CN115983534A - Method and system for evaluating state of sewage treatment process - Google Patents
Method and system for evaluating state of sewage treatment process Download PDFInfo
- Publication number
- CN115983534A CN115983534A CN202310183342.0A CN202310183342A CN115983534A CN 115983534 A CN115983534 A CN 115983534A CN 202310183342 A CN202310183342 A CN 202310183342A CN 115983534 A CN115983534 A CN 115983534A
- Authority
- CN
- China
- Prior art keywords
- state
- online
- evaluation
- ers
- variable
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02W—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO WASTEWATER TREATMENT OR WASTE MANAGEMENT
- Y02W10/00—Technologies for wastewater treatment
- Y02W10/10—Biological treatment of water, waste water, or sewage
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a state evaluation method, a state evaluation system and a storage medium in a sewage treatment process. The state evaluation method comprises the following steps: acquiring online process data and acquiring a corresponding online economic index value; calculating mutual information between each process variable dimension and the online economic index value so as to divide online process data of each process variable dimension into an ERS variable space and an EIS variable space; performing regular correlation analysis on online process data in an ERS variable space, and extracting economic information in the online process data to construct ERS evaluation statistics; performing slow characteristic analysis on online process data in an EIS variable space, and extracting dynamic information in the online process data to construct EIS evaluation statistics; and determining an operation state label corresponding to the online economic index value according to the Bayesian theory fusion ERS evaluation statistic and EIS evaluation statistic so as to determine the operation state of the sewage treatment process.
Description
Technical Field
The present invention relates to the field of sewage treatment, and in particular, to a method and a system for evaluating a state of a sewage treatment process, and a computer-readable storage medium thereof.
Background
Municipal sewage treatment plants have been a field of interest and research and can be viewed as a complex nonlinear system. There are various biochemical and physical reactions in the process, with large flow and load perturbations. The nature of the wastewater treatment process dictates its need to work continuously, with increasingly stringent treatment standards. For modern enterprises aiming at improving economic benefits, a good process running state is required to ensure the product quality, reduce the production cost and reduce the economic loss on the basis of normal process running. However, in an actual industrial production process, the process operation state often deviates from the optimal operating point gradually due to factors such as equipment aging and parameter drift, so that the process is operated under an unhealthy working condition for a long time, which results in poor product quality and affects enterprise income. Although process monitoring technology has been widely used in industrial fields, merely distinguishing the operational status of a process by normality/malfunction is far from satisfying the production requirements of modern enterprises. This is because the process may already be in a non-optimal operating state before the anomaly occurs. Therefore, from the perspective of reducing production cost and improving economic benefits, the quality of the process running state needs to be further evaluated, and once the process deviates from the optimal running state, a regulation and control measure needs to be immediately taken to ensure that the process is always kept in the optimal running state.
As shown in fig. 1, the current process operation state evaluation method may be mainly classified into three types, a quantitative information method, a qualitative information method, and a method of combining quantitative and qualitative information. The evaluation method based on quantitative information refers to evaluation using variable information represented by numerical values, and a statistical learning method is mainly used. The evaluation method based on qualitative information refers to evaluation using variable information described semantically, and the most commonly used evaluation methods for processing qualitative information include Fuzzy theory (Fuzzy theory) and probabilistic Rough set theory (RS). Because the quantitative method relies on a large amount of process data, the reliability of the quantitative method is lost for production processes in which some process variables are difficult to obtain on-line or variable information cannot fully reflect process information. The traditional qualitative method needs sufficient process description information which is often discretized, and effective information is easily lost in the discretization process, so that the evaluation precision is reduced. The method solves the problem of single process characteristics, but most of the actual sewage treatment processes are composite variable non-Gaussian dynamic processes, and the variables are various and are coupled with one another. In addition, the process operation state evaluation is targeted at a complex industrial system, various process characteristics coexist, multiple different links and units are involved, and the relationship of mutual restriction and connection exists among the links and among the units, while the traditional process operation state evaluation method utilizes all variation information of process data to evaluate the state. Since the process operation state evaluation aims at reflecting the process operation state related to economic benefits, excessive process information unrelated to economic indicators can lead to the reduction of the performance of an evaluation model. Therefore, the traditional process running state evaluation method has the defects of large calculation amount and inaccurate evaluation result.
In order to overcome the above-mentioned defects in the prior art, there is a need in the art for a state evaluation technique for a sewage treatment process, which is used for simply, quickly and practically calculating the correlation between data to block the data, so as to effectively capture useful dynamic information in the process while overcoming data redundancy, shorten online evaluation time, improve evaluation accuracy and result interpretability, maintain continuous industrial process, obtain better evaluation accuracy, and reduce the cost of model maintenance.
Disclosure of Invention
The following presents a simplified summary of one or more aspects in order to provide a basic understanding of such aspects. This summary is not an extensive overview of all contemplated aspects, and is intended to neither identify key or critical elements of all aspects nor delineate the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that is presented later.
In order to overcome the defects in the prior art, the invention provides a state evaluation method of a sewage treatment process, a state evaluation system of the sewage treatment process and a corresponding computer readable storage medium, which can simply, quickly and practically use the correlation among mutual information calculation data to block the data, thereby effectively capturing useful dynamic information in the process while overcoming data redundancy, greatly shortening online evaluation time, and improving the accuracy of evaluation and the interpretability of results. In addition, mechanism knowledge is not needed in the experimental process of the method, so that the industrial process is not needed to be suspended, and meanwhile, the method can change the model according to actual needs and data changes so as to realize the online evaluation of the process running state, thereby obtaining better evaluation precision and reducing the cost of model maintenance.
Specifically, the method for evaluating the state of the sewage treatment process provided by the first aspect of the present invention comprises the following steps: acquiring online process data according to a plurality of process variable dimensions involved in the sewage treatment process, and acquiring corresponding online economic index values; calculating mutual information between each process variable dimension and the online economic index value so as to divide online process data of each process variable dimension into an ERS variable space and an EIS variable space, wherein the ERS variable space is related to economic performance, and the EIS variable space is not related to economic performance; performing regular correlation analysis on the online process data in the ERS variable space, and extracting economic information in the online process data to construct ERS evaluation statistics; performing slow characteristic analysis on the online process data in the EIS variable space, and extracting dynamic information in the online process data to construct EIS evaluation statistics; and fusing the ERS evaluation statistic and the EIS evaluation statistic according to Bayesian theory, and determining an operation state label corresponding to the online economic index value to determine the operation state of the sewage treatment process.
Further, in some embodiments of the invention,the sewage treatment process relates to a plurality of reaction tanks, including an anoxic tank, an aerobic tank and a secondary sedimentation tank, and the dimensions of a plurality of process variables are selected from BOD total amount of effluent, COD total amount of effluent, pump energy consumption, flow rate of effluent, internal reflux flow rate, soluble biodegradable organic nitrogen of effluent, and effluent
Concentration, effluent Kjeldahl nitrogen concentration, effluent dissolved oxygen concentration, effluent nitrate concentration, effluent nitrite concentration, effluent granular biodegradable organic nitrogen, dissolved oxygen content of each reaction tank, nitrate and nitrite concentration of each reaction tank, and concentration of each reaction tank
At least one of the concentration, the soluble biodegradable organic nitrogen of each reaction tank, the granular biodegradable organic nitrogen of each reaction tank, and the aeration intensity of the last reaction tank.
Further, in some embodiments of the present invention, the step of obtaining online process data according to a plurality of process variable dimensions involved in the sewage treatment process and obtaining corresponding online economic indicator values comprises: constructing an online data window; and acquiring multiple groups of online process data related to the multiple process variable dimensions and corresponding online economic index values thereof according to the size H of the online data window.
Further, in some embodiments of the present invention, after acquiring online process data according to a plurality of process variable dimensions involved in the sewage treatment process and acquiring corresponding online economic indicator values, the state evaluation method further comprises the steps of: standardizing multiple groups of on-line process data according to the following formula to construct a zero-mean unit standard deviation matrix X m×n ：
Wherein x is ij Is the mean unit standard deviation matrix X m×n M is the number of groups of the on-line process data, n is the number of groups of the on-line process dataA process variable dimension number in the online process data.
Further, in some embodiments of the present invention, mutual information between each of the process variable dimensions and the online economic indicator value is calculated by:
wherein, H (X) and H (Y) are information entropies of variables X and Y, H (X, Y) represents the joint entropy of the variables X and Y, p (X, Y) represents the joint probability density function of X and Y, and H (X), H (Y) and H (X, Y) are respectively calculated by the following formula:
H(X)＝-∫p(x)logp(x)dx
H(Y)＝-∫p(y)logp(y)dy
H(X,Y)＝-∫∫p(x,y)logp(x,y)dxdy。
further, in some embodiments of the present invention, the step of calculating mutual information between each of the process variable dimensions and the online economic indicator value to partition online process data of each of the process variable dimensions into an ERS variable space and an EIS variable space includes: comparing a mutual information value MI between each process variable dimension and the online economic index value with a preset first threshold value th; and if the mutual information value MI is larger than the first threshold value th, dividing the corresponding process variable dimension into an ERS variable space X E On the contrary, if the mutual information value MI is less than or equal to the first threshold value th, the corresponding process variable dimension is divided into the EIS variable space X D Wherein the total variable space X = [ X = D ,X E ]。
Further, in some embodiments of the present invention, the step of performing a canonical correlation analysis on the online process data in the ERS variable space to extract economic information therein to construct an ERS evaluation statistic includes:
constructing the ERS evaluation statistic according to the following formula
Wherein the content of the first and second substances,
and &>
Is N l,E Identity matrix of dimension, N l,E Indicates the number of variables in the ERS variable space in the status level of type I, <' > H>
Is ERS variable space->
One sample data of (1), J l And L l Respectively representing the ERS variable space in the first state level and the typical related variable, sigma l A matrix of singular values representing the ERS variable space within the l-th state class.
Further, in some embodiments of the present invention, the slow feature analysis of the online process data in the EIS variable space to extract dynamic information therein to construct the EIS evaluation statistic comprises:
constructing the EIS evaluation statistic according to the following formula
Wherein the content of the first and second substances,
is the EIS variable space->
One sample data of (2), W l Is a matrix of coefficients in the EIS variable space within the l-th state level.
Further, in some embodiments of the present invention, the step of determining the operation state label corresponding to the online economic index value by fusing the ERS evaluation statistic and the EIS evaluation statistic according to the bayesian theory to determine the operation state of the sewage treatment process includes:
let C l Represents the l operation state and transmits the on-line process data x k Is expressed as
/>
And x is calculated by the following formula k Probability within class i operating state:
wherein, pr (C) l ) Indicates that the current process state is at C l The probability of (a) of (b) being,
the representation indicates that the current process state is not at C l Probability of (g), pr [ g ] l (x k )|C l ]And &>
Can be calculated from:
wherein the content of the first and second substances,
offline evaluation criterion which represents the status rating of the l < th > species >>
And
and (3) obtaining global evaluation statistics in different state levels by Bayesian inference:
Further, in some embodiments of the present invention, before the ERS evaluation statistic and the EIS evaluation statistic are fused according to the bayesian theory, and the operation state label corresponding to the online economic index value is determined to determine the operation state of the sewage treatment process, the state evaluation method further includes the following steps: selecting a plurality of process variables related in the sewage treatment process, and acquiring a plurality of groups of off-line process data and off-line economic index values corresponding to the off-line process data according to the process variables; dividing each group of the off-line process data into a plurality of state levels related to economic benefit indexes by utilizing a Gaussian mixture model; calculating mutual information between each process variable dimension and the offline economic index value, and blocking the process variables according to the mutual information to establish the ERS variable space and the EIS variable space; performing regular correlation analysis on the ERS variable space, and extracting economic information in the ERS variable space to construct ERS evaluation statistics; and performing slow characteristic analysis on the EIS variable space, and extracting dynamic information in the EIS variable space to construct EIS evaluation statistics.
Further, in some embodiments of the present invention, the step of classifying the sets of offline process data into a plurality of state levels related to economic benefit indicators by using a gaussian mixture model comprises: selecting an overall cost index OCI as the economic benefit index, and calculating the economic benefit index according to the following formula:
OCI＝AE+PE+5*SP+3*EC+ME
wherein AE represents aeration energy consumption, PE represents pump energy consumption, SP represents sludge production, EC represents external carbon source consumption, and ME represents mixing energy consumption; differentiating, via a clustering method, multiple operating state levels { X) of the multiple sets of offline process data from a probabilistic perspective using the Gaussian mixture model l ,y l L =1, 2.., L, wherein, X l And y l Respectively representing the process data and the economic indicator OCI within the state set l,
N l representing the number of samples in the state set l; according to >>
Determining a rank label for each state set in which>
The smaller the representative state set l, the better the behavior, and the better the corresponding rank label.
Further, according to a second aspect of the present invention, there is provided a system for evaluating a state of a sewage treatment process, comprising: a memory; and the processor is connected with the memory and is configured with the state evaluation method of the sewage treatment process.
Further, according to a third aspect of the present invention, there is provided a computer-readable storage medium having stored thereon computer instructions, wherein the computer instructions, when executed by a processor, implement the above-described method for evaluating a state of a wastewater treatment process.
Drawings
The above features and advantages of the present disclosure will be better understood upon reading the detailed description of embodiments of the disclosure in conjunction with the following drawings. In the drawings, components are not necessarily drawn to scale, and components having similar relative characteristics or features may have the same or similar reference numerals.
FIG. 1 illustrates a process operating state evaluation method classification diagram provided in accordance with some embodiments of the invention;
FIG. 2 illustrates a schematic diagram of a modeling strategy for evaluating a model provided in accordance with some embodiments of the invention;
FIG. 3 illustrates a system framework diagram of a wastewater treatment process operational status evaluation method according to some embodiments of the invention;
FIG. 4 illustrates a simplified schematic diagram of a simulation model provided in accordance with some embodiments of the invention;
5A-5C illustrate schematic views of online evaluation results of operating condition evaluations provided in accordance with some embodiments of the present invention;
6A-6D illustrate schematic diagrams of online evaluation results of operating condition evaluations provided in accordance with some embodiments of the present invention;
fig. 7A-7D are schematic diagrams illustrating online evaluation results of operating condition evaluations provided according to some embodiments of the present invention.
Detailed Description
The following description is given by way of example of the present invention and other advantages and features of the present invention will become apparent to those skilled in the art from the following detailed description. While the invention will be described in connection with the preferred embodiments, there is no intent to limit the features of the invention to those embodiments. On the contrary, the invention is described in connection with the embodiments for the purpose of covering alternatives or modifications that may be extended based on the claims of the present invention. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The invention may be practiced without these particulars. Moreover, some of the specific details have been left out of the description in order to avoid obscuring or obscuring the focus of the present invention.
In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Additionally, the terms "upper," "lower," "left," "right," "top," "bottom," "horizontal," "vertical" and the like as used in the following description are to be understood as referring to the segment and the associated drawings in the illustrated orientation. The relative terms are used for convenience of description only and do not imply that the described apparatus should be constructed or operated in a particular orientation and therefore should not be construed as limiting the invention.
It will be understood that, although the terms first, second, third, etc. may be used herein to describe various elements, regions, layers and/or sections, these elements, regions, layers and/or sections should not be limited by these terms, but rather are used to distinguish one element, region, layer and/or section from another element, region, layer and/or section. Thus, a first component, region, layer or section discussed below could be termed a second component, region, layer or section without departing from some embodiments of the present invention.
As described above, the current process operation state evaluation methods are mainly classified into three types, i.e., a quantitative information method, a qualitative information method, and a method of combining quantitative and qualitative information, as shown in fig. 1. However, the methods all solve the problem of single process characteristics, but the actual sewage treatment process is mostly a non-gaussian dynamic process with complex variables, and the variables are various and are coupled with each other. In addition, the process operation state evaluation is targeted at a complex industrial system, various process characteristics coexist, multiple different links and units are involved, and the relationship of mutual restriction and connection exists among the links and among the units, while the traditional process operation state evaluation method utilizes all variation information of process data to evaluate the state. Since the process operation state evaluation aims at reflecting the process operation state related to economic benefit, excessive process information unrelated to economic indicators can lead to the reduction of the evaluation model performance. Therefore, the traditional process running state evaluation method has the defects of large calculation amount and inaccurate evaluation result.
In order to overcome the defects in the prior art, the invention provides a state evaluation method of a sewage treatment process, a state evaluation system of the sewage treatment process and a corresponding computer readable storage medium thereof, which can simply, quickly and practically calculate the correlation among data by using mutual information so as to block the data, thereby effectively capturing useful dynamic information in the process while overcoming data redundancy, shortening online evaluation time, improving evaluation accuracy and interpretability of results, maintaining continuous industrial process, obtaining better evaluation precision and reducing model maintenance cost.
In some non-limiting embodiments, the method for evaluating the state of the wastewater treatment process provided by the first aspect of the present invention may be implemented by the system for evaluating the state of the wastewater treatment process provided by the second aspect of the present invention. Specifically, the state evaluation system of the sewage treatment process is provided with a memory and a processor. The memory includes, but is not limited to, the above-described computer-readable storage medium provided by the third aspect of the invention having computer instructions stored thereon. The processor is connected with the memory and is configured to execute the computer instructions stored on the memory so as to implement the state evaluation method of the sewage treatment process provided by the first aspect of the invention.
Please refer to fig. 2 and fig. 3 first. FIG. 2 illustrates a schematic diagram of a modeling strategy for evaluating a model provided in accordance with some embodiments of the invention. Fig. 3 illustrates a system block diagram of a sewage treatment process operation state evaluation method according to some embodiments of the present invention.
As shown in fig. 2 and 3, in some embodiments of the present invention, the method for evaluating the operation state of the sewage treatment process includes the steps of:
the method comprises the following steps: selecting process variables, and dividing the offline process data into different state grades by using a Gaussian Mixture Model (GMM);
step two: judging the quality of the state grade according to the size of the economic benefit index, and standardizing the data;
step three: calculating mutual information between every two variables by using the mutual information;
step four: partitioning the process variable according to the mutual information obtained by calculation, and respectively establishing a variable space ERS related to economic performance and a variable space EIS unrelated to economic performance;
step five: extracting economic information and dynamic information in the data by using a regular correlation analysis and slow characteristic analysis method;
step six: constructing an online data window to obtain a data matrix of the current moment and a corresponding online economic index value;
step seven: respectively standardizing the data matrix and the corresponding online economic index value according to the mean value and the variance of the offline data in different state levels;
step eight: according to different state grades, sub-spaces relevant to the economic performance indexes and irrelevant to the economic performance indexes are divided;
step nine: constructing statistics related to the economic performance indexes and unrelated to the economic performance indexes in each window;
step ten: and according to Bayesian theory, fusing the online statistics of the two subspaces to determine the most possible state grade of the online data, thereby realizing the online evaluation of the process running state.
These steps will be described in detail below. It is to be understood that within the scope of the present invention, the above-described technical features of the present invention and the technical features specifically described below (e.g., examples) may be combined with each other to constitute a preferred embodiment.
1. Selecting process variables, and dividing off-line data state grades by using Gaussian mixture model
In the first step, the sewage treatment process conforms to a Long-Term Model No.1 Benchmark Simulation Model (Long-Term Benchmark Simulation Model No.1, BSM 1). FIG. 4 shows a simplified diagram of a long-term model No.1 baseline simulation model. The long-term model No.1 reference simulation model is well known in the art and comprises two anoxic tanks, three aerobic tanks and a secondary sedimentation tank.
It is understood that in step one, the process variables are selected to reflect the close relationship between the process operating conditions and the economic indicators of wastewater treatment. In the present invention, the process variable refers to the kind of parameter to be collected. The selectable process variables comprise BOD total amount of discharged water, COD total amount of discharged water, pump energy consumption, flow rate of discharged water, internal reflux flow rate, soluble biodegradable organic nitrogen of discharged water and discharged water
Concentration, effluent Kjeldahl nitrogen concentration, effluent dissolved oxygen concentration, effluent nitrate and nitrite concentration, effluent granular biodegradable organic nitrogen, dissolved oxygen content of five reaction tanks, nitrate and nitrite concentration of five reaction tanks, and/or>
Concentration, soluble biodegradable organic nitrogen in five reaction tanks, granular biodegradable organic nitrogen in five reaction tanks, and aeration intensity of the fifth reaction tank.
In the invention, the method distinguishes different operation states of process data from the angle of probability by adopting a Gaussian mixture model clustering method in consideration of the fact that the characteristics of the process can change when the process is in different production conditions and different operation states can have different distribution conditions for complex industrial processes. Suppose a set of data x ∈ R n Subject to a gaussian distribution, its probability density function can then be expressed as:
where μ is an n-dimensional mean vector and Σ is an n × n-dimensional covariance matrix. Assuming an n-dimensional sample x from multiple operating states, its probability density function can be seen as a combination of multiple gaussian distributions, i.e.:
wherein M represents the number of Gaussian components, μ m Sum-sigma m Parameter, α, representing the m-th Gaussian component m A mixture coefficient representing the m-th Gaussian component, satisfies 0 alpha < 1 and
the number of Gaussian components is known a priori, so the most critical step for establishing the Gaussian mixture model is to solve the model parameters theta = { (alpha) m ,μ m ,Σ m ) I1M M, the parameter theta can be solved through maximum likelihood estimation and expectation maximization algorithm in general.
2. Judging whether the state grade is good or bad
In the second step, in order to better evaluate the performance condition of the BSM1 for treating sewage under different water inlet data and operation conditions, the model provides two different performance indexes: considering that the process running state evaluation focuses on reflecting the close relation between the process running state and the economic Index, the OCI is selected as the economic benefit Index for judging the grade of each process running state, and can be calculated by the following formula:
OCI＝AE+PE+5*SP+3*EC+ME
where AE represents aeration energy consumption, PE represents pump energy consumption, SP represents sludge production, EC represents external carbon source consumption, and ME represents mixing energy consumption.
After the GMM is used for dividing the offline process data X into different state sets, the corresponding economic indicators y can be correspondingly divided into different state sets, and expert knowledge and production experience are combined to know that the smaller the quantitative economic indicator value is, such as the production cost or the running cost, the better the production process running in the state is represented under the normal running condition, so the economic indicator mean value of each state set is used for determining the grade of each state set. Suppose a process is divided into L different production processes, namely { X } l ,y l L, representing process data and economic indicators within a state set L, wherein
N l Representing the number of samples in a state set l, the level of each state set being available->
Is determined and/or is taken up>
Smaller states represent better behavior of the state set l, so that the process can be classified into different class labels such as "good", "medium", "bad", and the like.
In the second step, in order to eliminate the influence of different dimensions of the variables, the original data are converted into a matrix with zero mean unit standard deviation, and M groups of data { X ] are assumed to exist m Each set of data is N-dimensional, thus constituting a matrix X m×n The formula for normalizing the data is:
3. computing mutual information between variables
Mutual Information (MI) can represent the interdependence between two variables, and is derived from the concept of entropy in Information theory, and entropy is also called Information entropy, and can represent the uncertainty of the value of a variable, and is further used for describing the Information quantity contained in the variable, and a larger entropy means a larger uncertainty of the variable, that is, a larger Information content of the variable.
In step three, the mutual information I (X, Y) between two variables is calculated as follows:
wherein H (X) and H (Y) are information entropies of variables X and Y, H (X, Y) represents the joint entropy of the variables X and Y, and p (X, Y) represents the joint probability density function of X and Y; wherein H (X), H (Y) and H (X, Y) are calculated as follows:
H(X)＝-∫p(x)logp(x)dx
H(Y)＝-∫p(y)logp(y)dy
H(X,Y)＝-∫∫p(x,y）logp（x,y）dxdy
4. blocking process variables, establishing ERS and EIS
And in the fourth step, the process variables are partitioned by utilizing mutual information.
Suppose that there is a set of input data matrices X = [ X ] at a state level 1 ,x 2 ,...,x m ] T ∈R m×n M represents the number of process variables, n represents the number of samples, and the corresponding performance index variable y CEI ∈R l×n Where the index CEI represents the overall economic indicator, then the MI values between the m process variables and the performance indicator variable can be determined, with a larger MI value meaning the input variable x i (1. Ltoreq. I.ltoreq.m) the more relevant the performance index variable is, the more economic information is contained in the variable.A threshold value th may thus be defined that classifies the process variable to ERS as soon as the MI value exceeds this threshold value, and else to EIS, as shown in the following equation:
after blocking, the original set of variables is divided into two parts: variable space ERS related to economic performance and variable space EIS, namely X [ X ], not related to economic performance D ,X E ]。
In step four, the threshold for MI is determined in combination with empirical and actual requirements. In certain embodiments, the threshold th for MI is set to 0.6.
5. Regular correlation analysis is used for extracting economic information, and slow characteristic analysis is used for extracting dynamic information
In the fifth step, because a large amount of economic information is contained in the ERS, even if certain dynamic information is contained, the economic information is also taken as the leading factor, so that typical relevant features most relevant to economic performance indexes are extracted through the CCA, and an evaluation statistic is constructed by using the CCA and calculated according to the following formula:
wherein the content of the first and second substances,
is N l,E Identity matrix of dimension, N l,E The number of variables which indicate the status class species ERS of the l th species>
Is->
One sample data of seed, J l And L l Respectively representing ERS in the first state level and the typical related variable, sigma, of the corresponding output data l And (3) a singular value matrix representing the L-th state class species ERS.
In step five, although a small amount of economic information is distributed in the EIS, the EIS mainly takes the dynamic information of the process, so the dynamic characteristics in the EIS are extracted by using the SFA, and the SFA construction evaluation statistic is calculated as follows:
wherein the content of the first and second substances,
is->
One sample data of (1), W l Is the coefficient matrix in the l-th state level EIS.
In some embodiments, the statistics in the training data are referred to as offline statistics, and the statistics in the real-time data are referred to as online statistics. In the ninth step, the online statistics need to be compared by using the corresponding offline statistics obtained in the fifth step as a reference evaluation index to form a final evaluation result.
6. Building an online window
In step six, assuming that the current time is k, the data matrix of the current timeIs X on ＝[x on (k-H+1),x on (k-H+2),...,x on (k)]And the corresponding online economic index value y on . In general, the size of the online window H may be predetermined based on a number of experiments and experience. In certain embodiments, the size of the online window H is 10.
7. Respectively standardizing the data matrix and the corresponding online economic index value according to the mean value and the variance of the offline data in different state levels
In the seventh step, the standardization method is the same as that in the second step, and the standardized online data is recorded as
8. ERS and EIS are classified according to different state grades
Step eight, the ERS and EIS divided in different state grades according to different state grades are the same as the ERS and EIS divided in different state grades according to off-line data in step four, and X is on ∈R m×H Also divided into two different subspaces, i.e.
ERS and @, representing online data>
EIS representing online data, where m E Is the number of variables in ERS, m D Is the number of variables in EIS, and satisfies mm E +m D 。/>
9. Constructing statistics of ERS and EIS at each window
In step nine, assuming that the current time is k, the statistic at the time of k is the mean statistic in the unit window. For X on, The mean statistic of the online data at the time k in the ERS is calculated as follows:
for X on, The mean statistic of the online data at the time k in the EIS is calculated as follows:
10. according to Bayes theory, the online statistics of the two subspaces are fused to determine the most possible state grade of the online data, and the online evaluation of the process running state is realized
In some embodiments, assuming that the process has L operating states in common, 2L online statistics need to be constructed, and the final evaluation result is formed by comparing the online statistics with the corresponding offline statistics as the reference evaluation index, which undoubtedly increases the amount of calculation and complexity in the online evaluation stage.
In the invention, two subspace evaluation results of each state grade are fused, and the evaluation information of each subblock is integrated from the perspective of probability to obtain a global evaluation result. Let C l Indicating the l-th operating state, on-line data x k Can be expressed as
Then x k The probability of being within the class of the i operating condition is calculated by:
wherein, pr (C) l ) Indicates that the current process state is at C l The probability of (a) of (b) being,
indicating that the current process state is not at C l Probability of (g), pr [ g ] l (x k )|C l ]And &>
Can be calculated from the following equation:
wherein the content of the first and second substances,
offline evaluation criterion which represents the status rating of the l th species->
Finally, obtaining global evaluation statistics under different state grades by Bayesian inference:
in the tenth step, if the evaluation index is larger than a certain threshold value alpha, then BIC l (k) And alpha represents that the online data at the time k is in the l operating state level, otherwise, the online data is in other operating state levels.
In general, the magnitude of the threshold α of the online evaluation index may be predetermined according to a great deal of experiments and experience. In certain embodiments, the threshold value for the online evaluation index α =0.95.
When the process running state evaluation is carried out, firstly, the method in the first step of the invention is adopted to divide the state grade of the off-line data, the method in the second step is utilized to judge the quality of the state grade, then the method in the third step is adopted to calculate the correlation between variables by utilizing mutual information, the variables are partitioned according to the correlation, the economic correlation subspace and the economic irrelevant subspace are divided, and then the method in the fourth step and the fifth step is utilized to construct evaluation statistics to extract the economic information and the dynamic information; and after the online process data are obtained, the same blocking method is adopted, evaluation statistics are constructed, the evaluation results of the two subspaces of each state grade are fused according to the method described in the step ten, the overall evaluation result of the real-time data is obtained, and the quality condition of the process state is monitored.
The invention has the following beneficial effects:
the process running state evaluation method is simple, rapid and practical, and the data are partitioned by utilizing the correlation among mutual information calculation data. Compared with the traditional process running state evaluation method, the method overcomes the data redundancy, effectively captures useful dynamic information in the process, greatly shortens the online evaluation time, considers the correlation among variables, and improves the evaluation accuracy and the interpretability of the result. Mechanism knowledge is not needed in the experimental process, and the industrial process is not needed to be suspended; meanwhile, the method can change the model according to actual needs and data changes so as to realize online evaluation of the process running state, obtain better evaluation precision and reduce the cost of model maintenance.
The present invention will be specifically described below by way of examples. It should be noted that the following examples are only for illustrating the present invention and should not be construed as limiting the scope of the present invention, and any insubstantial modifications and adaptations by those skilled in the art based on the teachings of the present invention are still within the scope of the present invention. Algorithms and methods not specifically described in the examples are those well known in the art or described herein. Algebra not explicitly described in the examples have the meaning known in the art or described herein.
Example 1
The method for evaluating the running state of the sewage treatment process based on the block canonical correlation analysis and the slow feature analysis is described by taking an example of evaluating the running state of the sewage treatment process as follows, and comprises the following specific steps:
the method comprises the following steps: selecting 38 variables shown in table 1 as process variables, adopting a constant water inlet file as water inlet data, and collecting training data of two weeks; changing the dissolved oxygen solubility of a fifth reaction tank which has influences on the whole energy consumption and the operation cost OCI to simulate each operation state which is easy to appear in the sewage treatment process, wherein three simulated operation conditions are shown in table 2, so that the mixed component number M =3 of GMM is set, and then, the GMM is used for dividing the process data into three data sets;
step two: judging grades through different state sets of the OCI mean value corresponding to each set, wherein the smaller the OCI mean value is, the more optimal the corresponding process running state is, and the larger the OCI mean value is, the worse the corresponding process running state grade is, so that the three running states are divided into three grades of 'superior, intermediate and inferior', each grade is marked by corresponding 1,2,3, and the sample amount and the corresponding state grade in each state set in the training data are shown in a table 3;
calculating the OCI:
OCI＝AE+PE+5*SP+3*EC+ME
wherein AE represents aeration energy consumption, PE represents pump energy consumption, SP represents sludge production, EC represents external carbon source consumption, and ME represents mixing energy consumption;
step three: calculating mutual information I (x) of variables 1 ,x 2 )：
Wherein H (X) and H (Y) are information entropies of variables X and Y, H (X, Y) represents the joint entropy of the variables X and Y, and p (X, Y) represents the joint probability density function of X and Y; wherein H (X), H (Y) and H (X, Y) are calculated as follows:
H(X)＝-∫p(x)logp(x)dx
H(Y)＝-∫p(y)logp(y)dy
H(X,Y)＝-∫∫p(x,y)logp(x,y)dxdy；
step four: setting a threshold value of mutual information to be 0.6, and partitioning variables according to mutual information rules to obtain ERS and EIS;
step five: establishing an evaluation statistic of ERS by utilizing canonical correlation analysis, and establishing an evaluation statistic of EIS by utilizing slow characteristic analysis;
step six: constructing an online window;
using the size of the online window described herein H =10;
step seven: the first step is that 38 variables shown in the table 1 are selected as process variables, a constant water inlet file is used as water inlet data, test data of one week are collected, three simulated operation conditions are shown in the table 4, and the process data are divided into three data sets by using the GMM;
step eight: calculating mutual information by using the method in the third step, and dividing ERS and EIS of online data;
step nine: constructing evaluation statistics of ERS and EIS by adopting the method in the fifth step;
step ten: according to Bayes theory, the online statistics of the two subspaces are fused to determine the most possible state grade of the online data, and the online evaluation of the process running state is realized.
If the evaluation index is greater than the predetermined threshold α (0.95 in this embodiment), the BIC is determined l (k) And alpha is greater than alpha, the online data at the moment k is in the l type operation state level, otherwise, the online data is in other operation state levels.
Example 1 the results of the experiments for evaluating the operating conditions of the sewage treatment process are shown in FIGS. 5A to 5C and Table 5. The process operation state evaluation experiment was performed by the conventional Canonical Correlation Analysis (CCA) and Slow Feature Analysis (SFA) in the art using the same training data and detection data as in example 1, and the experimental results are shown in fig. 6A to 6D, fig. 7A to 7D, and table 6.
Table 1: selection of variables
Table 2: operating conditions of training data
Table 3: run state rankings for training data
| Group number | Status rating | Rating label | Sample interval | OCI mean 10 8 |
| |
Youyou (an instant noodle) | 1 | 1-489 | 1.54 |
| Group 2 | In (1) | 2 | 490-965 | 1.62 |
| Group 3 | Difference (D) | 3 | 966-1344 | 1.74 |
Table 4: operating conditions of test data
Table 5: on-line identification result of test data running state grade
| Group number | Status rating | Rating label | Sample section before evaluation | Sample section before |
| Group | ||||
| 1 | |
1 | 1-193 | 1-199 |
| Group 2 | In | 2 | 194-385 | 244-397 |
| Group 3 | Difference between | 3 | 386-672 | 434-672 |
Fig. 6A to 6D and fig. 7A to 7D are online evaluation result diagrams obtained by respectively performing process running state evaluation by using the canonical correlation analysis method and the slow feature method according to embodiment 1, where 6A, 6B, 6C, 7A, 7B, and 7C respectively represent similarities between online data and each state level, an upper dotted line represents a threshold of 0.95, and the total evaluation result can be seen in fig. 6D and fig. 7D, it can be seen that the state levels of the test data still change from good to medium to poor in the evaluation of SFA and CCA, but under the limitation of the same threshold, the SFA and the CCA methods cannot correctly evaluate all data belonging to the state levels.
Table 6 compares the evaluation results of the three methods, and it is found by comparison that the proposed method has a more accurate evaluation result when the state rank is middle. The SFA and the CCA cannot be provided independently, and meanwhile, the fact that the process is distinguished correctly from the economic information and the dynamic information in the process is also shown to be beneficial to improving the sensitivity of the process to the change of economic indexes, the internal characteristics of the process in each state grade can be better mined, the process running state can be reflected more truly and accurately, and effective evaluation information is provided for operators.
Table 6: run state ranking results for training data
While, for purposes of simplicity of explanation, the methodologies are shown and described as a series of acts, it is to be understood and appreciated that the methodologies are not limited by the order of acts, as some acts may, in accordance with one or more embodiments, occur in different orders and/or concurrently with other acts from that shown and described herein or not shown and described herein, as would be understood by one skilled in the art.
The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples and designs described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (13)
1. A state evaluation method for a sewage treatment process is characterized by comprising the following steps:
acquiring online process data according to a plurality of process variable dimensions involved in the sewage treatment process, and acquiring corresponding online economic index values;
calculating mutual information between each process variable dimension and the online economic index value so as to divide online process data of each process variable dimension into an ERS variable space and an EIS variable space, wherein the ERS variable space is related to economic performance, and the EIS variable space is not related to economic performance;
performing regular correlation analysis on the online process data in the ERS variable space, and extracting economic information in the online process data to construct ERS evaluation statistics;
performing slow characteristic analysis on the online process data in the EIS variable space, and extracting dynamic information in the online process data to construct EIS evaluation statistics; and
and fusing the ERS evaluation statistic and the EIS evaluation statistic according to Bayesian theory, and determining an operation state label corresponding to the online economic index value to determine the operation state of the sewage treatment process.
2. The state evaluation method according to claim 1, wherein the sewage treatment process involves a plurality of reaction tanks including an anoxic tank, an aerobic tank, and a secondary sedimentation tank,
the multiple process variable dimensions are selected from BOD total amount of effluent, COD total amount of effluent, pump energy consumption, effluent flow rate, internal reflux flow rate, soluble biodegradable organic nitrogen of effluent and effluent
Concentration, effluent Kjeldahl nitrogen concentration, effluent dissolved oxygen concentration, effluent nitrate concentration, effluent nitrite concentration, effluent granular biodegradable organic nitrogen, dissolved oxygen content of each reaction tank, nitrate and nitrite concentration of each reaction tank, and/or>
At least one of concentration, soluble biodegradable organic nitrogen in each reaction tank, particulate biodegradable organic nitrogen in each reaction tank, and aeration intensity of last reaction tank.
3. A state evaluation method according to claim 1, wherein the step of obtaining on-line process data according to a plurality of process variable dimensions involved in the sewage treatment process and obtaining corresponding on-line economic indicator values comprises:
constructing an online data window; and
and acquiring multiple groups of online process data related to the multiple process variable dimensions and corresponding online economic index values thereof according to the size H of the online data window.
4. A state evaluation method according to claim 1, wherein after acquiring on-line process data according to a plurality of process variable dimensions involved in the sewage treatment process and acquiring corresponding on-line economic indicator values, the state evaluation method further comprises the steps of:
standardizing multiple groups of on-line process data according to the following formula to construct a zero-mean unit standard deviation matrix X m×n ：
Wherein x is ij Is the mean unit standard deviation matrix X m×n M is the group number of the online process data, and n is the process variable dimension number in the online process data.
5. A state estimation method according to claim 1, wherein mutual information between each of said process variable dimensions and said online economic indicator value is calculated by the following formula:
wherein, H (X) and H (Y) are information entropies of variables X and Y, H (X, Y) represents the joint entropy of the variables X and Y, p (X, Y) represents the joint probability density function of X and Y, and H (X), H (Y) and H (X, Y) are respectively calculated by the following formula:
H(x)＝-∫p(x)logp(x)dx
H(Y)＝-∫p(y)logp(y)dy
H(X，Y)＝-∫∫p(x，y)logp(x，y)dxdy。
6. a state evaluation method according to claim 5, wherein said step of calculating mutual information between each of said process variable dimensions and said online economic indicator value to partition online process data for each of said process variable dimensions into an ERS variable space and an EIS variable space comprises:
comparing a mutual information value MI between each process variable dimension and the online economic index value with a preset first threshold value th; and
if the mutual information value MI is larger than the first threshold value th, dividing the corresponding process variable dimension into an ERS variable space X E On the contrary, if the mutual information value MI is less than or equal to the first threshold value th, the corresponding process variable dimension is divided into the EIS variable space X D Wherein the total variable space X = [ X = D ，X E ]。
7. A state estimation method according to claim 6, wherein said step of performing canonical correlation analysis on the online process data in the ERS variable space to extract economic information therein to construct an ERS estimation statistic comprises:
constructing the ERS evaluation statistic according to the following formula
Wherein, the first and the second end of the pipe are connected with each other,
and &>
Is N l，E The identity matrix of the dimension(s),N l，E indicates the number of variables in the ERS variable space in the status level of type I, <' > H>
Is ERS variable space->
One sample data of (1), J l And L l Respectively representing ERS variable space in the first state level and typical related variable, sigma, corresponding to output data l A matrix of singular values representing the ERS variable space within the l-th state class.
8. The state evaluation method of claim 7, wherein the step of performing slow feature analysis on the online process data in the EIS variable space to extract dynamic information therein to construct the EIS evaluation statistic comprises:
constructing the EIS evaluation statistic according to the following formula
Wherein the content of the first and second substances,
is the EIS variable space>
Wl is a coefficient matrix in EIS variable space within the l-th state level.
9. The state evaluation method according to claim 8, wherein the step of determining the operation state label corresponding to the online economic indicator value by fusing the ERS evaluation statistic and the EIS evaluation statistic according to the bayesian theory to determine the operation state of the sewage treatment process comprises:
let C l Indicating the l-th operating state, the on-line process data x k Is expressed as
And x is calculated by the following formula k Probability within class i operating state:
wherein, pr (C) l ) Indicates that the current process state is at C l The probability of (a) of (b) being,
indicating that the current process state is not at C l Probability of (e), pr [ g l (x k )|C l ]And &>
Can be calculated from the following equation:
wherein，
Offline evaluation criterion which represents the status rating of the l < th > species >>
And
and obtaining global evaluation statistics in different state levels by Bayesian inference:
if BIC l (k) If the current time is greater than the preset second threshold value alpha, the online process data at the moment k is judged to be in the class of the l-th running state, otherwise, if the current time is greater than the preset second threshold value alpha, the online process data at the moment k is judged to be in the class of the l-th running state, and if the current time is not greater than the preset second threshold value alpha, the online process data at the moment k is judged to be in the class of the l-th running state l (k) And if the current value is less than or equal to the second threshold value alpha, the online process data at the moment k is judged to be in other operation state grades.
10. The state evaluation method according to claim 1, wherein before the ERS evaluation statistic and the EIS evaluation statistic are fused according to bayesian theory to determine the operation state label corresponding to the online economic index value so as to determine the operation state of the sewage treatment process, the state evaluation method further comprises the following steps:
selecting a plurality of process variables involved in the sewage treatment process, and acquiring a plurality of groups of off-line process data and off-line economic index values corresponding to the off-line process data according to the process variables;
classifying each group of the off-line process data into a plurality of state levels related to economic benefit indexes by using a Gaussian mixture model;
calculating mutual information between each process variable dimension and the offline economic index value, and partitioning the process variables according to the mutual information to establish the ERS variable space and the EIS variable space;
performing regular correlation analysis on the ERS variable space, and extracting economic information in the ERS variable space to construct ERS evaluation statistics; and
and performing slow characteristic analysis on the EIS variable space, and extracting dynamic information in the EIS variable space to construct EIS evaluation statistics.
11. A state estimation method according to claim 10, wherein said step of classifying each set of said off-line process data into a plurality of state classes with respect to economic benefit indicators using a gaussian mixture model comprises:
selecting an overall cost index OCI as the economic benefit index, and calculating the economic benefit index according to the following formula:
OCI＝AE+PE+5*SP+3*EC+ME
wherein AE represents aeration energy consumption, PE represents pump energy consumption, SP represents sludge production, EC represents external carbon source consumption, and ME represents mixing energy consumption;
differentiating, by means of clustering, multiple operating state levels { X } of the multiple sets of off-line process data from a probabilistic perspective using the Gaussian mixture model l ，y l L =1, 2.., L, wherein, X l And y l Respectively representing the process data and the economic indicator OCI within the state set l,
N l representing the number of samples in the state set l;
according to
Determining a rank label for each state set in which>
The better the ground represents the behavior of the state set l, the better its corresponding class label.
12. A state evaluation system of a sewage treatment process, characterized by comprising:
a memory having computer instructions stored thereon; and
a processor coupled to the memory and configured to execute computer instructions stored on the memory to implement the method of assessing the status of a wastewater treatment process according to any of claims 1-11.
13. A computer-readable storage medium having stored thereon computer instructions, wherein the computer instructions, when executed by a processor, implement the method for evaluating a state of a sewage treatment process according to any one of claims 1 to 11.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310183342.0A CN115983534A (en) | 2023-03-01 | 2023-03-01 | Method and system for evaluating state of sewage treatment process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310183342.0A CN115983534A (en) | 2023-03-01 | 2023-03-01 | Method and system for evaluating state of sewage treatment process |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115983534A true CN115983534A (en) | 2023-04-18 |
Family
ID=85962678
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310183342.0A Pending CN115983534A (en) | 2023-03-01 | 2023-03-01 | Method and system for evaluating state of sewage treatment process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115983534A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808216A (en) * | 2024-03-01 | 2024-04-02 | 四川省铁路建设有限公司 | Energy saving and emission reduction effect evaluation method for sewage treatment |
-
2023
- 2023-03-01 CN CN202310183342.0A patent/CN115983534A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808216A (en) * | 2024-03-01 | 2024-04-02 | 四川省铁路建设有限公司 | Energy saving and emission reduction effect evaluation method for sewage treatment |
| CN117808216B (en) * | 2024-03-01 | 2024-05-07 | 四川省铁路建设有限公司 | Energy saving and emission reduction effect evaluation method for sewage treatment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106355030B (en) | A kind of fault detection method based on analytic hierarchy process (AHP) and Nearest Neighbor with Weighted Voting Decision fusion | |
| CN110379463B (en) | Marine algae cause analysis and concentration prediction method and system based on machine learning | |
| CN111160776A (en) | Method for detecting abnormal working condition in sewage treatment process by utilizing block principal component analysis | |
| WO2019071384A1 (en) | Intelligent systems and methods for process and asset health diagnosis, anomoly detection and control in wastewater treatment plants or drinking water plants | |
| CN109685366A (en) | Equipment health state evaluation method based on mutation data | |
| Cheng et al. | Monitoring influent conditions of wastewater treatment plants by nonlinear data-based techniques | |
| CN111626821B (en) | Product recommendation method and system for realizing customer classification based on integrated feature selection | |
| CN109085805B (en) | Industrial process fault detection method based on multi-sampling-rate factor analysis model | |
| CN110033141B (en) | Method for establishing desulfurization system operation condition database | |
| CN112446597B (en) | Storage tank quality assessment method, storage tank quality assessment system, storage medium, computer equipment and application | |
| CN112904810B (en) | Process industry nonlinear process monitoring method based on effective feature selection | |
| CN115983534A (en) | Method and system for evaluating state of sewage treatment process | |
| CN114764682B (en) | Rice safety risk assessment method based on multi-machine learning algorithm fusion | |
| CN104376370A (en) | Dam horizontal displacement prediction method | |
| Atukeren | The relationship between the F-test and the Schwarz criterion: implications for Granger-causality tests | |
| CN115600102B (en) | Abnormal point detection method and device based on ship data, electronic equipment and medium | |
| CN116661410A (en) | Large-scale industrial process fault detection and diagnosis method based on weighted directed graph | |
| Bongiorno et al. | Evaluating the complexity of some families of functional data | |
| ATKINSON | Fitting mixtures of regression lines with the forward search | |
| CN112733903B (en) | SVM-RF-DT combination-based air quality monitoring and alarming method, system, device and medium | |
| Wakiru | A decision tree-based classification framework for used oil analysis applying random forest feature selection | |
| Himani et al. | A comparative study on machine learning based prediction of citations of articles | |
| Ren et al. | Fault detection and process monitoring of industrial process based on spherical kernel T-PLS | |
| CN114266401B (en) | Chassis engine fault prediction method and system | |
| CN115423148B (en) | Agricultural machinery operation performance prediction method and device based on Ke Li jin method and decision tree |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |