 1, the closer the indicator status is to the optimal ideal solution under the corresponding benchmark at that moment, and the better the health status; conversely, the closer it is to 0, the worse the status is.

[0085] The third and fourth Euclidean distances can be calculated using the Euclidean distance formula. The second health score for the j-th indicator... It can be represented as = Fourth Euclidean distance / (Third Euclidean distance + Fourth Euclidean distance).

[0086] This invention utilizes historical health averages and deviations to construct positive and negative ideal solutions and calculates Euclidean distance. This accurately reflects the degree of deviation of the current state from its historical best state, greatly enhancing the ability to capture subtle and slowly progressive early deterioration trends. By setting positive and negative ideal solutions based on safety threshold boundaries and calculating spatial proximity, it can quickly and accurately reflect instantaneous drastic anomalies or over-limit risks of equipment parameters, providing a reliable and rapid quantitative judgment benchmark for monitoring sudden failures.

[0087] Based on the above calculations, this invention can output two complete health score time-series curves. The mean method reflects the deviation trend of the current state of the equipment relative to its historical health baseline, and is more sensitive to slow performance degradation, but may lag in responding to sudden anomalies. The threshold method reflects the distance of the current state of the equipment from the absolute safety boundary, and responds quickly to instantaneous parameter over-limits or drastic fluctuations, but is prone to false alarms due to fluctuations in normal operating conditions. The dual benchmarks also provide a basis for the dynamic fusion of risk game theory in the following sections.

[0088] In one embodiment, step S4 may further include: The first risk factor corresponding to the first health score and the second risk factor corresponding to the second health score are determined according to a preset exponential function relationship. The risk factor is used to characterize the significance of the abnormal risk of the corresponding health score, and the closer the health score is to the abnormal limit, the larger the corresponding risk factor is. A risk game model is constructed based on the first risk factor and the second risk factor. The first risk factor is divided by the sum of the first risk factor and the second risk factor to obtain the first dynamic weight. The second dynamic weight is obtained by dividing the second risk factor by the sum of the first and second risk factors. A single health index is obtained by weighting and summing the first health score and the second health score using the first dynamic weight and the second dynamic weight.

[0089] Specifically, a risk function can be constructed using formulas, and health status can be determined by risk values, thus converting health scores into risk factors.

[0090] The first risk factor corresponding to the j-th indicator It can be represented as: ; in, To adjust the parameters.

[0091] The second risk factor corresponding to the j-th indicator It can be represented as: ; To measure the risk warning effectiveness of any alliance of participants, this invention defines a characteristic function using the following formula: ; That is, for any alliance, its characteristic function value simplifies to the sum of the risk values ​​of all participants in the alliance at that moment. The Shapley value is used to fairly quantify the contribution of each participant. Shapley value The average of its marginal contribution to all possible alliances: ; First dynamic weight It can be represented as: ; Second dynamic weight It can be represented as: ; The j-th indicator's single health index S j It can be represented as: ; By dynamically assigning weights based on the real-time health status of the equipment, the robustness and sensitivity of the evaluation system are ensured, effectively solving the problem that the traditional fixed-weight method cannot simultaneously take into account all aspects, and achieving true dynamic integration.

[0092] The formula for calculating the PCHI (Positive Health Index) is as follows: ; Where n is the total number of indicators contained in a single health assessment indicator set. Let j be the static weight of the j-th indicator. Let j be a single health index for the j-th indicator.

[0093] This invention effectively achieves dynamic adaptive adjustment of the assessment model by constructing a risk function to calculate dynamic weights and fusing dual-benchmark health scores. It retains the sensitivity of the mean method to early deterioration while taking into account the robustness of the threshold method to sudden anomalies, thus reducing false alarms and missed alarms.

[0094] Existing pump health monitoring systems mostly rely on fixed operating cycles or static thresholds for comparative alarms. Their indicator weights are static values ​​and cannot adaptively adjust to changes in equipment operating status. This method struggles to effectively capture slow performance degradation trends caused by wear and tear, is insensitive to early signs of failure, and is prone to missed alarms. Furthermore, it may generate false alarms due to normal operating fluctuations or measurement noise. In addition, traditional methods struggle to effectively integrate and quantify multi-source monitoring data, resulting in ambiguous decision-making criteria for the comprehensive health index and failing to provide accurate and adaptive quantitative support for predictive maintenance.

[0095] The present invention has the following significant advantages over the prior art: First, this invention improves the ability to identify early, progressive degradation of equipment. Traditional health assessment methods often rely on instantaneous alarms when benchmark values ​​are exceeded. This approach is prone to false alarms or equipment damage due to late risk detection, representing a typical reactive alarm mode. This invention introduces a historical average baseline assessment method to compare the current state of the equipment with its historical health baseline. More importantly, it innovatively proposes a risk game fusion mechanism that can dynamically identify risky faults. When equipment exhibits early degradation such as slow wear or efficiency decline but before exceeding a threshold, the score under the average method will decrease first, indicating an increase in risk value. This causes the dynamic fusion weight to automatically increase, and the comprehensive health index sensitively reflects this slow but continuous deterioration trend.

[0096] Secondly, this invention effectively solves the problems of robustness and sensitivity in the assessment results, reducing false alarms and missed alarms. While mean-based methods are stable, they are slow to respond to sudden anomalies; while threshold-based methods are sensitive, they are prone to false alarms due to normal operating condition fluctuations or measurement noise. This invention takes the TOPSIS method as an example, using a dynamic fusion of dual-benchmark quantification and risk game theory to analyze the risk intensity under both methods in real time. This dynamic weight allocation mechanism makes the fluctuations in the final comprehensive health index assessment curve more reliable.

[0097] Finally, this invention forms a complete health status assessment system. The output comprehensive health index transforms complex, multi-dimensional monitoring data into a single health level with clear physical meaning, which helps extend the service life of large water pump units and reduce the overall life-cycle operation and maintenance costs.

[0098] The centrifugal pump health assessment device provided by the present invention is described below. The centrifugal pump health assessment device described below can be referred to in correspondence with the centrifugal pump health assessment method described above.

[0099] like Figure 2 As shown, the present invention also provides a centrifugal pump health assessment device, comprising: The standardization module is used to standardize and reduce the dimensionality of multiple health assessment index sets for centrifugal pumps, resulting in multiple standardized index sets. The first determination module is used to determine the static weights of each indicator based on each standardized indicator set. The second determining module is used to determine the first health score of the indicator based on the historical average baseline corresponding to the indicator, and to determine the second health score of the indicator based on the safety threshold boundary corresponding to the indicator. The dynamic fusion module is used to assign dynamic weights to the first health score and the second health score by defining risk factors, and to perform weighted fusion of the first health score and the second health score according to the dynamic weights to obtain a single health index of the indicator. The static fusion module is used to perform weighted fusion of the individual health indices of each indicator according to the static weights corresponding to the standardized indicator set, so as to obtain the comprehensive health index of the centrifugal pump. The grading module is used to determine the health status level of the centrifugal pump based on the comprehensive health index.

[0100] Figure 3 A schematic diagram of the physical structure of an electronic device is provided. This device may include a processor, a communications interface, memory, and a communication bus, wherein the processor, communications interface, and memory communicate with each other via the communication bus. The processor can invoke logical instructions from the memory to execute a centrifugal pump health assessment method.

[0101] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0102] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to perform the centrifugal pump health assessment method provided by the above methods.

[0103] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the centrifugal pump health assessment methods provided by the methods described above.

[0104] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0105] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0106] 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for assessing the health of a centrifugal pump, characterized in that, include: Multiple health assessment index sets for centrifugal pumps are standardized and dimensionality reduced to obtain multiple standardized index sets. The static weights of each indicator are determined based on the various standardized indicator sets. The first health score of the indicator is determined based on the historical mean baseline corresponding to the indicator, and the second health score of the indicator is determined based on the safety threshold boundary corresponding to the indicator. By defining risk factors to assign dynamic weights to the first health score and the second health score, and by weighting and fusing the first health score and the second health score according to the dynamic weights, a single health index of the indicator is obtained. The single health index of each indicator is weighted and fused according to the static weights corresponding to the standardized indicator set to obtain the comprehensive health index of the centrifugal pump. The health status level of the centrifugal pump is determined based on the comprehensive health index.

2. The centrifugal pump health assessment method according to claim 1, characterized in that, The set of health assessment indicators includes, but is not limited to, rotational speed, head, flow rate, pump outlet pipe pressure, pump inlet pipe pressure, peak-to-peak value of spindle swing, effective value of vibration velocity at the drive end, effective value of vibration velocity at the non-drive end, effective value of vibration displacement at the drive end, and effective value of vibration displacement at the non-drive end.

3. The centrifugal pump health assessment method according to claim 1, characterized in that, The health assessment indicator set includes flow rate and rotation speed, and the standardized indicator set includes flow rate coefficient; The dimensionality reduction of the health assessment indicator set includes: The flow coefficient is determined based on the flow rate and the rotational speed.

4. The centrifugal pump health assessment method according to claim 1, characterized in that, The health assessment index set includes head and speed, and the standardized index set includes head coefficient; The dimensionality reduction of the health assessment indicator set includes: The head coefficient is determined based on the head and the rotational speed.

5. The centrifugal pump health assessment method according to claim 1, characterized in that, The standardization process is performed on multiple sets of health assessment indicators, including: Determine the deviations between the first type of indicators and their corresponding theoretical values ​​in each of the aforementioned health assessment indicator sets; The deviations of each indicator are normalized based on the maximum and minimum values ​​of the indicator deviations, thereby completing the standardization of the first type of indicators.

6. The centrifugal pump health assessment method according to claim 1, characterized in that, The standardization process is performed on multiple sets of health assessment indicators, including: Determine the maximum and minimum values ​​of each second-category indicator in each set of health assessment indicators; The second category of indicators is normalized based on the maximum and minimum values ​​of the second category of indicators, thus completing the standardization of the second category of indicators.

7. The centrifugal pump health assessment method according to claim 1, characterized in that, The static weights of each indicator are determined based on the various standardized indicator sets, including: The static weights of each standardized index set are obtained by calculating the weights using the entropy weight method or the Critic method.

8. The centrifugal pump health assessment method according to claim 1, characterized in that, The historical average baseline includes a first positive ideal solution and a first negative ideal solution, and the safety threshold boundary includes a second positive ideal solution and a second negative ideal solution; Under the historical average baseline, the average value of each indicator of the equipment during the historical healthy operation phase is selected as the first positive ideal solution, and the boundary value that deviates from the first positive ideal solution to a certain extent is set as the first negative ideal solution; Under the safety threshold boundary, for hydraulic performance indicators, a very narrow allowable interval near the optimal curve is selected as the second positive ideal solution, and a limit line of ±20% deviation from the optimal curve is selected as the second negative ideal solution. For mechanical stability indicators, the second positive ideal solution is set to 0, and the value specified in relevant national standards or engineering designs is used as the second negative ideal solution. The first health score of the indicator is determined based on the historical mean baseline corresponding to the indicator, and the second health score is determined based on the safety threshold boundary corresponding to the indicator, including: The first Euclidean distance between the target index in the standardized index set and the corresponding first positive ideal solution, the second Euclidean distance between the target index and the corresponding first negative ideal solution, the third Euclidean distance between the target index and the corresponding second positive ideal solution, and the fourth Euclidean distance between the target index and the corresponding second negative ideal solution are determined respectively. The first health score of the target indicator is determined based on the first Euclidean distance and the second Euclidean distance, and the second health score of the target indicator is determined based on the third Euclidean distance and the fourth Euclidean distance.

9. The centrifugal pump health assessment method according to claim 1, characterized in that, By defining risk factors and assigning dynamic weights to the first and second health scores, and then weighting and fusing the first and second health scores according to the dynamic weights, a single health index is obtained, including: The first risk factor corresponding to the first health score and the second risk factor corresponding to the second health score are determined according to a preset exponential function relationship; wherein, the risk factor is used to characterize the significance of the abnormal risk of the corresponding health score, and the closer the health score is to the abnormal limit, the larger the corresponding risk factor is. A risk game model is constructed based on the first risk factor and the second risk factor. The first risk factor is divided by the sum of the first risk factor and the second risk factor to obtain the first dynamic weight. Divide the second risk factor by the sum of the first risk factor and the second risk factor to obtain the second dynamic weight; The first health score and the second health score are weighted and summed using the first dynamic weight and the second dynamic weight to obtain the single health index.

10. The centrifugal pump health assessment method according to claim 1, characterized in that, The health status level of the centrifugal pump is determined based on the comprehensive health index, including: The health status level of the centrifugal pump is determined based on the technical indicators of the comprehensive health index, risk assessment requirements, and long-term operating experience.

Images