Performance index abnormity pre-judgment method and system

Abstract

The invention discloses a performance index abnormity pre-judgment method and system, and the method comprises the steps: collecting the historical time series data of a first performance index and a first main control variable, and building a relation model between the first performance index and the first main control variable; on the basis of the relation model, whether abnormity exists or not is judged by calculating the deviation between the current first performance index and an expected value; and according to a deviation result of the current first performance index and an expected value, combining trend analysis of historical time series data, and predicting time required for the first performance index to be degraded from a current state to an abnormal state. The method can accurately predict the abnormal time and degree of the unit performance index through the establishment of the relation model and the calculation of the abnormal pre-judgment, and improves the pre-judgment accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormal prediction of unit performance, and particularly to a method and system for abnormal prediction of performance indicators. Background Art

[0002] In the field of abnormal prediction of unit performance, the existing technologies mainly rely on a series of tools based on abstract ideal mathematical models and neural network algorithms. These technologies attempt to predict the abnormalities of unit performance by analyzing the large data of the unit. However, due to the huge amount of data and the unclear relationships between data, these technologies often fail to obtain satisfactory results in practical applications.

[0003] Specifically, classical fitting models usually rely on certain preset assumptions and parameters, which may not fully conform to the actual operating conditions of the unit, resulting in limited prediction accuracy of the model. Although the tools based on neural network algorithms have strong self - adaptability and learning ability, due to their complexity and high requirements for the amount of data, they often fail to achieve the best results in the analysis of unit big data. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method and system for abnormal prediction of performance indicators to solve the problems in the existing technologies that due to relying on fixed assumptions and complex algorithms, it is difficult to flexibly adapt to the changes of actual working conditions when dealing with unit big data and non - solidified parameter indicators, resulting in limited prediction accuracy and poor actual application effects.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for abnormal prediction of performance indicators, including:

[0008] Collecting historical time - series data of the first performance indicator and the first main control variable, and establishing a relationship model between the first performance indicator and the first main control variable;

[0009] Based on the relationship model, judging whether there is an abnormality by calculating the deviation between the current first performance indicator and the expected value;

[0010] According to the deviation result between the current first performance indicator and the expected value, and combining with the trend analysis of historical time - series data, predicting the time required for the first performance indicator to deteriorate from the current state to an abnormal state.

[0011] As a preferred solution of the method for abnormal prediction of performance indicators according to the present invention, wherein:

[0012] Establishing the relationship model between the first performance index and the first main control variable includes the following steps:

[0013] Collect historical time-series data of the first performance index X and the first main control variable Y;

[0014] Find the period T(s) when the first main control variable Y is in a stable state;

[0015] Extract all samples of the first performance index within the period T(s);

[0016] Within the range of T(s), associate the sample value of each first performance index with a corresponding value of the first main control variable to form the first set C(x, y);

[0017] In the first set C(x, y), classify according to the value of the first main control variable Y, and group all samples of the first performance index associated with the same value of the first main control variable into one category to obtain the second set C(X y );

[0018] For each second set C(X y ), calculate the expected value XP of the first performance index y , and obtain the corresponding relationship model M(XP y , Y) between the expected value of the first main control variable and the first performance index.

[0019] As a preferred solution of the performance index anomaly prediction method described in the present invention, wherein:

[0020] Judging whether there is an anomaly by calculating the deviation between the current first performance index and the expected value includes:

[0021] If the deviation between the current first performance index X t and the expected value XP of the first performance index y is greater than the first threshold Rxf, it is directly determined as an anomaly and the process is terminated;

[0022] If the deviation between the current first performance index X t and the expected value XP of the first performance index y is less than the first threshold Rxf, continue with the trend analysis.

[0023] As a preferred solution of the performance index anomaly prediction method described in the present invention, wherein:

[0024] The trend analysis of the historical time-series data in combination with the deviation result between the current first performance index and the expected value includes the following steps:

[0025] Select the one containing the current first performance index X tand the third set Cn(X including the first N historical values t, X t-1 , X t-2 ,..., X t-n );

[0026] Filter out the values corresponding to the steady state stage of the first main control variable Y from the third set Cn(X t, X t-1 , X t-2 ,..., X t-n ) to obtain the fourth set Cmx(X t1 , X t2 ,..., X tm ), and record the corresponding time points T(TX t1 , TX t2 ,..., TX tm ) and the values of the first main control variable Y to obtain the fifth set

[0027] Cmy(Y t1 , Y t2 ,..., Y tm );

[0028] Using the relationship model, denote the expected values XP of the first performance indicators corresponding to the elements in the fifth set Cmy(Y t1 , Y t2 ,..., Y tm ) as the sixth set Cmxp(XP y , XP t1 ,..., XP t2 ,..., XP[[ID=?]] tm );

[0029] Calculate the differences between the corresponding elements of the fourth set Cmx(X t1 [[ID=6?]], X t2 ,..., X tm ) and the sixth set Cmxp(XP t1 , XP t2 ,..., XP tm ) to obtain the seventh set Cmd(X d1 , X d2 ,..., X dm );

[0030] Calculate the seventh set Cmd(X d1 , X d2 ,..., X dm ) and the sixth set Cmxp(XP t1 , XP t2 ,..., XP tm It seems there is a missing number in the original text at . Please check and correct it if needed. Also, there are some "?" marks added in the translation where the original text seems to have some formatting or numbering issues that might be typos.) the ratio between the corresponding elements, and get the eighth set Cmdr(R1,R2,...,R m );

[0031] Calculate the eighth set Cmdr(R1,R2,...,R m ) of the element polarity specific volume J m ;

[0032] Based on the polar specific volume J m The calculation result is used to determine whether to start predicting the time Td required for degradation from the current time to abnormality.

[0033] As a preferred solution of the method for predicting abnormal performance indicators of the present invention, wherein:

[0034] The determining whether to start predicting the time required for degradation from the current time to abnormality includes:

[0035] If the absolute value of the polar specific volume |J m | is greater than a preset tolerance parameter R and the tolerance parameter R is greater than 1, then start predicting the time Td required for the first performance indicator to deteriorate from the current time to an abnormality;

[0036] If the absolute value of the polar specific volume |J m If the value is less than the preset tolerance parameter R, the prediction of the time Td required for the first performance indicator to deteriorate from the current time to abnormality will not be started.

[0037] As a preferred solution of the method for predicting abnormal performance indicators of the present invention, wherein:

[0038] The time Td required from the current time degradation to abnormality is expressed as:

[0039] Td=Rxf / AVG(R1,R2,...,R m )*N

[0040] Among them, AVG(R1,R2,...,R m ) represents the eighth set Cmdr(R1,R2,...,R m ) is the element-wise mean of .

[0041] As a preferred solution of the method for predicting abnormal performance indicators of the present invention, wherein:

[0042] The expected value XP of the first performance indicator is calculated y This involves calculating the second set C(X y ) is the weighted mean of the elements in .

[0043] In a second aspect, the present invention provides a performance indicator abnormality prediction system, comprising:

[0044] An acquisition module, configured to acquire historical time-series data of a first performance indicator and a first main control variable, and establish a relationship model between the first performance indicator and the first main control variable;

[0045] A calculation module, configured to determine whether there is an abnormality based on the relationship model by calculating the deviation between the current first performance indicator and the expected value;

[0046] A prediction module, configured to predict the time required for the first performance indicator to deteriorate from the current state to an abnormal state based on the deviation result between the current first performance indicator and the expected value and in combination with the trend analysis of the historical time-series data.

[0047] In a third aspect, the present invention provides a computing device, including:

[0048] A memory, configured to store a program;

[0049] A processor, configured to execute the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the performance indicator abnormality prediction method are implemented.

[0050] In a fourth aspect, the present invention provides a computer-readable storage medium, including: when the program is executed by a processor, the steps of the performance indicator abnormality prediction method are implemented.

[0051] Advantages of the present invention: By establishing a relationship model and calculating abnormality prediction, the method of the present invention can accurately predict the abnormal time and degree of the unit performance indicator, improving the accuracy of prediction. By accurately predicting the abnormality of the unit performance indicator, the operation and maintenance personnel can timely carry out maintenance design and planning, improving the utilization rate of the unit and reducing losses. It is not limited to the understanding and assumption of the data correlation relationship, can flexibly process the big data of the unit, has strong adaptability, clear steps, is easy to understand and operate, and is convenient for the operation and maintenance personnel to use in actual applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for description in the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. Among them:

[0053] Figure 1 It is a basic process schematic diagram of a performance indicator abnormality prediction method provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0055] Embodiment 1

[0056] Referring to Figure 1 , for an embodiment of the present invention, a method for predicting abnormal performance indicators is provided, including:

[0057] S1: Collect historical time-series data of the first performance indicator and the first main control variable, and establish a relationship model between the first performance indicator and the first main control variable;

[0058] In the embodiments of the present application, establishing a relationship model M(XP y , Y) between the first performance indicator X and its first main control variable Y includes the following steps:

[0059] (1) Collect time-series data of the first performance indicator X and its first main control variable Y in history;

[0060] (2) Find out the time period T(s) when the first main control variable Y is in a stable state;

[0061] (3) Extract all samples of the first performance indicator X within the time period T(S);

[0062] (4) According to the principle of time consistency within the range of T(S), match each sample value of X with a value of Y to form the first set C(x, y);

[0063] (5) In C(x, y), classify according to the value of the first main control variable Y, and group all samples of X associated with the same Y value into one category, thereby forming a second set C(X y ) corresponding to different Y values;

[0064] (6) For each C(X y ), find the expected value XP y of X, and obtain a one-to-one correspondence relationship model M(XP y , Y) between the first main control variable Y and the expected value of X;

[0065] It should be noted that S1 provides a basic model for subsequent steps, enabling the prediction of the expected value XP of the indicator X based on the current value of Y. y .

[0066] S2: Based on the relational model, determine whether there is an anomaly by calculating the deviation between the current first performance indicator and the expected value;

[0067] In the embodiment of the present application, determining whether an abnormality exists by calculating the deviation between the current first performance indicator and the expected value includes:

[0068] Let the current value of X be X t , and it is consistent with the model M(XP y ,Y) determines the expected value of X, XP y The deviation is less than Rxf; (If the deviation is already greater than Rxf, the following calculation is not performed.)

[0069] It should be noted that by comparing the current value X t and expected value XP y If the deviation is small, the historical data is further analyzed to determine whether a forecast should be initiated.

[0070] S3: Based on the deviation between the current first performance indicator and the expected value and combined with trend analysis of historical time series data, predict the time required for the first performance indicator to deteriorate from the current state to an abnormal state.

[0071] In the embodiment of the present application, based on the deviation result between the current first performance indicator and the expected value, combined with the trend analysis of historical time series data, the following is included:

[0072] Select the current value X t And the third set of values of all first performance indicators X with N+1 values within a time length N is Cn(X t ,X t-1 ,X t-2 ,...,X t-n );

[0073] Take the set C(X t-1 ,X t-2 ,...,X t-n ) in the steady state phase of the first main control variable Y, forming a fourth set Cmx(X t1 ,X t2 ,...,X tm ), the time corresponding to these element samples is recorded as T(TX t1 ,TX t2 ,...,TX tm ), the fifth set Cmy(Y t1 ,Y t2 ,...,Y tm );

[0074] It should be noted that this step provides a more comprehensive perspective to evaluate the changing trend of the current first performance metric by reviewing historical data, helping to identify the historical performance of the first performance metric X under the steady state of the first main control variable Y.

[0075] Find the model M(XP y ,Y), and denote the expected values XP of the first performance metrics corresponding to the elements in the fifth set Cmy(Y t1 ,Y t2 ,...,Y tm ) as the sixth set Cmxp(XP y ,XP t1 ,...,XP t2 ,...,XP tm );

[0076] According to the principle of time consistency, calculate the difference between the corresponding elements of the fourth set Cmx(X t1 ,X t2 ,...,X tm ) and the sixth set Cmxp(XP t1 ,XP t2 ,...,XP tm ) to obtain the seventh set Cmd(X d1 ,X d2 ,...,X dm );

[0077] Calculate the ratio of the corresponding elements between the seventh set Cmd(X d1 ,X d2 ,...,X dm ) and the sixth set Cmxp(XP t1 ,XP t2 ,...,XP tm ), that is, R m = X dm / XP tm to obtain the eighth set Cmdr(R1, R2,..., R m );

[0078] Calculate the element polarity ratio J m of the eighth set Cmdr(R1, R2,..., R m ), J m = number of positive samples / number of negative samples or its reciprocal; and |J m | ≥ 1. m

[0079] Based on the calculation result of the polarity ratio J m , determine whether to start predicting the time Td required to deteriorate from the current time to an abnormal state.

[0080] In the embodiment of the present application, when |J m | > the preset tolerance parameter R and R > 1, the deterioration time Td required for predicting the first performance indicator X to deteriorate from the current time to an abnormal state is started; otherwise, it is not started.

[0081] It should be noted that this step synthesizes the results of all the previous steps. By deeply analyzing the historical data and the current data, the time Td required for predicting the first performance indicator X to deteriorate to an abnormal state depends on the previously established relationship model, the deviation between the current value and the expected value, and the trend analysis of the historical data.

[0082] In the embodiment of the present application, the deterioration time Td required for deteriorating from the current time to an abnormal state is expressed as:

[0083] Td = Rxf / AVG(R1, R2,..., R m ) * N

[0084] Where AVG(R1, R2,..., R m ) represents the element mean of the eighth set Cmdr(R1, R2,..., R m ).

[0085] In the embodiment of the present application, the expected value XP of X in step S1-(6) y = the weighted mean or the mean of the elements in the set C(X y ).

[0086] In the embodiment of the present application, taking the vibration index of a large generator set as an example, assume that there is a certain relationship between the index X and its main control variable Y (such as load). Through the method proposed in this patent, a relationship model between X and Y can be established, and based on the deviation between the current vibration index value Xt and the expected value, as well as the set of vibration index values over a past period of time, the time required for the index to deteriorate to an abnormal state can be predicted.

[0087] This embodiment also provides a system for predicting abnormal performance indicators, including:

[0088] An acquisition module, configured to acquire historical time-series data of the first performance indicator and the first main control variable, and establish a relationship model between the first performance indicator and the first main control variable;

[0089] A calculation module, configured to determine whether there is an abnormality by calculating the deviation between the current first performance indicator and the expected value based on the relationship model;

[0090] A prediction module, configured to predict the time required for the first performance indicator to deteriorate from the current state to an abnormal state according to the deviation result between the current first performance indicator and the expected value, in combination with the trend analysis of the historical time-series data.

[0091] Furthermore, it further includes:

[0092] A memory for storing programs;

[0093] A processor for loading the program to execute the method for predicting abnormal performance indicators.

[0094] This embodiment also provides a computer-readable storage medium storing a program, which when executed by a processor, implements the method for predicting abnormal performance indicators.

[0095] The storage medium proposed in this embodiment and the method for predicting abnormal performance indicators proposed in the above embodiment belong to the same inventive concept. Technical details not described in detail in this embodiment can be referred to in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0096] From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general-purpose hardware, and of course, it can also be implemented by hardware. However, in many cases, the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk, or optical disc of a computer, and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0097] Embodiment 2

[0098] This is an embodiment of the present invention, which provides a system for predicting abnormal performance indicators, including an acquisition module, a calculation module, and a prediction module;

[0099] In the embodiment of the present application, the acquisition module includes acquiring historical time-series data of the first performance indicator and the first main control variable, and establishing a relationship model between the first performance indicator and the first main control variable;

[0100] In the embodiment of the present application, the calculation module includes judging whether there is an abnormality based on the relationship model by calculating the deviation between the current first performance indicator and the expected value;

[0101] In the embodiment of the present application, the prediction module includes predicting the time required for the first performance indicator to deteriorate from the current state to an abnormal state based on the deviation result between the current first performance indicator and the expected value and combining the trend analysis of the historical time-series data.

[0102] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for predicting abnormal performance indicators, characterized in that, Including: Collect the historical time-series data of the first performance indicator and the first main control variable, and establish a relationship model between the first performance indicator and the first main control variable; Based on the relationship model, determine whether there is an abnormality by calculating the deviation between the current first performance indicator and the expected value; According to the deviation result between the current first performance indicator and the expected value, combined with the trend analysis of the historical time-series data, predict the time required for the first performance indicator to deteriorate from the current state to an abnormal state.

2. The performance index abnormal prediction method according to claim 1, wherein: The establishment of the relationship model between the first performance indicator and the first main control variable includes the following steps: Collect the historical time-series data of the first performance indicator X and the first main control variable Y; Find the time period T(s) when the first main control variable Y is in a stable state; Extract all samples of the first performance indicator within the time period T(s); Within the range of T(s), associate the sample value of each first performance indicator with a corresponding first main control variable value to form the first set C(x,y); In the first set C(x, y), classification is performed according to the value of the first main control variable Y, and all samples of the first performance index that are also associated with the same value of the first main control variable are grouped into one class, resulting in the second set C(X y ); For each second set C(X y ), calculate the expected value XP of the first performance metric y , and obtain a corresponding relationship model M(XP y , Y) between the first main control variable and the expected value of the first performance metric.

3. The method for pre-judging abnormal performance indicators as described in claim 1 or 2, characterized in that: The determination of whether there is an abnormality by calculating the deviation between the current first performance indicator and the expected value includes: If the current first performance indicator X t has a deviation from the expected value XP of the first performance indicator y greater than the first threshold Rxf, it is directly determined as abnormal and the process is terminated; If the current first performance metric X t has a deviation from the expected value XP y of the first performance metric that is less than the first threshold Rxf, then continue with the trend analysis.

4. The method for predicting abnormal performance indicators according to claim 3, characterized in that: The combination of the deviation result between the current first performance indicator and the expected value and the trend analysis of the historical time-series data includes the following steps: Select the third set Cn(X that includes the current first performance metric X t and the first N historical values t, X t-1 , X t-2 ,..., X t-n ); From the third set Cn(X t, X t-1 ,X t-2 ,...,X t-n ), select the values corresponding to the steady state stage of the first main control variable Y to obtain the fourth set Cmx(X t1 ,X t2 ,...,X tm ). At the same time, record the corresponding time points T(TX t1 ,TX t2 ,...,TX tm ) and the values of the first main control variable Y to obtain the fifth set Cmy(Y t1 ,Y t2 ,...,Y tm ); Using the relational model M(XP y , Y), the expected value XP of the first performance indicator corresponding to the elements in the fifth set Cmy(Y t1 , Y t2 ,..., Y tm ) is denoted as the sixth set Cmxp(XP y , XP t1 ,..., XP t2 ,..., XP tm ); Calculate the difference between the corresponding elements of the fourth set Cmx(X t1 , X t2 ,..., X tm ) and the sixth set Cmxp(XP t1 , XP t2 ,..., XP tm ) to obtain the seventh set Cmd(X d1 , X d2 ,..., X dm ); Calculate the ratio between the corresponding elements of the seventh set Cmd(X d1 ,X d2 ,...,X dm ) and the sixth set Cmxp(XP t1 ,XP t2 ,...,XP tm ) to obtain the eighth set Cmdr(R1,R2,...,R m ); Calculate the element polarity ratio J of the eighth set Cmdr(R1, R2,..., R m ) m ; Based on the calculation result of the polar specific volume J m judge whether to start predicting the time Td required for deterioration from the current time to abnormality.

5. The method for predicting abnormal performance indicators according to claim 4, characterized in that: The determination of whether to start predicting the time required to deteriorate from the current time to an abnormal state includes: If the absolute value |J m | of the polar specific volume is greater than a preset tolerance parameter R and the tolerance parameter R is greater than 1, then start predicting the time Td required for the first performance index to deteriorate from the current time to an abnormal state; If the absolute value of the polarity specific volume |J m | is less than the preset tolerance parameter R, the time Td required to predict the deterioration of the first performance index from the current time to the abnormal state is not started.

6. The performance index abnormal prediction method according to claim 5, wherein: The time Td required to deteriorate from the current time to an abnormal state is expressed as: Td = Rxf / AVG(R1, R2,..., R m ) * N where AVG(R1, R2,..., R m ) represents the element mean of the eighth set Cmdr(R1, R2,..., R m ).

7. The method for predicting abnormal performance indicators according to claim 6, characterized in that: Calculating the expected value XP of the first performance metric y includes calculating the weighted mean of the elements within the second set C(X y ).

8. An abnormal prediction system for performance indicators, which applies the method according to any one of claims 1-7, is characterized in that, Including: A collection module for collecting the historical time-series data of the first performance indicator and the first main control variable, and establishing a relationship model between the first performance indicator and the first main control variable; A calculation module for determining whether there is an abnormality by calculating the deviation between the current first performance indicator and the expected value based on the relationship model; A prediction module for predicting the time required for the first performance indicator to deteriorate from the current state to an abnormal state according to the deviation result between the current first performance indicator and the expected value, combined with the trend analysis of the historical time-series data.

9. A computing device, characterized in that, Including: A memory for storing programs; A processor for loading the program to execute the steps of the performance indicator abnormality prediction method according to any one of claims 1-7.

10. A computer-readable storage medium stores a program, characterized in that, When the program is executed by the processor, the steps of the performance indicator abnormality prediction method according to any one of claims 1-7 are implemented.

Images