CN113656267B

CN113656267B - Device energy efficiency calculation method and device, electronic device and storage medium

Info

Publication number: CN113656267B
Application number: CN202110860028.2A
Authority: CN
Inventors: 易存道
Original assignee: Beijing Baolande Software Co ltd
Current assignee: Beijing Baolande Software Co ltd
Filing date: 2021-07-28
Publication date: 2024-05-31
Anticipated expiration: 2041-07-28

Abstract

The invention provides a device energy efficiency calculation method, a device, an electronic device and a storage medium, wherein the method comprises the following steps: acquiring data values of a plurality of running indexes of equipment to be predicted and data values of a plurality of asset attributes; based on the data values of a plurality of running indexes and the data values of a plurality of asset attributes, calculating to obtain the energy efficiency result of the equipment by using a pre-constructed equipment energy efficiency calculation model. The method can be used for determining the energy efficiency of the equipment in a multi-index and multi-dimensional manner and efficiently, has an accurate energy efficiency result, and can provide an effective reference basis for equipment management.

Description

Device energy efficiency calculation method and device, electronic device and storage medium

Technical Field

The invention relates to the technical field of informatization energy conservation, in particular to a method and a device for calculating equipment energy efficiency, electronic equipment and a storage medium.

Background

The evaluation of energy efficiency of various asset devices is gradually paid attention to by enterprises and related parties at the moment of advocating green and low carbon. The device energy efficiency is generally related to the device operation condition and the property of the device, wherein the device operation condition can be represented by a plurality of operation indexes. For example, if the equipment has good running condition (each running index is normal, etc.) and the equipment asset attribute is good (the equipment is newer or has no maintenance record, etc.), the energy efficiency of the equipment is normal; if the running condition of the equipment is poor (a few running indexes are abnormal, etc.) and the property of the equipment asset is poor (the equipment is older and has more maintenance records), the energy efficiency of the equipment is lower; if the device is not running directly or the device asset attributes are too old, the device energy efficiency is not valid.

The traditional management method for various computer devices is to store basic device management data in an EXCEL table file which is established manually, and then record and check the difference between the data in the data table and the actual device data by means of manual operation so as to find out the problems of the devices and determine the energy efficiency of the devices. But this approach has several drawbacks: 1. the data amount of the data table is numerous and complicated, the data updating is fast, the manual query operation is complex, time and labor are wasted, and the efficiency is low; 2. when the energy efficiency of the equipment is evaluated manually, the factors or the dimensions which can be considered are too small, and the evaluation integrity and the accuracy are poor.

There is also a device energy efficiency determination method: the method comprises the steps of collecting real-time data of operation indexes of equipment to be detected aiming at each operation index, correspondingly comparing the real-time data of the operation indexes with a preset threshold value of the operation indexes, judging that the operation indexes are abnormal if the real-time data exceeds the preset threshold value, and judging that the energy efficiency of the equipment is lower.

The method can only perform single judgment on the result of direct comparison between each operation index and the threshold value, and has low energy efficiency determination accuracy. Moreover, an operation index is merely an instantaneous representation of the performance of a device on the one hand, and the energy efficiency of the device is generally influenced by the combination of various operation indexes of the whole performance of the device on the plurality of aspects. For example, the CPU utilization rate of the device is particularly high in a period of time, and exceeds the CPU utilization threshold, and according to the method, the device is judged to be in a high-energy-consumption running state, and the energy efficiency of the device is low. This determination is obviously one-sided. Because, in this period, the CPU usage rate is particularly high, in fact, a large amount of service TPS enters in this period, so that both the CPU usage rate and the IO usage rate exceed their own thresholds, but after a period of time, the CPU usage rate and the IO usage rate can resume to be normal, so this situation belongs to the normal fluctuation of the device operation index when the device processes the service request. Obviously, the judgment result of the method in this case is erroneous.

Therefore, aiming at the problems of poor accuracy, less evaluation dimension, low efficiency and the like in the energy efficiency determination method in the prior art, a more effective solution is lacking at present.

Disclosure of Invention

The invention provides a method and a device for calculating energy efficiency of equipment, electronic equipment and a storage medium, which are used for overcoming the defects of poor accuracy, less evaluation dimension, low efficiency and the like in the prior art energy efficiency determination method, and can accurately and efficiently determine the energy efficiency of the equipment in multiple indexes and multiple dimensions.

The invention provides a method for calculating equipment energy efficiency, which comprises the following steps:

acquiring data values of a plurality of running indexes of equipment to be predicted and data values of a plurality of asset attributes;

based on the data values of a plurality of running indexes and the data values of a plurality of asset attributes, calculating to obtain the energy efficiency result of the equipment by using a pre-constructed equipment energy efficiency calculation model.

According to the method for calculating the energy efficiency of the equipment provided by the invention, the energy efficiency result of the equipment is calculated and obtained by applying a pre-constructed equipment energy efficiency calculation model based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes, and the method specifically comprises the following steps:

Based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes, executing calculation according to the calculation strategy of each level of energy efficiency calculation scene in the equipment energy efficiency calculation model, and respectively obtaining each level of energy efficiency result of the equipment;

Based on the data values of a plurality of running indexes and the data values of a plurality of asset attributes, executing calculation according to the calculation strategy of each secondary energy efficiency calculation scene in the equipment energy efficiency calculation model, and respectively obtaining each secondary energy efficiency result of the equipment;

the equipment energy efficiency calculation model aggregates the primary energy efficiency results and the secondary energy efficiency results according to a result aggregation strategy to obtain the energy efficiency results of the equipment;

And the method for calculating the energy efficiency of the equipment further comprises the step of determining the elimination level of the equipment by combining a preset elimination strategy when the energy efficiency result of the equipment is that the equipment is invalid or the equipment is low.

According to the method for calculating the equipment energy efficiency, which is provided by the invention, the construction process of the equipment energy efficiency calculation model comprises the following steps:

periodically collecting data values of a plurality of operation indexes and data values of a plurality of asset attributes of each device in a plurality of devices;

The data values of a plurality of operation indexes and the data values of a plurality of asset attributes of each device are respectively calculated according to a preset calculation strategy of each level of energy efficiency calculation scene, and each level of energy efficiency result of each device is obtained;

the data values of a plurality of operation indexes and the data values of a plurality of asset attributes of each device are respectively calculated according to a preset calculation strategy of each secondary energy efficiency calculation scene, and each secondary energy efficiency result of each device is obtained;

aggregating the primary energy efficiency results and the secondary energy efficiency results of each device according to a result aggregation strategy to obtain an energy efficiency calculation result of each device;

collecting each item of data in the above steps to be used as basic data of modeling;

And performing machine learning training by applying xgboost algorithm based on the basic data to construct the equipment energy efficiency calculation model.

According to the method for calculating the equipment energy efficiency provided by the invention, the first-level energy efficiency calculation scene comprises the following steps: at least one of a device maintenance scene, a device maintenance expiration scene, a device bearing analysis scene, a device elimination scene and a device product expiration scene;

And, the secondary energy efficiency computation scenario includes: gao Weibao, a high-fault scenario, a low-utilization scenario, and a high-energy scenario.

According to the method for calculating the energy efficiency of the equipment provided by the invention, the operation index comprises the following steps: at least one of CPU utilization rate, memory utilization rate, IO utilization rate, connection channel utilization rate and storage space utilization rate;

And/or, the asset attributes comprising: at least one of maintenance manufacturer, maintenance time and maintenance period, bearing assembly condition, manufacturer and equipment model, equipment on-line date and equipment effective period, maintenance expense, equipment type, fault alarm times and equipment use power.

According to the method for calculating the energy efficiency of the equipment, which is provided by the invention, the calculation strategy of the equipment maintenance scene comprises the following steps: judging whether the equipment meets the requirement of having a maintenance manufacturer or not based on the data value of the maintenance manufacturer in the asset attribute of the current equipment; if the first-level energy efficiency result of the equipment in the equipment maintenance scene is met, the equipment is low-efficiency; if not, the equipment is normal;

And/or, the computing strategy of the device maintenance expiration scene comprises: judging whether the equipment meets the requirement of exceeding the maintenance period limit or not based on the maintenance time and the data value of the maintenance period in the asset attribute of the current equipment; if the first-level energy efficiency result of the equipment in the equipment maintenance out-of-date scene is low-efficiency; if not, the equipment is normal;

And/or, the device bears a calculation strategy of the analysis scene, including: judging whether the equipment meets the condition of bearing the component or not based on the data value of the condition of bearing the component in the asset attribute of the current equipment; if yes, the primary energy efficiency result of the equipment in the equipment bearing analysis scene is that the equipment is invalid; if not, the equipment is normal;

And/or, the computing strategy of the device elimination scene comprises the following steps: judging whether the equipment is in the obsolete equipment catalogue or not based on the data values of manufacturers and equipment models in the asset attributes of the current equipment and the pre-acquired obsolete equipment catalogue; if the device is in the first-level energy efficiency result of the device in the device elimination scene is that the device is invalid; if not, the equipment is normal;

and/or, the computing strategy of the equipment product out-of-date scene comprises the following steps: judging whether the equipment meets the condition that the equipment service date exceeds the equipment effective period or not based on the data value of the equipment on-line date and the equipment effective period in the asset attribute of the current equipment; if the first-level energy efficiency result of the equipment in the equipment product out-of-date scene is low-efficiency; if not, the equipment is normal;

And/or, the computing strategy of Gao Weibao scenes comprises: judging whether the maintenance fee of the current equipment meets a preset reference value exceeding the maintenance fee of the equipment or judging whether the maintenance fee of the current equipment meets a preset multiple exceeding the average maintenance fee of each equipment under the same equipment type based on the data value of the maintenance fee in the asset attribute of the current equipment; if one of the two energy efficiency results meets the requirement, the second energy efficiency result of the device in the Gao Weibao scene is that the device is low-efficiency; if the two types of the information are not satisfied, the equipment is normal;

And/or, the calculation strategy of the high fault scene comprises the following steps: judging whether the average value of the fault alarming times of the current equipment meets a preset reference value exceeding the fault alarming times or whether the average value of the fault alarming times of the current equipment meets a preset multiple exceeding the average fault alarming times of all equipment in the same equipment type or not based on the data value of the fault alarming times in the asset attribute of the current equipment in a preset time period; if one of the two energy efficiency results meets the requirement, the secondary energy efficiency result of the equipment in the high-fault scene is that the equipment is invalid; if the two types of the information are not satisfied, the equipment is normal;

And/or, the calculation strategy of the low-utilization scene comprises: judging whether the data value of the running index of the current equipment meets a preset reference value lower than the running index or whether the data value of the running index of the current equipment meets a preset multiple lower than the average value of the running indexes of all the equipment in the same equipment type or not based on the data value of each running index of the current equipment in a preset time period; if all operation indexes of the current equipment meet one of the operation indexes, the secondary energy efficiency result of the equipment in the low-utilization-rate scene is equipment inefficiency; otherwise, the equipment is normal;

and/or, the calculation strategy of the high-energy consumption scene comprises the following steps: judging operation indexes: judging whether the data value of the running index of the current equipment meets a preset reference value lower than the running index or whether the data value of the running index of the current equipment meets a preset multiple lower than the average value of the running indexes of all the equipment in the same equipment type or not based on the data value of each running index of the current equipment in a preset time period; asset attribute judging step: judging whether the data value of the property of the current equipment meets a preset reference value exceeding the property of the property or judging whether the data value of the property of the current equipment meets a preset multiple exceeding the average value of the property of each equipment under the same equipment type based on the data value of each property of the current equipment; if all running indexes of the current equipment meet one of the running indexes or any asset attribute of the current equipment meets one of the running indexes, the secondary energy efficiency result of the equipment in the high-energy consumption scene is equipment inefficiency; otherwise, the device is normal.

According to the method for calculating the energy efficiency of the equipment, the result aggregation strategy comprises the following steps: analyzing each primary energy efficiency result and each secondary energy efficiency result of each device, and if at least one result is that the device is invalid, the energy efficiency result of the device is that the device is invalid; if no device is invalid but at least one result is that the device is inefficient, the energy efficiency result of the device is that the device is inefficient; if the results are all equipment normal, the energy efficiency result of the equipment is equipment normal.

The invention also provides a device energy efficiency calculating device, which comprises:

the data acquisition module is used for acquiring data values of a plurality of running indexes of the equipment to be predicted and data values of a plurality of asset attributes;

and the energy efficiency calculation module is used for calculating and obtaining an energy efficiency result of the equipment by applying a pre-constructed equipment energy efficiency calculation model based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes.

The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing all or part of the steps of a method of computing energy efficiency according to any one of the above claims when the computer program is executed.

The invention also provides a computer readable storage medium having stored thereon a computer program which when executed by a processor performs all or part of the steps of a method of calculating energy efficiency according to any of the above described devices.

The invention provides a method, a device, an electronic device and a storage medium for calculating the energy efficiency of equipment, wherein the method obtains the energy efficiency result of the equipment by obtaining the data values of a plurality of running indexes and the data values of a plurality of asset attributes of the equipment to be predicted and combining with a pre-built equipment energy efficiency calculation model, so that the energy efficiency of the equipment can be determined in a multi-index, multi-dimensional and efficient manner, the energy efficiency result is accurate, and an effective reference basis can be provided for equipment management.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings that are necessary for the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention and that other drawings can be obtained from these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of a method for calculating energy efficiency of a device according to the present invention;

FIG. 2 is a second flow chart of the method for calculating the energy efficiency of the device according to the present invention;

FIG. 3 is a third flow chart of the method for calculating the energy efficiency of the device according to the present invention;

FIG. 4 is a schematic flow chart of a process for constructing an energy efficiency calculation model of a device in the method for calculating energy efficiency of the device provided by the invention;

FIG. 5 is a second schematic flow chart of a process for constructing an energy efficiency calculation model of a device in the method for calculating energy efficiency of a device according to the present invention;

FIG. 6 is a schematic diagram of a computing device for device energy efficiency provided by the present invention;

Fig. 7 is a schematic structural diagram of an electronic device provided by the present invention.

Reference numerals:

610: a data acquisition module; 620: an energy efficiency calculation module; 710: a processor;

720: a communication interface; 730: a memory; 740: a communication bus.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be fully described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

The following describes a method, a device, an electronic device and a storage medium for calculating energy efficiency of a device according to the present invention with reference to fig. 1 to 7.

The invention provides a method for calculating energy efficiency of equipment, fig. 1 is one of flow diagrams of the method for calculating energy efficiency of equipment provided by the invention, as shown in fig. 1, the method comprises the following steps:

210. Acquiring data values of a plurality of running indexes of equipment to be predicted and data values of a plurality of asset attributes;

220. Based on the data values of a plurality of running indexes and the data values of a plurality of asset attributes, calculating to obtain the energy efficiency result of the equipment by using a pre-constructed equipment energy efficiency calculation model.

Modern enterprises, in order to expand their own business and cope with various new service demands, introduce a large number of asset devices, such as: a server, a network device, a storage device, an application device, a host device, and the like. Because of the complicated manual management, and the reasons of untimely management, elimination, and updating, a large number of inefficient or invalid devices may exist in the devices, which cannot well play a role in enterprise service, occupy space and management resources of enterprises, further analysis of the assets is urgently needed to find out the inefficient or invalid devices from the assets, and effectively classify, manage and eliminate updating.

The multiple devices of each device class are usually stored in the same AMDB system resource pool, and one device needing energy efficiency calculation is obtained from the AMDB system resource pool according to the service requirement to be predicted, namely one device under one device class is used as the device to be predicted. Of course, if there are multiple devices that need to perform energy efficiency calculation, one device is selected at a time to perform energy efficiency calculation, and the currently selected device is used as the device to be predicted, which is also referred to as the current device. And acquiring the data values of a plurality of running indexes of the current equipment and the data values of a plurality of asset attributes of the current equipment. The operation index comprises one or more of CPU utilization rate, memory utilization rate, IO utilization rate, connection channel utilization rate and storage space utilization rate, and the asset attribute comprises one or more of maintenance manufacturer, maintenance time and maintenance period, bearing component condition, manufacturer and equipment model, equipment on-line date and equipment validity period, maintenance expense, equipment type, fault alarm times and equipment use power. The adjustment may be specifically performed according to the type of device to which the current device belongs or according to the specific situation of the current device. And then, inputting the obtained data values of a plurality of running indexes of the current equipment and the obtained data values of a plurality of asset attributes of the current equipment into a pre-built equipment energy efficiency calculation model, and obtaining an energy efficiency result of the current equipment through calculation of the equipment energy efficiency calculation model. The energy efficiency result of the current device may include three situations, that is, the current device is invalid, that the current device is inefficient, or that the current device is normal. And directly judging whether the current equipment is low-efficiency or invalid equipment or not according to the energy efficiency condition of the current equipment indicated in the finally obtained energy efficiency result of the current equipment, so that relevant management personnel can manage the current equipment conveniently.

The invention provides a calculation method of equipment energy efficiency, which is characterized in that the method obtains the energy efficiency result of equipment by obtaining the data values of a plurality of running indexes and the data values of a plurality of asset attributes of the equipment to be predicted and combining with a pre-built equipment energy efficiency calculation model, so that the energy efficiency of the equipment can be determined in a multi-index, multi-dimensional and efficient manner, the energy efficiency result is accurate, and an effective reference basis can be provided for equipment management.

According to the method for calculating the energy efficiency of the device provided by the present invention, fig. 2 is a second flow chart of the method for calculating the energy efficiency of the device provided by the present invention, as shown in fig. 2, in the step 220, based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes, an energy efficiency result of the device is calculated and obtained by applying a pre-built device energy efficiency calculation model, which specifically includes:

221. Based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes, executing calculation according to the calculation strategy of each level of energy efficiency calculation scene in the equipment energy efficiency calculation model, and respectively obtaining each level of energy efficiency result of the equipment;

222. based on the data values of a plurality of running indexes and the data values of a plurality of asset attributes, executing calculation according to the calculation strategy of each secondary energy efficiency calculation scene in the equipment energy efficiency calculation model, and respectively obtaining each secondary energy efficiency result of the equipment;

223. and the equipment energy efficiency calculation model aggregates the primary energy efficiency results and the secondary energy efficiency results according to a result aggregation strategy to obtain the energy efficiency results of the equipment.

After the obtained data values of a plurality of running indexes of the current equipment and the obtained data values of a plurality of asset attributes of the current equipment are input into a pre-built equipment energy efficiency calculation model, the energy efficiency calculation model can be divided into a simpler primary energy efficiency calculation scene and a more complex secondary energy efficiency calculation scene, each scene is an energy efficiency calculation scene preset according to historical experience data, and each scene corresponds to an independent calculation strategy. And determining which energy efficiency calculation scenes are suitable for the current equipment according to the obtained data values of the running indexes of the current equipment and the obtained data values of the asset attributes of the current equipment, wherein the running indexes are specific, the asset attributes are specific, and correspondingly executing the calculation of one or more energy efficiency calculation scenes according to the actual calculation requirements. Of course, only the calculation of the primary energy efficiency calculation scene may be performed, or both the calculation of the primary energy efficiency calculation scene and the calculation of the secondary energy efficiency calculation scene may be performed. When the current device performs calculation of multiple energy efficiency calculation scenes, the results obtained by each energy efficiency calculation scene need to be combined correspondingly, for example, aggregation of the results is performed according to a preset result aggregation policy.

Specifically, firstly, based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes, executing calculation according to the calculation strategy of each level of energy efficiency calculation scene in the equipment energy efficiency calculation model, and respectively obtaining each level of energy efficiency result of the equipment; based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes, executing calculation according to the calculation strategy of each secondary energy efficiency calculation scene in the equipment energy efficiency calculation model, and respectively obtaining each secondary energy efficiency result of the equipment; and finally, the equipment energy efficiency calculation model aggregates the primary energy efficiency results and the secondary energy efficiency results according to a result aggregation strategy to obtain the comprehensive energy efficiency result of the current equipment.

Each level of energy efficiency calculation scene preset in the model comprises: one or more of a device maintenance scene, a device maintenance expiration scene, a device bearing analysis scene, a device elimination scene and a device product expiration scene. The preset secondary energy efficiency calculation scene in the model comprises the following steps: gao Weibao, a high-fault scenario, a low-utilization scenario, and a high-energy scenario. The secondary energy efficiency calculation scene is often judged by the energy efficiency of a plurality of asset attributes and/or angles of a plurality of operation indexes, and the judgment result is more accurate. And, for each energy efficiency calculation scene and specific calculation strategies thereof, reference can be made to the detailed description of the subsequent model construction process.

And wherein the result aggregation policy comprises: analyzing each primary energy efficiency result and each secondary energy efficiency result of each device, and if at least one result is that the device is invalid, the energy efficiency result of the device is that the device is invalid; if no device is invalid but at least one result is that the device is inefficient, the energy efficiency result of the device is that the device is inefficient; if the results are all equipment normal, the energy efficiency result of the equipment is equipment normal.

The energy efficiency calculation scenes of different levels are preset in the equipment energy efficiency calculation model, the calculation of the simple primary energy efficiency calculation scene is respectively carried out according to the equipment conditions, or the calculation of the secondary energy efficiency calculation scene is carried out in a combined mode, and finally the energy efficiency results of the calculation of the energy efficiency calculation scenes are combined and aggregated to obtain the total energy efficiency evaluation of the current equipment, so that various defects of single angle and large evaluation energy efficiency result error in the manual evaluation of the energy efficiency can be effectively avoided.

And the method for calculating the energy efficiency of the equipment further comprises the step of determining the elimination level of the equipment by combining a preset elimination strategy when the energy efficiency result of the equipment is that the equipment is invalid or the equipment is low. Fig. 3 is a third flow chart of the method for calculating the energy efficiency of the device according to the present invention, as shown in fig. 3, i.e. on the basis of the embodiment shown in fig. 2, the method further includes the following steps:

230. And when the energy efficiency result of the equipment is that the equipment is invalid or the equipment is low, determining the elimination level of the equipment by combining a preset elimination strategy.

And when the energy efficiency result of the current equipment finally determined in the step 223 is that the equipment is invalid or the equipment is low, determining the elimination level of the current equipment according to the elimination strategy. Specifically, according to an elimination strategy comprising three items of high elimination level, medium elimination level and low elimination level, determining the elimination level to which the current equipment belongs, and marking the elimination level to the current equipment in a label form for viewing and reference by related management staff.

In the elimination strategy, corresponding elimination conditions are respectively set for each elimination level, and specific contents of the elimination conditions are elimination limiting ranges of all the asset attributes and all the operation indexes of the equipment, for example, the elimination limiting ranges of the asset attributes such as manufacturers, attribution scenes, maintenance expenses, service life, fault warning times and the like, or the elimination threshold limiting ranges of all the operation indexes and the like can be set.

Comparing the property and operation index of the current equipment with the corresponding elimination conditions of the elimination level from high to low, if the elimination conditions of the high elimination level are met, the elimination level of the current equipment is the high elimination level, otherwise, the current equipment is continuously compared with the corresponding elimination conditions of the medium elimination level, if the elimination conditions of the medium elimination level are met, the elimination level of the current equipment is the medium elimination level, and the current equipment is the medium elimination level, so that the current equipment is not required to be eliminated if all the three elimination conditions of the medium elimination level are not met, and the current equipment is regarded as 0.

According to the method for calculating the energy efficiency of any device provided by the present invention, that is, on the basis of the embodiment shown in any one of fig. 1 to 3, fig. 4 is one of flow diagrams of a process for constructing an energy efficiency calculation model of a device in the method for calculating the energy efficiency of a device provided by the present invention, and as shown in fig. 4, the process for constructing an energy efficiency calculation model of a device specifically includes:

110. periodically collecting data values of a plurality of operation indexes and data values of a plurality of asset attributes of each device in a plurality of devices;

120. the data values of a plurality of operation indexes and the data values of a plurality of asset attributes of each device are respectively calculated according to a preset calculation strategy of each level of energy efficiency calculation scene, and each level of energy efficiency result of each device is obtained;

130. the data values of a plurality of operation indexes and the data values of a plurality of asset attributes of each device are respectively calculated according to a preset calculation strategy of each secondary energy efficiency calculation scene, and each secondary energy efficiency result of each device is obtained;

140. Aggregating the primary energy efficiency results and the secondary energy efficiency results of each device according to a result aggregation strategy to obtain an energy efficiency calculation result of each device;

150. collecting each item of data in the above steps to be used as basic data of modeling;

160. And performing machine learning training by applying xgboost algorithm based on the basic data to construct the equipment energy efficiency calculation model.

In order to increase the accuracy of the constructed device energy efficiency calculation model, a model basic data collection stage, a model training stage and a model testing stage are required to be controlled respectively.

Model base data collection phase:

a plurality of data values for a plurality of operational metrics for a plurality of devices and a plurality of data values for a plurality of asset attributes for the same device are collected. Wherein, each device is collected one by one, or a plurality of devices are collected in a crossing way.

After a large amount of data of a large number of devices are collected, the collected data are subjected to data aggregation according to a large amount of data values of different time sequences of each operation index of each device, and the modeling efficiency can be effectively improved in the preprocessing process according to data aggregation of different time dimensions such as a day dimension, a Zhou Weidu dimension and a month dimension.

The specific data aggregation algorithm is to sum the original values of the data values of each operation index of each of the plurality of devices in a period of time, and take the maximum value as a new data value after summation. And then adding new data value each time, and carrying out data aggregation again, wherein each time of data aggregation, the maximum value of each original value of all data values at this time is compared with the maximum value taken by the previous aggregation, and if the maximum value is smaller than the maximum value taken by the previous aggregation, the maximum value taken by the previous aggregation is still taken; and taking the maximum value of the original values of all the data values of the aggregation at the time under the other conditions. The maximum value is taken as the new data value of the operation index, so that the maximum utilization rate of the operation index of the current equipment in the period of time can be accurately reflected, and the maximum utilization rate is used for being compared with the set reference value of the operation index and the like in the calculation process of the equipment energy efficiency model, so that the energy efficiency of the equipment can be accurately estimated.

And meanwhile, counting the number of data values of the running index, summing up the original values, dividing the sum by the number to obtain a corresponding data average value of the running index for later use, and even combining the equipment type, recording the average value of each data value of the running index in the period of time of each equipment in the equipment type to which the current equipment belongs for later use.

Of course, the data can be stored in a relational database, so that the management and the calling are convenient.

The method comprises the steps of defining a plurality of energy efficiency calculation scenes in advance, wherein two types of energy efficiency calculation scenes are mainly used, and one type of energy efficiency calculation scenes is a simple scene, such as a device maintenance scene, a device maintenance exceeding scene, a device bearing analysis scene, a device elimination scene, a device product exceeding scene and the like; another class is slightly complex, such as Gao Weibao scenes, high-fault scenes, low-utilization scenes, and high-energy-consumption scenes. The secondary energy efficiency calculation scene is often judged by the energy efficiency of a plurality of asset attributes and/or angles of a plurality of operation indexes, and the judgment result is more accurate. And each energy efficiency calculation scene has an independent calculation strategy, so that the energy efficiency of the equipment is judged according to different standards or angles. Of course, for each energy efficiency calculation scenario and its specific calculation strategy, reference may be made to the detailed description of the subsequent model construction process.

And respectively executing calculation according to a preset calculation strategy of each level of energy efficiency calculation scene to obtain each level of energy efficiency result of each device. And then the data values of a plurality of operation indexes and the data values of a plurality of asset attributes of each device are respectively calculated according to the preset calculation strategy of each secondary energy efficiency calculation scene, and each secondary energy efficiency result of each device is obtained. And aggregating the primary energy efficiency results and the secondary energy efficiency results of each device according to a result aggregation strategy to obtain the energy efficiency results of each device. This is performed for each device, whereby the energy efficiency results for each device are determined.

Of course, the step of verifying and confirming by the enterprise user can be added, namely, the comprehensive energy efficiency result of each device is displayed to the enterprise user, and whether the energy efficiency result is accurately evaluated or not is finally confirmed by the enterprise user. The energy efficiency result approved by the enterprise user mark is an accurate energy efficiency result, and the energy efficiency result not approved by the enterprise user mark is an inaccurate energy efficiency result. Therefore, the steps of enterprise user confirmation and screening are added to the energy efficiency results of the devices, and only relevant data of the energy efficiency results can be screened out for constructing the device energy efficiency calculation model. And each device corresponding to each energy efficiency result which is not approved by the enterprise user needs to be replaced into the resource pool, so that the energy efficiency result is re-determined later.

And collecting various data in the steps, including a large amount of data values of a plurality of running indexes and data values of a plurality of asset attributes of each device, various primary energy efficiency calculation scenes, calculation strategies and corresponding primary energy efficiency results, various secondary energy efficiency calculation scenes, calculation strategies and corresponding secondary energy efficiency results, comprehensive energy efficiency results of each device and the like, which are used as basic data for modeling. It may also be understood that the data of each device and each collected operation index before determining the energy efficiency, the data of each asset attribute, the data of each operation index or the reference value/threshold value of each asset attribute, and the like, as well as the intermediate energy efficiency result, each energy efficiency calculation scene, calculation strategy and final energy efficiency result, are all collected and used as the modeling basic data together.

And after all data in the basic data are serialized into json character strings, the json character strings are input into an AI model through RESTfulApi programmable programs, and a higher-order xgboost algorithm is applied to perform machine learning training by combining a cross verification method and a parameter grid search method so as to construct and optimize a required equipment energy efficiency calculation model.

The secondary energy efficiency calculation scene is often judged by the energy efficiency of a plurality of asset attributes and/or angles of a plurality of operation indexes, and the judgment result is more accurate. And each energy efficiency calculation scene has an independent calculation strategy, so that the energy efficiency of the equipment is judged according to different standards or angles.

According to the method for calculating the energy efficiency of the device provided by the invention, before the step 160, the machine learning training is performed by applying xgboost algorithm based on the basic data to construct the energy efficiency calculation model of the device, the method further comprises the step of performing data conversion processing on the basic data, and specifically comprises the following steps:

Determining the data values of a plurality of operation indexes of each device in the basic data and the Chinese fields in the data values of a plurality of asset attributes;

And carrying out data conversion processing on each Chinese field based on a single-heat coding characteristic extraction method so as to convert each Chinese field into a plurality of variable values corresponding to each Chinese field.

That is, between the step 150 and the step 160, a step of performing data conversion processing on the basic data is further included, which specifically includes the following steps:

there may be a plurality of data types such as a string type, a chinese field type, etc. after the preliminary acquisition of the basic data for modeling. And the underlying data may also have missing or outliers. Therefore, the data filling or exception removing preprocessing is required to be performed on the basic data, and even the unified conversion processing is required to be performed on the data types of the basic data, for example, the Chinese field type data can be converted into the character string type data.

(1) Preprocessing the basic data to fill data or remove anomalies:

(1-1) analyzing the missing values in each field of the basic data in a summarized manner, and filling the missing values in each field with 0 based on conventional business knowledge of the operation and maintenance management of the equipment. (1-2) summarizing the outliers in each piece of data of the analysis base data, and carrying out averaging treatment on more than 99% of the quantile data by a mean method so as to remove the outliers.

The acquired basic data involves a large amount of Chinese field type data such as "equipment manufacturer", "equipment holder", etc. While using machine learning algorithms generally requires that all input data must be numerical. Therefore, the unified conversion processing of the data types needs to be performed on the basic data, which mainly refers to converting all the data of the chinese field type in the basic data into the data of the character string type.

(2) And adopting a single-heat coding characteristic extraction method (OneHot method for short) to perform data conversion treatment:

(2-1) introduction of OneHot method:

The OneHot method belongs to a method of 'dummy variable' conversion to convert a variable into multiple columns. Taking the "manufacturer" as an example, according to its data value (the data value may be Hua Cheng, cisco and Langchao), the data value is respectively converted into three variables of "whether the manufacturer is Hua Cheng", "whether the manufacturer is Cisco", "whether the manufacturer is Langchao", that is, it will be converted according to each unique value of the data value, how many unique values are converted into how many variables, and the data value of the "manufacturer" has three unique values of Hua Cheng, cisco and Langchao, so that it is finally converted into three variables. Similarly, for example: the manufacturer has 1000 different data values, wherein 10 unique values exist, and then 10 variables exist after conversion, and one variable occupies one column. However, some operation indexes or asset attributes have more unique values of data values, so that variables are more after the operation indexes or asset attributes are converted by OneHot method, and therefore, the characteristics are sparse. Under the condition, summarizing and counting all unique values of the data value, selecting TOP-N unique values, and performing data conversion processing on the data value according to the TOP-N unique values to obtain N variables. For example, N is 10, when the unique value of the data value has 100 values, the first 10 unique values in the 100 values are selected as conversion targets, the data value is subjected to data conversion processing, 10 variables are obtained, and the variables after conversion of the other 90 data values are all recorded as "others". See in particular table 1 below.

TABLE 1

Manufacturer(s)	Whether the manufacturer is	Whether the manufacturer is Cisco	Whether the manufacturer is wave
				Hua Cheng Ji (Chinese character)	1	0	0
Hua Cheng Ji (Chinese character)	1	0	0
				Cisco (Cisco)	0	1	0
Langchao tide	0	0	1
				......	......	......	......

(2-2) Determining the data of all Chinese fields in the data values of a plurality of operation indexes and the data values of a plurality of asset attributes of each device in the basic data, and then performing data conversion processing on the data based on the OneHot method in (2-1) so as to convert each Chinese field into a plurality of corresponding variable values.

(2-3) Further performing feature derivation processing on the basic data based on conventional business knowledge of device operation and maintenance management, for example, taking "IOThreshold" and "IOResult" data as examples, and deriving features of whether the current device IO usage is greater than the IO usage reference value based on the conventional business knowledge of device operation and maintenance management.

The basic data processed by the steps (1) and (2) are used as new basic data for constructing an energy efficiency calculation model of the equipment.

According to the method for calculating the energy efficiency of the device provided by the present invention, fig. 5 is a second flowchart of a process for constructing a model for calculating the energy efficiency of the device in the method for calculating the energy efficiency of the device provided by the present invention, as shown in fig. 5, on the basis of the embodiment shown in fig. 4, step 160, performing machine learning training based on the basic data and using xgboost algorithm to construct the model for calculating the energy efficiency of the device, specifically includes the following steps:

161. Dividing the basic data into K sub-samples based on a K-fold cross validation method;

162. Performing machine learning training on a training data set formed by the K-1 sub-samples by using xgboost algorithm, and constructing an equipment energy efficiency calculation model;

163. Verifying the equipment energy efficiency calculation model based on a verification data set formed by 1 sub-sample;

164. And repeatedly executing the cross training and verification for K times to obtain a final equipment energy efficiency calculation model.

I.e. the model training process in turn comprises in particular the steps 161-164, in particular as follows. The basic idea of the cross-validation method is to divide the original data (dataset) into groups in a sense, one group as training dataset (track set) and the other group as validation dataset (validation set or test set). The AI classifier is first learned and trained using a training dataset, and then a model obtained from a previous training is validated and tested using a validation dataset to serve as a performance index for evaluating the classifier, and the model is optimized accordingly. In this embodiment, a K-fold cross-validation method (K-fold cross-validation) is used for cross-validation, specifically, 10-fold cross-validation is used for cross-validation, and the raw data in this embodiment refers to basic data for model training.

The base data of the initial sample is divided into K sub-samples, i.e. into 10 sub-samples. Wherein each sub-sample may be obtained by random segmentation.

And reserving 1 independent sub-sample as data of a subsequent verification model, forming training data sets by other K-1 sub-samples, particularly other 9 sub-samples, performing machine learning training on the training data sets by using a high-order xgboost algorithm, and constructing a preliminary equipment energy efficiency calculation model.

And then verifying and testing the preliminary equipment energy efficiency calculation model based on a verification data set formed by the 1 sub-samples reserved in advance, and adjusting xgboost algorithm parameters and the like according to verification and test results to optimize the model.

And, 10 times of cross training and verification are repeatedly performed, that is, 10 times of cross training and verification are performed, 1 different independent sub-sample is selected as a verification data set each time, the rest 9 sub-samples are used as training data sets, the training is performed, the verification is performed again after the training, … … times of cross training and verification are performed, that is, each sub-sample is verified once, and the result of 10 times is averaged, so that a final optimized equipment energy efficiency calculation model is obtained.

The K-fold cross validation method has the advantages that different randomly generated subsamples are repeatedly used for training and validation respectively, and each result can be validated once, so that more optimized and accurate results are obtained.

It should be noted that, when the 10-fold cross-validation is adopted for cross-validation for cross-training and validation in this embodiment, a parameter grid search method may be introduced in the model training process to adjust xgboost algorithm parameters.

The description of the parameter grid search method is as follows:

The parameter grid search method refers to that each parameter possibility is tried through cyclic traversal in the selection of all candidate parameters, and finally the best-performing parameter is taken as a final result. In order to improve learning efficiency of xgboost algorithm, in general, a xgboost algorithm max_depth parameter and a learning_rate parameter are selected for key adjustment and optimization, and search spaces of the two parameters are shown in table 2 below.

TABLE 2

Parameter name	Search space
		max_depth	3、4、5
learning_rate	0.001、0.01、0.02、0.05

A number of operational metrics for each device, comprising: any one or more of CPU usage, memory usage, IO usage, connection channel usage, storage space usage, and the like. And/or, a number of asset attributes for each device, including: any one or more of maintenance manufacturer, maintenance time and maintenance period, bearing component condition, manufacturer and equipment model, equipment number, equipment on-line date and equipment validity period, maintenance expense, equipment type, equipment major class, equipment minor class, equipment id, equipment ip, fault alarm times, equipment use power, equipment holder and the like.

And/or, the device bears a calculation strategy of the analysis scene, including: judging whether the equipment meets the requirement of bearing the component or not based on the data value of the condition of bearing the component in the asset attribute of the current equipment, wherein the component comprises one or more of a sub-component, an instance and an application; if yes, the primary energy efficiency result of the equipment in the equipment bearing analysis scene is that the equipment is invalid; if not, the equipment is normal;

Wherein, each operation index meeting one of the two conditions belongs to the same relationship, and each operation index must meet one of the two conditions to judge that the equipment is low-efficiency, otherwise, the equipment is normal. Such as: and observing two operation indexes of the CPU utilization rate and the memory utilization rate of the equipment A. The data value of the CPU utilization rate is selected as a new data value by the maximum value after data aggregation, and the data value of the memory utilization rate is also selected as a new data value by the maximum value after data aggregation. And judging that the new data value of the CPU utilization rate of the equipment A is lower than the preset reference value of the operation index, but the memory utilization rate of the equipment A is the new data value but higher than the threshold value. At this time, if the data value of each operation index is not lower than the preset reference value, the energy efficiency result of the equipment A in the scene is that the equipment A is normal.

And the operation index judging step and the asset attribute judging step have a relationship of or, and if only one step judges that the equipment is low-efficiency, the comprehensive result is that the equipment is low-efficiency.

According to the method for calculating the equipment energy efficiency provided by the invention, for each result obtained by each primary energy efficiency calculation scene and each secondary energy efficiency calculation scene, the result aggregation strategy comprises the following steps: analyzing each primary energy efficiency result and each secondary energy efficiency result of each device, and if at least one result is that the device is invalid, the energy efficiency result of the device is that the device is invalid; if no device is invalid but at least one result is that the device is inefficient, the energy efficiency result of the device is that the device is inefficient; if the results are all equipment normal, the energy efficiency result of the equipment is equipment normal.

The following describes a device for calculating energy efficiency of an apparatus, where the device for calculating energy efficiency may be understood as a device for executing a method for calculating energy efficiency of the apparatus, and the application principles of the device and the device are the same, and may be referred to each other and not described herein.

The invention also provides a device energy efficiency calculating device, which comprises: a data acquisition module 610, and an energy efficiency calculation module 620, wherein,

The data obtaining module 610 is configured to obtain data values of a plurality of operation indexes and data values of a plurality of asset attributes of the device to be predicted;

The energy efficiency calculation module 620 is configured to calculate and obtain an energy efficiency result of the device by applying a pre-built device energy efficiency calculation model based on the data values of the operation indexes and the data values of the asset attributes.

The invention provides a device for calculating the energy efficiency of equipment, which comprises a data acquisition module 610 and an energy efficiency calculation module 620, wherein the modules work cooperatively, so that the device can determine the energy efficiency of the equipment in a multi-index and multi-dimensional manner and efficiently, the energy efficiency result is accurate, and an effective reference basis can be provided for equipment management.

The present invention also provides an electronic device, fig. 7 is a schematic structural diagram of the electronic device provided by the present invention, and as shown in fig. 7, the electronic device may include: processor 710, communication interface (Communications Interface) 720, memory 730, and communication bus 740, wherein processor 710, communication interface 720, memory 730 communicate with each other via communication bus 740. Processor 710 may invoke logic instructions in memory 730 to perform all or part of the steps of a method for computing energy efficiency of the device, the method comprising:

Further, the logic instructions in the memory 730 described above may be implemented in the form of software functional units and may be stored in a computer readable storage medium when sold or used as a stand alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method for calculating the energy efficiency of the device according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform all or part of the steps of the method for calculating energy efficiency of a device according to the above embodiments, the method comprising:

In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements all or part of the steps of a method for computing energy efficiency of a device as described in the above embodiments, the method comprising:

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment of the invention. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the method of calculating the device energy efficiency of the respective embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for computing device energy efficiency, comprising:

Based on the data values of a plurality of running indexes and the data values of a plurality of asset attributes, calculating to obtain an energy efficiency result of the equipment by using a pre-constructed equipment energy efficiency calculation model;

the construction process of the equipment energy efficiency calculation model comprises the following steps:

2. The method for calculating the energy efficiency of the device according to claim 1, wherein the calculating, based on the data values of the operation indexes and the data values of the asset attributes, the energy efficiency result of the device by using a pre-constructed device energy efficiency calculation model specifically comprises:

3. The method of computing device energy efficiency of claim 1, wherein the primary energy efficiency computing scenario comprises: at least one of a device maintenance scene, a device maintenance expiration scene, a device bearing analysis scene, a device elimination scene and a device product expiration scene;

4. The method for calculating the energy efficiency of the device according to claim 3, wherein the operation index includes: at least one of CPU utilization rate, memory utilization rate, IO utilization rate, connection channel utilization rate and storage space utilization rate;

5. The method for computing energy efficiency of a device of claim 4,

The computing strategy of the equipment maintenance scene comprises the following steps: judging whether the equipment meets the requirement of having a maintenance manufacturer or not based on the data value of the maintenance manufacturer in the asset attribute of the current equipment; if the first-level energy efficiency result of the equipment in the equipment maintenance scene is met, the equipment is low-efficiency; if not, the equipment is normal;

6. The method of computing device energy efficiency of claim 5, wherein the result aggregation policy comprises: analyzing each primary energy efficiency result and each secondary energy efficiency result of each device, and if at least one result is that the device is invalid, the energy efficiency result of the device is that the device is invalid; if no device is invalid but at least one result is that the device is inefficient, the energy efficiency result of the device is that the device is inefficient; if the results are all equipment normal, the energy efficiency result of the equipment is equipment normal.

7. A computing device for device energy efficiency, comprising:

The energy efficiency calculation module is used for calculating and obtaining an energy efficiency result of the equipment by applying a pre-constructed equipment energy efficiency calculation model based on the data values of a plurality of operation indexes and the data values of a plurality of asset attributes;

8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements all or part of the steps of the method for computing energy efficiency of a device according to any of claims 1-6 when the computer program is executed by the processor.

9. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, carries out all or part of the steps of a method of calculating the energy efficiency of a device according to any one of claims 1-6.