Cloud native multi-terminal intercommunication power plant industrial equipment management method and system

Abstract

The invention belongs to the technical field of smart power plants and data medium stations, and provides a cloud native multi-terminal intercommunication power plant industrial equipment management method and system, and the method specifically comprises the steps: enabling a smart power plant scene to comprise a data medium station and a data application terminal, and enabling the data medium station to be provided with a load monitor; a lagging feature group is obtained in real time through a load monitor; calculating the load pollution risk of each data service according to the lagging feature group; and finally, in combination with the load pollution risk of each middle station service, judging the middle station load abnormity of the base data. The data feedback quality of each intermediate station service in the base data intermediate stations in the smart power plant scene is monitored in real time, so that the risk of low reading effectiveness of the base data intermediate stations in the process of providing the intermediate station service for the data application end is effectively quantified, and the problem of data pollution in the process of acquiring the data service by the data application end is avoided; the stability of data service is improved, and meanwhile, the risk of data pollution to a data application end is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart power plants and data middle platforms, and specifically relates to a cloud-native multi-terminal intercommunication power plant industrial equipment management method and system. Background Art

[0002] With the rapid development of information technology, power plants are gradually transforming and upgrading to intelligent smart power plants. Smart power plants integrate the base data middle platform and the data application end, which includes intelligent monitoring, data collection, power generation management or energy efficiency management. The base data middle platform plays a centralized data service role in the smart power plant, responsible for providing the required data to the data application end to realize data sharing and integration of the smart power plant.

[0003] In the actual application of smart power plants, the data application end obtains data services by calling the standardized interface of the data middle platform. In addition to the traditional basic functions of reading and storage, the data service also has middle platform functions such as interception, isolation and preprocessing. The risk of source data leakage is greatly reduced by using the data middle platform as the middle platform function. However, in the process of the data application end obtaining the middle platform function service, data pollution problems often occur on the data application end. One of the reasons for this problem is the widespread incompatibility of the collaborative working mode of multi-terminal intercommunication. The data formats and interface design standards adopted by different systems are inherently different, which makes data services often face incompatibility risks. As a result, there is a risk of low reading utility in the call data obtained by the data application end, that is, the timeliness of the feedback data does not meet the data request, which will cause application data pollution on the application end and undermine the accuracy of the application end processing program execution results.

[0004] At present, people's focus on data service acquisition is often concentrated on the interface adjustment process of data transmission, that is, whether data flow can be smoothly realized through standardized interfaces. However, in actual application, interface problems are not the only source of the risk of low read utility. Another important reason is that there are defects in the data processing results fed back by the base data platform in the power plant. These defects are often due to the fact that the data processing of the power plant requires extremely high real-time feedback. However, due to the cascade delay problem when multiple terminals work together, the feedback data has a lag phenomenon. This lag phenomenon has a strong correlation with the CPU occupancy rate and RAM occupancy rate in the server. When the service volume of individual middle-end services increases sharply, this lag phenomenon will also affect other middle-end services running on the same server. Data with lag belongs to non-request data in data applications with real-time requirements. The sending and receiving of non-request data causes the risk of low read utility on the data application side. In the process of data transmission, the defective data service is transmitted to the data application side, resulting in the problem of data quality pollution. Therefore, there is an urgent need for a cloud-native multi-terminal intercommunication power plant industrial equipment management method and system. Summary of the invention

[0005] The purpose of the present invention is to propose a cloud-native multi-terminal intercommunication power plant industrial equipment management method and system to solve one or more technical problems existing in the prior art and at least provide a beneficial choice or create conditions.

[0006] In order to achieve the above object, according to one aspect of the present invention, a cloud-native multi-terminal intercommunication power plant industrial equipment management method is provided, and the method comprises the following steps:

[0007] The smart power plant scenario includes a data middle station and a data application end. The data middle station is configured with a load monitor. The lag feature group is obtained in real time through the load monitor. The load contamination risk of each middle station service is calculated based on the lag feature group. The load anomaly of the base data middle station is determined based on the load contamination risk of each middle station service.

[0008] Furthermore, in the smart power plant scenario, which includes a data middle platform and a data application end, the method for configuring a load monitor for the data middle platform is: the data middle platform provides several types of data services, and the data application end obtains data services by calling the standardized interface of the data middle platform. In the data service, non-direct read data requests are defined as middle platform services; the load monitor is a load monitoring tool at the software level on the data middle platform.

[0009] Furthermore, a method for obtaining a hysteresis feature group in real time through a load monitor is as follows: for a data request process of a data application terminal under any middle-office service, the data that needs to be obtained from the middle-office service in the data request of the data application terminal is preset with a timestamp range, denoted as rq.tmk, and the timestamp range of the data actually sent by the middle-office service is denoted as st.tmk. The load monitor monitors the two: when st.tmk∈rq.tmk defines that the time domain of the data request result is normal, otherwise the time domain of the data request result overflows; the amount of data that has time domain overflow is denoted as the overflow data amount; and a hysteresis feature group is constructed according to the data request that has time domain overflow.

[0010] Furthermore, the method of constructing the hysteresis feature group according to the data request of the time domain overflow is as follows: setting the data collection interval as DF, DF∈[60,180] seconds, and forming a data order value once per interval time DF;

[0011] In the DF period, the percentage of the number of data requests with time domain overflow to the total number of data requests is recorded as the overflow occurrence ratio, and the accumulated value of the overflow data volume is recorded as the overflow cumulative value;

[0012] The tuple consisting of the overflow occurrence ratio and the overflow cumulative value is recorded as the lag feature group.

[0013] Furthermore, the method for calculating the load contamination risk of each middle station service based on the hysteresis feature group is:

[0014] The overflow occurrence threshold Ovth is preset to 0.5, and each time a hysteresis feature group is obtained is recorded as a measurement point;

[0015] For any middle-end service, the overflow ratio is recorded as pto. When pto>Ovth, the corresponding measurement point is recorded as a strong overflow point. It is defined that each measurement point between any strong overflow point and the first strong overflow point in the reverse time direction constitutes an overflow interval. The difference between any measurement point and the corresponding overflow cumulative value of the farthest measurement point in its overflow interval is the overflow differential value Ofd.

[0016] The principle of calculating the overflow differential value is that the overflow differential value is a quantitative measure of the delay in processing data requests, which characterizes the performance fluctuations and load conditions of the data service when processing requests, and is an important indicator reflecting the data processing efficiency and delay characteristics of the middle office service in a specific time period; the formation of the overflow differential value belongs to the quantification of the difference between the overflow cumulative value between any measurement point and the farthest measurement point in its overflow interval. This difference quantification makes full use of the timestamp monitoring angle of data requests, deeply explores the response characteristics of the system under high load conditions, and comprehensively constructs a dynamic monitoring model with the overflow differential value as the core by combining the overlap of data requests, processing delays and system load. It can effectively capture the fluctuation trend in the data processing process, making it sensitive to the time domain changes of data requests. This sensitivity makes the model more adaptable in a high-concurrency environment.

[0017] In any overflow interval, obtain the overflow cumulative value corresponding to the measurement point with the smallest overflow occurrence ratio and record it as the risk reference; calculate the root mean square value of the risk reference of all middle-office services under this measurement point and record it as the reference level;

[0018] The ratio of the cumulative value of the overflow cumulative values ​​in any overflow interval to the number of elements in each overflow cumulative value of the overflow interval that are greater than the corresponding reference level is the near-risk level Neri;

[0019] The calculation method is: Where k1 is the cumulative variable, nR is the number of elements in the overflow interval, OnR is the number of elements in the overflow interval whose cumulative overflow value is greater than the reference level, Ocv k1 is the overflow cumulative value of the k1th measuring point in the overflow interval corresponding to the measuring point;

[0020] The pollution risk s.Losc of each sub-load is calculated based on the overflow interval:

[0021]

[0022] Where k2 is the cumulative variable, nR is the number of elements in the overflow interval, FnR is the number of negative overflow difference values ​​in the overflow interval, and mid.Ofd is the median value of the overflow difference values ​​in the overflow interval.

[0023] The ratio of the sub-load contamination risk of the overflow interval to the number of elements in the overflow interval is taken as the load gradient, and the ratio of the load gradient of the current overflow interval to the average value of the load gradients of each overflow interval except the current overflow interval is taken as the load contamination risk;

[0024] Since the calculation of load contamination risk is obtained by graded processing based on the overflow intervals of each measuring point, the risk of low reading utility in the call data obtained by the application end of the application scenario data in the base data can be effectively quantified. However, due to the excessive reliance on the distribution of overflow occurrence ratios in the screening process of strong overflow points, the sensitivity of the overflow cumulative value in data analysis is reduced, and the problem of insufficient accuracy of load contamination risk arises, especially in time periods when strong overflow points occur more frequently. This problem is more prominent. However, the existing technology cannot effectively compensate for this sensitivity decline. In order to eliminate this influence, the present invention proposes a more preferred solution as follows.

[0025] Preferably, the method for calculating the load contamination risk of each middle station service according to the hysteresis feature group is: each time when the hysteresis feature group is obtained is recorded as a measurement point; for any middle station service, the median value of the overflow occurrence ratio of all measurement points is recorded as the overflow threshold; if the overflow occurrence ratio of a measurement point is greater than the overflow threshold and the overflow occurrence ratio of the measurement point is greater than the previous moment, then the measurement point is recorded as a marginal measurement point, otherwise it is a content measurement point;

[0026] The overflow accumulation values ​​at all measuring points are normalized to form a cumulative normalized value. The product of the overflow occurrence ratio at any measuring point and the cumulative normalized value is the hysteresis modulus.

[0027] The minimum value of the overflow cumulative values ​​of all marginal measurement points is taken as the marginal reference value, and the overflow critical condition of the measurement point is that the overflow cumulative value is greater than the marginal reference value; if there is a content measurement point under the current middle platform service that meets the overflow critical condition, then the average value of the overflow cumulative value of each content measurement point is calculated and recorded as the overflow critical threshold; if there is no content measurement point that meets the overflow critical condition under any middle platform service, then the overflow critical threshold is defined as the upper quartile of the overflow cumulative value of the content measurement point of the middle platform service;

[0028] Define the corresponding measuring point whose overflow cumulative value is greater than the overflow critical threshold as the first critical point; if a first critical point has a larger overflow cumulative value than both the previous and next first critical points, it is defined as the second critical point; each measuring point between any second critical point and the first second critical point in the reverse time direction constitutes a marginal interval;

[0029] The principle of calculating the overflow critical threshold is to utilize the changing trend of the overflow cumulative value, combined with the marginal reference value and quartile method, to form an effective threshold setting mechanism, which aims to search for the performance limit of the system under high load; if there are content measurement points that meet the overflow critical conditions, the process of calculating the critical threshold reveals the average performance of the system under the current load conditions, which can effectively reflect the carrying capacity of the system under high load conditions; and the marginal reference value sets a lower limit for the evaluation of the overflow critical threshold, so that it can tolerate the lowest load state; when the overflow critical conditions are not met, the upper quartile is used to replace the threshold, which ensures the rationality and adaptability of the threshold, so that the system can perform dynamic and autonomous adjustments.

[0030] The marginal subinterval is formed by any measuring point to the end measuring point of its corresponding marginal interval. If a measuring point has a larger overflow occurrence ratio and overflow cumulative value than the previous measuring point, then such a measuring point is defined as a sub-overflow measuring point, and the proportion of sub-overflow measuring points in the marginal subinterval is the central magnification ratio Rgin; the derivative coefficient φ of the marginal interval is calculated according to the overflow occurrence ratio of the first critical point: φ=ln(∑ j2＝1 Rgin j2 ×Spo j2 +1); where j2 is the cumulative variable, Spo j2 and Rgin j2 are the overflow occurrence ratio and center amplification ratio of the j2th first critical point in the marginal interval, respectively, and ln() is a logarithmic function with the natural number e as the base;

[0031] The calculation of the central amplification ratio is based on the dynamic monitoring of the sub-overflow measurement points in the marginal sub-interval. The identification mechanism of the sub-overflow measurement points in the marginal sub-interval ensures that the central amplification ratio only reflects abnormal data requests under high load conditions, so that the central amplification ratio has good dynamic response capabilities and can reflect the changes of the system under different load conditions in real time; an increase in the central amplification ratio indicates that the system is facing greater pressure, thereby causing performance degradation or delay, so it is used as an early warning mechanism indicator to effectively support the efficient operation of the system in a complex environment.

[0032] Define the measurement point corresponding to the maximum value of all hysteresis moduli as the reference measurement point; the number of the first critical points between any marginal interval and the reference measurement point is used as the reference weight Div;

[0033] The current load contamination risk Losc is calculated based on the derivative coefficient of the marginal interval and the benchmark weight:

[0034]

[0035] Where j1 is the cumulative variable, e is the natural constant, CT j1 and CB j1are the hysteresis characteristic groups with the maximum and minimum hysteresis modulus in the j1th marginal interval, Div j1 and SuDc j1 are the reference weight and cumulative mean of the j1th marginal interval respectively. The cumulative mean is the average of the overflow cumulative values ​​in the marginal interval; min.SuDc represents the minimum value of all overflow cumulative values; GRLdx() is the slope function, and its return value is the slope between two calling tuples.

[0036] Furthermore, the method of judging the abnormal load of the base data middle platform in combination with the load contamination risk of each middle platform service is: if the load contamination risk of a moment is greater than that of the previous moment, it is defined that the middle platform service has a progressive risk at that moment; if the z-score corresponding to the maximum value of each load contamination risk at the same moment exceeds 3, it is considered that there is a burst call at that moment;

[0037] The average value of the pollution risk of each load at the same time is recorded as the risk set value at that time. From the current time, search in reverse time for the time when the first risk set value appears to be the minimum. Each time between that time and the current time is recorded as a reference point. The ratio of the number of middle-office services with progressive risk at any reference point to the total number of middle-office services at that reference point is recorded as the progressive occurrence degree.

[0038] If the progressive occurrence degree at the current moment is higher than 50%, and no burst call occurs at each reference point at the current moment, it is determined that the load on the base data center is abnormal, and there is a risk of low read efficiency caused by server performance.

[0039] Preferably, when it is determined that there is a risk of low read utility caused by server performance, server resources are added to the base data center, and the server resources include one or more of server nodes, CPU, GPU, RAM memory, and bandwidth.

[0040] Preferably, all undefined variables in the present invention, if not clearly defined, can be manually set thresholds.

[0041] The present invention also provides a cloud-native multi-terminal intercommunication power plant industrial equipment management system, the cloud-native multi-terminal intercommunication power plant industrial equipment management system comprises: a processor, a memory, and a computer program stored in the memory and executable on the processor, the processor implements the steps in the cloud-native multi-terminal intercommunication power plant industrial equipment management method when executing the computer program, the cloud-native multi-terminal intercommunication power plant industrial equipment management system can be run on computing devices such as desktop computers, laptop computers, PDAs, and cloud data centers, and the executable system may include, but is not limited to, a processor, a memory, and a server cluster, and the processor executes the computer program to run in the following system units:

[0042] A scenario construction unit is used to include a data middle platform and a data application end in a smart power plant scenario, and the data middle platform is configured with a load monitor;

[0043] A data acquisition unit, used for acquiring a hysteresis characteristic group in real time through a load monitor;

[0044] A load contamination risk construction unit, used to calculate the load contamination risk of each data service according to the lag feature group;

[0045] The data base resource monitoring unit is used to determine the abnormal load of the base data middle platform based on the load contamination risk of each middle platform service.

[0046] The beneficial effects of the present invention are as follows: by real-time monitoring of the data feedback quality of each middle platform service in the base data middle platform in the smart power plant scenario, the risk of low reading utility that occurs in the process of the base data middle platform providing middle platform services to the data application end is effectively quantified, and the timeliness of the data obtained by each data application end from the data base is identified abnormally, and then the server resource configuration strategy of the data middle platform is warned. Through the final warning and adjustment, data pollution problems that occur in the process of the data application end obtaining data services are avoided. This not only effectively guarantees that the data middle platform reasonably allocates the configuration of server resources in the process of providing data services, improves the stability of data services, and ensures the normal execution of each business process or function in the smart power plant that is highly sensitive to the timeliness of business data, but also effectively improves the integration efficiency of the data middle platform and the data application end in the smart power plant scenario and ensures the efficient operation of the entire smart power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The above and other features of the present invention will become more obvious by describing in detail the embodiments shown in the accompanying drawings. The same reference numerals in the accompanying drawings of the present invention represent the same or similar elements. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other accompanying drawings can be obtained based on these accompanying drawings without creative work. In the accompanying drawings:

[0048] Figure 1 Shown is a flow chart of a cloud-native multi-terminal intercommunication method for industrial equipment management in power plants;

[0049] Figure 2 Shown is a structural diagram of a cloud-native multi-terminal interoperable power plant industrial equipment management system. DETAILED DESCRIPTION

[0050] The following will be combined with the embodiments and drawings to clearly and completely describe the concept, specific structure and technical effects of the present invention, so as to fully understand the purpose, scheme and effect of the present invention. It should be noted that the embodiments and features in the embodiments of this application can be combined with each other without conflict.

[0051] like Figure 1 The following is a flowchart of a cloud-native multi-terminal interoperable power plant industrial equipment management method. Figure 1 A cloud-native multi-terminal intercommunication power plant industrial equipment management method according to an embodiment of the present invention is described, and the method includes the following steps:

[0052] The smart power plant scenario includes a data middle station and a data application end. The data middle station is configured with a load monitor. The lag feature group is obtained in real time through the load monitor. The load contamination risk of each middle station service is calculated based on the lag feature group. The load anomaly of the base data middle station is determined based on the load contamination risk of each middle station service.

[0053] Furthermore, in the smart power plant scenario, which includes a data middle platform and a data application end, the method for configuring a load monitor for the data middle platform is: the data middle platform provides several types of data services, and the data application end obtains data services by calling the standardized interface of the data middle platform. In the data service, non-direct read data requests are defined as middle platform services; the load monitor is a load monitoring tool at the software level on the data middle platform.

[0054] The data middle platform runs several middle platform services at the same time. Each middle platform service generates data transmission during execution. The direction of data transmission is from the data middle platform to the corresponding data application end. The load monitor is configured in the data middle platform. The load monitor is a software-level load monitoring tool that is responsible for monitoring the data transmission information between the data middle platform and the data application end. The load monitor options include Prometheus, Grafana or New Relic. The data middle platform needs to rely on server resources to provide middle platform services. Server resources include CPU, bandwidth, and RAM memory provided by several servers.

[0055] The data middle platform refers to the base data middle platform, which plays a centralized data service role in the smart power plant and is responsible for collecting, processing and storing data from other data application terminals. The data application terminal is responsible for providing specific functions and operations, including intelligent monitoring and data collection applications, power generation management applications, energy efficiency management applications, etc. The standardized interfaces usually provided by the data middle platform include RESTful API or GraphQL API. The data application terminal can obtain data services through these interfaces. The data services include the basic functions of reading and storage, as well as middle platform functions such as interception, isolation and preprocessing. The data obtained by the data application terminal through the standardized interface and the data transmitted by the data middle platform through the standardized interface have the same transmission speed, and the data content read after format conversion is the same.

[0056] Furthermore, a method for obtaining a hysteresis feature group in real time through a load monitor is as follows: for a data request process of a data application terminal under any middle-office service, the data that needs to be obtained from the middle-office service in the data request of the data application terminal is preset with a timestamp range, denoted as rq.tmk, and the timestamp range of the data actually sent by the middle-office service is denoted as st.tmk. The load monitor monitors the two: when st.tmk∈rq.tmk defines that the time domain of the data request result is normal, otherwise the time domain of the data request result overflows; the amount of data that has time domain overflow is denoted as the overflow data amount; and a hysteresis feature group is constructed according to the data request that has time domain overflow.

[0057] Each data request returns several records from the base data center. Each record has a timestamp. The range of timestamps corresponding to all records under this data request is the timestamp range of the data actually sent by the center service.

[0058] Furthermore, the method of constructing the hysteresis feature group according to the data request of the time domain overflow is as follows: setting the data collection interval as DF, DF∈[60,180] seconds, and forming a data order value once per interval time DF;

[0059] In the DF period, the percentage of the number of data requests with time domain overflow to the total number of data requests is recorded as the overflow occurrence ratio, and the accumulated value of the overflow data volume is recorded as the overflow cumulative value;

[0060] The tuple consisting of the overflow occurrence ratio and the overflow cumulative value is recorded as the lag feature group.

[0061] There are several preset middle-office services in the data middle-office, and each middle-office service performs different data processing functions, including interception, isolation and preprocessing methods in different preset data application environments, as well as data conversion methods based on data models. Each data application end often needs to make data requests for the same middle-office service in the data middle-office. These data requests have overlapping work tasks, but due to data isolation and different request parameters, it is impossible to directly intercept and use the processed data at the timestamp, and server resources are preset and allocated for each middle-office service. Therefore, the more data requests occur at the same time, the greater the probability of delayed data.

[0062] Furthermore, the method for calculating the load contamination risk of each middle station service based on the hysteresis feature group is:

[0063] Limit the measurement point range to the 24-72 hours before the current time;

[0064] The overflow occurrence threshold Ovth is preset to 0.5, and each time a hysteresis feature group is obtained is recorded as a measurement point;

[0065] For any middle-end service, the overflow occurrence ratio is recorded as pto. When pto>Ovth, the corresponding measurement point is recorded as a strong overflow point. It is defined that each measurement point between any strong overflow point and the first strong overflow point in the reverse time direction constitutes an overflow interval. The difference between any measurement point and the corresponding overflow cumulative value of the farthest measurement point in its overflow interval is the overflow difference value Ofd. If the number of elements in the overflow interval is less than or equal to 30, it will be removed. The current moment is the strong overflow point by default. If the time interval between the current moment and the strong overflow point searched in reverse time is less than 30 minutes, the searched strong overflow point will be ignored in the subsequent procedures.

[0066] Preferably, the preset overflow threshold Ovth has a first constraint condition that the number of strong overflow points is less than 20% of the total, and a second constraint condition that Ovth ≥ 0.5; when the search obtains a value that satisfies both the first constraint condition and the second constraint condition, the minimum value therebetween is used as the overflow threshold, otherwise the overflow threshold Ovth is set to 0.5.

[0067] The farthest measuring point in the overflow interval refers to the measuring point farthest from the current moment.

[0068] In any overflow interval, obtain the overflow cumulative value corresponding to the measurement point with the smallest overflow occurrence ratio and record it as the risk reference; calculate the root mean square value of the risk reference of all middle-office services under this measurement point and record it as the reference level;

[0069] The ratio of the cumulative value of the overflow cumulative values ​​in any overflow interval to the number of elements in each overflow cumulative value of the overflow interval that are greater than the corresponding reference level is the near-risk level Neri;

[0070] The calculation method is: Where k1 is the cumulative variable, nR is the number of elements in the overflow interval, OnR is the number of elements in the overflow interval whose cumulative overflow value is greater than the reference level, Ocv k1 is the overflow cumulative value of the k1th measuring point in the overflow interval corresponding to the measuring point;

[0071] The pollution risk s.Losc of each sub-load is calculated based on the overflow interval:

[0072]

[0073] Where k2 is the cumulative variable, nR is the number of elements in the overflow interval, FnR is the number of negative overflow difference values ​​in the overflow interval, and mid.Ofd is the median value of the overflow difference values ​​in the overflow interval; pto k2 is the overflow ratio of the k2th measuring point in the overflow interval, Ofd k2 is the overflow difference value of the k2th measuring point in the overflow interval, and ln() is the logarithmic function with the natural number e as the base.

[0074] The ratio of the sub-load contamination risk of the overflow interval to the number of elements in the overflow interval is taken as the load gradient, and the ratio of the load gradient of the current overflow interval to the average value of the load gradients of each overflow interval except the current overflow interval is taken as the load contamination risk;

[0075] The current measurement point refers to the closest measurement point at the current moment that has an overflow interval;

[0076] Preferably, the method for calculating the load contamination risk of each middle station service according to the hysteresis feature group is: limit the measurement point range to within 6-12 hours before the current time; each time the hysteresis feature group is obtained is recorded as a measurement point; for any middle station service, the median value of the overflow occurrence ratio of all measurement points is recorded as the overflow threshold; if the overflow occurrence ratio of a measurement point is greater than the overflow threshold and the overflow occurrence ratio of the measurement point is greater than the previous moment, then the measurement point is recorded as a marginal measurement point, otherwise it is a content measurement point;

[0077] The overflow accumulation values ​​at all measuring points are normalized to form a cumulative normalized value. The product of the overflow occurrence ratio at any measuring point and the cumulative normalized value is the hysteresis modulus.

[0078] The minimum value of the overflow cumulative values ​​of all marginal measurement points is taken as the marginal reference value, and the overflow critical condition of the measurement point is that the overflow cumulative value is greater than the marginal reference value; if there is a content measurement point under the current middle platform service that meets the overflow critical condition, then the average value of the overflow cumulative value of each content measurement point is calculated and recorded as the overflow critical threshold; if there is no content measurement point that meets the overflow critical condition under any middle platform service, then the overflow critical threshold is defined as the upper quartile of the overflow cumulative value of the content measurement point of the middle platform service;

[0079] Define the corresponding measuring point whose overflow cumulative value is greater than the overflow critical threshold as the first critical point; if a first critical point has a larger overflow cumulative value than both the previous and next first critical points, it is defined as the second critical point; each measuring point between any second critical point and the first second critical point in the reverse time direction constitutes a marginal interval;

[0080] The marginal subinterval is formed by any measuring point to the end measuring point of its corresponding marginal interval. If a measuring point has a larger overflow occurrence ratio and overflow cumulative value than the previous measuring point, then such a measuring point is defined as a sub-overflow measuring point, and the proportion of sub-overflow measuring points in the marginal subinterval is the central magnification ratio Rgin; the derivative coefficient φ of the marginal interval is calculated according to the overflow occurrence ratio of the first critical point: φ=ln(∑ j2＝1 Rgin j2 ×Spo j2 +1); where j2 is the cumulative variable, Spo j2 and Rgin j2 are the overflow occurrence ratio and center amplification ratio of the j2th first critical point in the marginal interval, respectively, and ln() is a logarithmic function with the natural number e as the base;

[0081] The end point of the marginal interval is the point in the marginal interval that is farthest from the current moment;

[0082] Define the measurement point corresponding to the maximum value of all hysteresis moduli as the reference measurement point; the number of the first critical points between any marginal interval and the reference measurement point is used as the reference weight Div; the measurement points between a marginal interval and the reference measurement point do not include the measurement points of the marginal interval itself;

[0083] The current load contamination risk Losc is calculated based on the derivative coefficient of the marginal interval and the benchmark weight:

[0084]

[0085] Where j1 is the cumulative variable, e is the natural constant, CT j1 and CB j1 are the hysteresis characteristic groups with the maximum and minimum hysteresis modulus in the j1th marginal interval, Div j1 and SuDc j1 are the reference weight and cumulative mean of the j1th marginal interval respectively. The cumulative mean is the average of the overflow cumulative values ​​in the marginal interval; min.SuDc represents the minimum value of all overflow cumulative values; GRLdx() is the slope function, and its return value is the slope between two calling tuples.

[0086] Furthermore, the method of judging the abnormal load of the base data middle platform in combination with the load contamination risk of each middle platform service is: if the load contamination risk of a moment is greater than that of the previous moment, it is defined that the middle platform service has a progressive risk at that moment; if the z-score corresponding to the maximum value of each load contamination risk at the same moment exceeds 3, it is considered that there is a burst call at that moment;

[0087] The average value of the pollution risk of each load at the same time is recorded as the risk set value at that time. From the current time, search in reverse time for the time when the first risk set value appears to be the minimum. Each time between that time and the current time is recorded as a reference point. The ratio of the number of middle-office services with progressive risk at any reference point to the total number of middle-office services at that reference point is recorded as the progressive occurrence degree.

[0088] If the progressive occurrence degree at the current moment is higher than 50%, and no burst call occurs at each reference point at the current moment, it is determined that the load on the base data center is abnormal, and there is a risk of low read efficiency caused by server performance.

[0089] Preferably, when it is determined that there is a risk of low read utility caused by server performance, server resources are added to the base data center, and the server resources include one or more of server nodes, CPU, GPU, RAM memory, and bandwidth.

[0090] An embodiment of the present invention provides a cloud-native multi-terminal interoperable power plant industrial equipment management system, such as Figure 2 What is shown is a structural diagram of a cloud-native multi-terminal intercommunication power plant industrial equipment management system of the present invention. A cloud-native multi-terminal intercommunication power plant industrial equipment management system of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the above-mentioned cloud-native multi-terminal intercommunication power plant industrial equipment management method embodiment.

[0091] The system comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to run in the following units of the system:

[0092] A scenario construction unit is used to include a data middle platform and a data application end in a smart power plant scenario, and the data middle platform is configured with a load monitor;

[0093] A data acquisition unit, used for acquiring a hysteresis characteristic group in real time through a load monitor;

[0094] A load contamination risk construction unit, used to calculate the load contamination risk of each data service according to the lag feature group;

[0095] The data base resource monitoring unit is used to determine the abnormal load of the base data middle platform based on the load contamination risk of each middle platform service.

[0096] The cloud-native multi-terminal intercommunication power plant industrial equipment management system can be run on computing devices such as desktop computers, laptops, PDAs and cloud servers. The cloud-native multi-terminal intercommunication power plant industrial equipment management system, the executable system may include, but is not limited to, processors, memories. Those skilled in the art will understand that the example is only an example of a cloud-native multi-terminal intercommunication power plant industrial equipment management system, and does not constitute a limitation on a cloud-native multi-terminal intercommunication power plant industrial equipment management system, which may include more or fewer components than the example, or a combination of certain components, or different components, for example, the cloud-native multi-terminal intercommunication power plant industrial equipment management system may also include input and output devices, network access devices, buses, etc.

[0097] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the cloud-native multi-terminal intercommunication power plant industrial equipment management system operation system, and uses various interfaces and lines to connect the various parts of the entire cloud-native multi-terminal intercommunication power plant industrial equipment management system operation system.

[0098] The memory can be used to store the computer program and / or module, and the processor realizes the various functions of the cloud-native multi-terminal intercommunication power plant industrial equipment management system by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the storage data area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0099] Although the description of the present invention has been quite detailed and has been described in particular with respect to several described embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, so as to effectively cover the intended scope of the present invention. In addition, the present invention is described above with the embodiments foreseeable by the inventors, and its purpose is to provide a useful description, and those non-substantial changes to the present invention that are not currently foreseen may still represent equivalent changes of the present invention.

Claims

1. A cloud-native multi-terminal intercommunication power plant industrial equipment management method, characterized in that: The method comprises the following steps: in a smart power plant scenario, a data middle station and a data application end are included, and a load monitor is configured in the data middle station; a hysteresis feature group is obtained in real time through the load monitor; the load contamination risk of each middle station service is calculated according to the hysteresis feature group; and the load abnormality of the base data middle station is determined in combination with the load contamination risk of each middle station service; The method for calculating the load contamination risk of each middle-office service based on the lagging feature group is: identifying the overflow occurrence ratio and the overflow cumulative value from the lagging feature group, identifying the strong overflow points according to the overflow occurrence ratio, using the strong overflow points to form an overflow interval, and calculating the near-risk level through each overflow cumulative value in the overflow interval, and obtaining each sub-load contamination risk based on the near-risk level, using the sub-load contamination risk to calculate the load gradient, and quantifying the load contamination risk by the load gradient.

2. According to claim 1, a cloud-native multi-terminal intercommunication power plant industrial equipment management method is characterized in that: In the smart power plant scenario, there are data middle platform and data application end. The method of configuring load monitor in data middle platform is as follows: data middle platform provides several types of data services, and data application end obtains data services by calling standardized interfaces of data middle platform. In data services, data requests that are not directly read are defined as middle platform services. The load monitor is a load monitoring tool at the software level on the data center.

3. According to claim 1, a cloud-native multi-terminal intercommunication power plant industrial equipment management method is characterized in that: The method of obtaining the lagging feature group in real time through the load monitor is: for the data request process of the next data application end of any middle-office service, the data that needs to be obtained from the middle-office service in the data request of the data application end is preset with a timestamp range, denoted as rq.tmk, and the timestamp range of the data actually sent by the middle-office service is denoted as st.tmk. The load monitor monitors the two: when st.tmk∈rq.tmk defines that the time domain of the data request result is normal, otherwise the time domain of the data request result overflows; the amount of data with time domain overflow is denoted as the overflow data amount; and the lagging feature group is constructed according to the data request with time domain overflow.

4. According to claim 3, a cloud-native multi-terminal intercommunication power plant industrial equipment management method is characterized in that: The method of constructing a hysteresis feature group based on data requests with time domain overflow is as follows: set the data collection interval to DF, DF∈[60,180] seconds, and form a data order value every time interval DF; in the DF period, the percentage of data requests with time domain overflow to the total data request volume is recorded as the overflow occurrence ratio, and the accumulated value of the overflow data volume is recorded as the overflow cumulative value; the tuple consisting of the overflow occurrence ratio and the overflow cumulative value is recorded as the hysteresis feature group.

5. According to claim 1, a cloud-native multi-terminal intercommunication power plant industrial equipment management method is characterized in that: The method for calculating the load contamination risk of each middle station service based on the hysteresis feature group is: preset the overflow threshold Ovth as 0.5, and record each time when the hysteresis feature group is obtained as a measurement point; for any middle station service, the overflow occurrence ratio is recorded as pto, and when pto>Ovth, the corresponding measurement point is recorded as a strong overflow point; define each measurement point between any strong overflow point and its first strong overflow point in the reverse time direction as an overflow interval, and the difference between any measurement point and the corresponding overflow cumulative value of the farthest measurement point in its overflow interval is the overflow difference value Ofd; In any overflow interval, obtain the overflow cumulative value corresponding to the measurement point with the smallest overflow occurrence ratio and record it as the risk reference; calculate the root mean square value of the risk reference of all middle-office services under this measurement point and record it as the reference level; The ratio of the cumulative value of the overflow cumulative value in any overflow interval to the number of elements in each overflow cumulative value of the overflow interval that are greater than the corresponding reference level is the near-risk level Neri. Where k1 is the cumulative variable, nR is the number of elements in the overflow interval, OnR is the number of elements in the overflow interval whose cumulative overflow value is greater than the reference level; the pollution risk s.Losc of each sub-load is calculated based on the overflow interval: Where k2 is the accumulated variable, nR is the number of elements in the overflow interval, FnR is the number of negative overflow difference values ​​in the overflow interval, and mid.Ofd is the median value of the overflow difference values ​​in the overflow interval. The ratio of the sub-load contamination risk of the overflow interval to the number of elements in the overflow interval is taken as the load gradient, and the ratio of the load gradient of the current overflow interval to the average value of the load gradients of each overflow interval except the current overflow interval is taken as the load contamination risk.

6. According to claim 1, a cloud-native multi-terminal intercommunication power plant industrial equipment management method is characterized in that: The method for calculating the load contamination risk of each middle-office service based on the lag feature group is as follows: each time a lag feature group is obtained is recorded as a measurement point; for any middle-office service, the median value of the overflow occurrence ratio of all measurement points is recorded as the overflow threshold; if the overflow occurrence ratio of a measurement point is greater than the overflow threshold and the overflow occurrence ratio of the measurement point is greater than that of the previous moment, then the measurement point is recorded as a marginal measurement point, otherwise it is a content measurement point; the overflow cumulative values ​​under all measurement points are normalized to form a cumulative normalized quantity, and the product of the overflow occurrence ratio of any measurement point and the cumulative normalized quantity is the lag modulus; the minimum value of the overflow cumulative values ​​of all marginal measurement points is taken as the marginal reference value, and the overflow critical condition of the measurement point is the overflow cumulative value. The product value is greater than the marginal reference value; if there are content measurement points under the current middle platform service that meet the overflow critical condition, then the average value of the overflow cumulative values ​​of each content measurement point is calculated and recorded as the overflow critical threshold; when there are no content measurement points that meet the overflow critical condition under any middle platform service, the overflow critical threshold is defined as the upper quartile of the overflow cumulative value of the content measurement point of the middle platform service; the corresponding measurement point whose overflow cumulative value is greater than the overflow critical threshold is defined as the first critical point; if a first critical point is larger than the overflow cumulative value of the previous and next first critical points, it is defined as the second critical point; the marginal interval is composed of each measurement point between any second critical point and the first second critical point in the reverse time direction; The marginal sub-interval is formed by any measuring point to the end measuring point of its corresponding marginal interval. If a measuring point has a larger overflow occurrence ratio and overflow cumulative value than the previous measuring point, such a measuring point is defined as a sub-overflow measuring point, and the proportion of sub-overflow measuring points in the marginal sub-interval is the central magnification ratio Rgin; the derivative coefficient of the marginal interval is calculated according to the overflow occurrence ratio of the first critical point; The measuring point corresponding to the maximum value of all hysteresis moduli is defined as the reference measuring point; the number of the first critical points between any marginal interval and the reference measuring point is used as the reference weight Div; the current load contamination risk is calculated based on the derivative coefficient of the marginal interval and the reference weight.

7. According to claim 1, a cloud-native multi-terminal intercommunication power plant industrial equipment management method is characterized in that: The method of judging the abnormal load of the base data middle platform in combination with the load contamination risk of each middle platform service is: if the load contamination risk of a moment is greater than that of the previous moment, it is defined that the risk of the middle platform service at that moment is progressive; if the z-score corresponding to the maximum value of each load contamination risk at the same moment exceeds 3, it is considered that there is a burst call at that moment; The average value of the pollution risks of each load at the same moment is recorded as the risk set value at that moment. Starting from the current moment, the moment when the first risk set value appears to be a minimum is searched in reverse time direction. Each moment between that moment and the current moment is recorded as a reference point. The ratio of the number of middle-office services that experience risk progression at any reference point to the total number of middle-office services at that reference point is recorded as the progression occurrence degree. If the progression occurrence degree at the current moment is higher than 50%, and no sudden calls occur at each reference point at the current moment, it is determined that the load of the base data middle-office is abnormal, and there is a risk of low read utility induced by server performance.

8. A cloud-native multi-terminal intercommunication power plant industrial equipment management method according to claim 7, characterized in that: When it is determined that there is a risk of low read efficiency caused by server performance, server resources are added to the base data center. The server resources include one or more of server nodes, CPU, GPU, RAM memory, and bandwidth.

9. A cloud-native multi-terminal interoperable power plant industrial equipment management system, characterized in that: The cloud-native multi-terminal intercommunication power plant industrial equipment management system includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the cloud-native multi-terminal intercommunication power plant industrial equipment management method described in any one of claims 1 to 8 are implemented. The cloud-native multi-terminal intercommunication power plant industrial equipment management system runs on desktop computers, laptop computers, PDAs, and computing devices in cloud data centers.

Images