CN106599568A - Abnormal temperature diagnosis device for greenhouse - Google Patents
Abnormal temperature diagnosis device for greenhouse Download PDFInfo
- Publication number
- CN106599568A CN106599568A CN201611132830.5A CN201611132830A CN106599568A CN 106599568 A CN106599568 A CN 106599568A CN 201611132830 A CN201611132830 A CN 201611132830A CN 106599568 A CN106599568 A CN 106599568A
- Authority
- CN
- China
- Prior art keywords
- abnormal temperature
- temperature
- node
- abnormal
- symptom
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses an abnormal temperature diagnosis device for a greenhouse, comprising an abnormal temperature compiling device, a management module, an observation operation module, a probability estimation module and an analysis diagnosis module. By adopting an intelligent method, the polling operation on the abnormal temperature of heating or cooling equipment or sensors is reduced, and the judgment time is shortened; and the diagnosis is implemented by adopting a more intelligent method.
Description
Technical field
The present invention relates to a kind of agricultural system design field, is related to a kind of mode by using adjustment to realize intelligent temperature
The intensification of room or the apparatus for diagnosis of abnormality of cooling system or sensor.
Background technology
For traditional greenhouse is used as an ecosystem, the abnormal temperature diagnosis first step is the state for obtaining greenhouse greenhouse
The acquisition of abnormal temperature information, i.e. thermal index.Note abnormalities temperature, needs to collect various greenhouse abnormal temperature information.One
As for, the temperature in greenhouse there occurs it is abnormal refer to that greenhouse respectively heats up or cooling system is in abnormal state of affairs, by obtaining
Equipment and various sensor-services and the abnormal state temperature information of application, it is possible to find the exception temperature occurred in greenhouse in time
Degree problem.Collecting abnormal temperature typically has two methods:One kind is to report to the police, when there is abnormal temperature, by generation abnormal temperature
Object is actively to the report of abnormal temperature management system；Another kind is active poll, abnormal warm by the inquiry of abnormal temperature system at regular intervals
The state of degree object.It is actively a kind of largely effective to the report of abnormal temperature management system from the greenhouse object of generation abnormal temperature
Abnormal temperature discovery mechanism, can find in time heat up or cooling system abnormal temperature, connection failure, intensifications or lower the temperature set
It is standby to restart, do not receive intensification or the response of cooling system, heat up or cooling system process exception isothermal chamber abnormal temperature
Or critical event, and limited communication frequency is only needed to, but this method is not always reliable, for example, if one rises
Temperature or cooling system or sensor there occurs the severely subnormal temperature such as power-off, and it can not send event, connected connection
Interruption can also prevent the transmission of relevant abnormalities temperature information.In these cases, it is necessary to manage exception using some devices
Temperature management system poll can help abnormal temperature management system by the method for pipe abnormal temperature object, the method for active poll
The abnormal temperature abnormal temperature having found that it is likely that.
The main object of the present invention be by a kind of greenhouse abnormal temperature diagnostic equipment, using intelligentized method, from
And reduce to heat up or cooling system or sensor abnormal temperature wrap count operation, reduce the judgement time,
And by going to realize using more intelligent method.Therefore, it can be said that by it is this be achieved in that it is necessary.
The content of the invention
In view of this, the technical problem to be solved in the present invention is to provide a kind of greenhouse abnormal temperature diagnostic equipment, for solving
The problem that certainly temperature intelligent in multiple greenhouses cannot be measured and controlled.It is of the invention to reach the effect of above-mentioned technical proposal
Technical scheme is：A kind of greenhouse abnormal temperature diagnostic equipment, including：
1. abnormal temperature compilation device, management module, observation operation module, probability Estimation module, analyzing and diagnosing module；
It is that event is generated one that abnormal temperature compilation device processes these methods for being caused symptom event by abnormal temperature
The Quick Response Code of individual mark abnormal temperature, the process of correlation analysiss is exactly the process being interpreted to symptom event, final to determine
Those abnormal temperatures that Quick Response Code is identified, that symptom event occur；
The mode of operation of abnormal temperature compilation device is divided into two stages；In the choice phase of two-dimentional code book, select to intercept
Event subset, the result of this process is to generate two-dimentional code book, and two-dimentional code book is optimum event subset, including making an uproar
The event that problem must be monitored is distinguished in acoustic capacitance mistake scope；In Greenhouse System, the change of abnormal temperature may be lost or postpone
Reach, the temperature warning of forgery can be also produced in greenhouse, even if this requires relevance algorithms noisy feelings in flow of event
Remain to correctly identify abnormal temperature under condition；In decoding stage, the event in the two-dimentional code book of monitoring is analyzed in real time,
Exactly by by symptom and observe symptom do it is closest match, find final abnormal temperature reason；
The basic module being connected with abnormal temperature compilation device is management module, and management module adapts to Greenhouse System and opens up automatically
The change of structure, the scope in greenhouse and complexity is flutterred, event in domain can be managed automatically, determine the mutual relation of event,
Call the predefined problem of management personnel, find cause symptom abnormal temperature, management module collect cooperate therewith other
Management module abnormal temperature information in abnormal temperature compilation device, reaches the mesh for safeguarding whole greenhouse LAN temperature jointly
's；
Using temperature as a temperature nodes variable, temperature nodes variable can be divided into following a few classes to management module:It is abnormal
Temperature symptom node set, abnormal temperature assume node set, observation running node set, diagnostic operation node set and other
Node；Other nodes refer to the set more than in addition to node；The value of temperature nodes variable is all finite discrete, different
There is significantly difference between value, it is impossible to ambiguous；Generally represent normal with state 0 (first statement value), other shapes
State is relevant with abnormal temperature；Value is normal (state 0) and two kinds of abnormal temperature (state 1) as abnormal temperature assumes node
State；Abnormal temperature symptom node refers to the node of greenhouse abnormal temperature symptom, is set up according to different abnormal temperature symptoms
Corresponding diagnostic method；After abnormal temperature symptom is detected, starting corresponding diagnostic cast carries out diagnostic reasoning, and by difference
The diagnostic cast of abnormal temperature symptom is integrated, is found with a diagnostic module；Abnormal temperature is assumed each in node table temperature indicating room
Different abnormal temperature patterns are planted, is that it defines diagnostic operation, if it is not for the abnormal temperature pattern of greenhouse managed object
Final abnormal temperature reason, then be considered profound abnormal temperature symptom by the abnormal temperature pattern；
Observation operation module is used for the Different Results of expression observation operation, and observation operation module acts on some operation section
Point is used to confirm that certain abnormal temperature whether there is, or the abnormal temperature information for obtaining and abnormal temperature hypothesis node is related is come
Source, it is the main canal for obtaining greenhouse abnormal temperature information to represent external observation phenomenon, result of test operation, observation running node
Road；
Probability Estimation module is used for for observation running node, observes the probability of running node value result, is used for
Judge various observation operating results；
Analyzing and diagnosing module reaches the Different Results of diagnostic operation in diagnostic operation node table, and the Different Results of diagnostic operation are led to
It is often that can operation determine specific certain abnormal temperature；Diagnostic operation node is expressed as abnormal temperature hypothesis in diagnostic cast
The child node of variable, diagnostic operation node estimates that mode is consistent with the conditional probability of observation running node, so as to be diagnosed to be
The probability of abnormal temperature；
Analyzing and diagnosing module in diagnosis when diagnosing to greenhouse according to greenhouse abnormal temperature symptom, seek most to have
Possible abnormal temperature reason, then carries out diagnostic operation test, and if not the abnormal temperature reason exception temperature is then excluded
Degree is possible, other possible abnormal temperature reasons is continued to look for, until eventually finding abnormal temperature reason.
Description of the drawings
Fig. 1 is a kind of structural representation of greenhouse abnormal temperature diagnostic equipment of the invention.
Specific embodiment
In order that the technical problem to be solved, technical scheme and beneficial effect become more apparent, below tie
Drawings and Examples are closed, the present invention will be described in detail.It should be noted that specific embodiment described herein is only used
To explain the present invention, it is not intended to limit the present invention, the product that can realize said function belongs to equivalent and improvement, includes
Within protection scope of the present invention.Concrete grammar is as follows：
Embodiment one：As shown in Figure 1 in practice, with the infiltration of technology of Internet of things, production-scale expansion is given birth to greenhouse
Producing the monitoring of overall process, the demand of control technique increasingly increases, between weighed, the speed that abnormal temperature finds is faster, needs
The greenhouse bandwidth to be taken is bigger.First be that the abnormal temperature information to acquisition is filtered, i.e., abnormal abnormal temperature letter
Breath is filtered.There is substantial amounts of abnormal temperature information redundancy in greenhouse network, to same abnormal temperature, certain sensor connects
Repeatedly alarm abnormal temperature information can be continuously transmitted after receiving, or it not is abnormal temperature management to send certain equipment for alerting
The key equipment that system is concerned about, in this case, abnormal temperature message-filter mechanism needs analysis of history to note down, and filters these
Redundancy abnormal temperature information, for the diagnosis of further abnormal temperature more valuable abnormal temperature information is provided.Greenhouse is extremely warm
Here we are set as mainly comprising some important rings such as abnormal temperature information processing, reasoning and diagnostic operation suggestions for degree diagnosis
Section.In Greenhouse operation abnormal state, by methods such as logic, model, decision-making theory and artificial intelligences, effectively combine and observe
The various abnormal temperature information relevant with greenhouse abnormal temperature, provide possible abnormal temperature and assume and diagnostic operation suggestion,
Diagnostic operation is performed, real abnormal temperature reason is found, diagnostic task terminates, and otherwise continues to search for new abnormal temperature information,
And other possible abnormal temperatures are excluded it is assumed that until finding abnormal temperature reason.From for other side, in greenhouse exception
Substantial amounts of original observed data or historical accumulation abnormal temperature letter may be by all means obtained in the practice of temperature diagnostic
Breath, these data or abnormal temperature information are often multi-source foreign peoples, and conditionally interdepend or depend on various controls
Factor processed.Meanwhile, during abnormal temperature diagnosis, the abnormal temperature information relevant with greenhouse abnormal temperature is a lot, such as some
How necessary dependent observation operation, enquirement, operation cost etc., effectively express above-mentioned numerous multi-source foreign peoples and relevant abnormalities
Temperature information, meanwhile, these abnormal temperature information how are excavated, the result of decision of most worthy is obtained, it is abnormal temperature diagnosis
The key issue for being faced.
Additionally, occur abnormal temperature in greenhouse inevitably brings huge economic loss to society and enterprise, how
Loss is reduced to into minimum degree, abnormal temperature is accurately found and positioned with minimum cost and shortest time, recover greenhouse just
Often operation is the abnormal temperature diagnosis another key issue to be solved.Therefore, the subject matter that abnormal temperature diagnosis in greenhouse faces
It is how by setting up the efficient abnormal temperature diagnostic equipment, obtains different from uncertain, multi-source foreign peoples's abnormal temperature information
The relevant abnormalities temperature information of normal temperature, accurately finds and positions abnormal temperature with minimum cost and shortest time.
The implementation procedure of abnormal temperature compilation device is：The generation of each abnormal temperature can cause a large amount of abnormal temperature diseases
The generation of shape event, the abnormal temperature symptom event that each object is produced is probably what the object itself abnormal temperature problem caused
Local event, it is also possible to which abnormal temperature occurs in relevant sensor object, is propagated through the event come and cause.Process
It is that event is generated a Quick Response Code for identifying abnormal temperature, dependency that these are caused the method for symptom event by abnormal temperature
The process of analysis is exactly the process being interpreted to symptom event, it is final determine Quick Response Code mark, there is symptom event that
A little abnormal temperatures.Technique of compiling is divided into two stages.In the choice phase of two-dimentional code book, the event subset intercepted is selected, this
The result of process is to generate two-dimentional code book, and two-dimentional code book is optimum event subset, including distinguishing in noise error tolerance
The event that other problem must be monitored.In Greenhouse System, the change of abnormal temperature may be lost or postpone to reach, and can also in greenhouse
The temperature warning forged is produced, even if this requires to remain to correctly know in the case of relevance algorithms are noisy in flow of event
Other abnormal temperature.In decoding stage, the event in the two-dimentional code book of monitoring is analyzed in real time, that is, by by symptom or
" compiling " with observe symptom do it is closest match, find final abnormal temperature reason.
The basic module of abnormal temperature compilation device is management module, automatically adapts to Greenhouse System topological structure, greenhouse
The change of scope and complexity, can be managed to event in domain automatically, determine the mutual relation of event, call management personnel pre-
The problem for first defining, finds the abnormal temperature for causing symptom, and management module collects other abnormal temperature compiling dresses cooperated therewith
Management module abnormal temperature information in putting, reaches the purpose for safeguarding whole greenhouse LAN temperature jointly.
Using temperature as a temperature nodes variable, such node can be divided into following a few classes to management module:Abnormal temperature
Node set, observation running node set, diagnostic operation node set and other sections are assumed in symptom node set, abnormal temperature
Point.Other nodes refer to the set more than in addition to node, for example for ease of the construction of greenhouse diagnostic cast, probability inference with
Decision-making application and auxiliary node for introducing etc..The value of all node variables is all finite discrete, has bright between different values
Aobvious difference, it is impossible to ambiguous.Generally with state 0 (first statement value) expression normally, other states and abnormal temperature
It is relevant.As abnormal temperature assumes node value for normal (state 0) and abnormal temperature (state 1) two states (hereinafter referred to as
For two state of value), if being not limited to a kind of abnormal temperature state, can be with the expression such as state 2, state 3.Abnormal temperature symptom section
Point abnormal temperature symptom node refers to the node of greenhouse abnormal temperature symptom, is typically set up according to different abnormal temperature symptoms
Corresponding diagnostic method.After abnormal temperature symptom is detected, starting corresponding diagnostic cast carries out diagnostic reasoning, it is also possible to will
The diagnostic cast of different abnormal temperature symptoms is integrated, is found with a diagnostic module.Abnormal temperature assumes that node expresses temperature
A variety of abnormal temperature patterns in room, for the abnormal temperature pattern of greenhouse managed object, to which define diagnostic operation,
If it is not final abnormal temperature reason, the abnormal temperature pattern is considered into profound abnormal temperature symptom, machine
Is classified and the node variable for expressing this abnormal temperature pattern is referred to as middle abnormal temperature hypothesis node.Final abnormal temperature reason
It is the diagnosis decision-making target finally to be solved, the node variable for expressing final abnormal temperature pattern is referred to as into general abnormal temperature false
If node.When especially not emphasizing, abnormal temperature is assumed to be two-valued variable, i.e., be normally false (state 0), and abnormal temperature is
True (state 1).
Observation operation module is used for the Different Results of expression observation operation, and observation operation module acts on some operation section
Point is used to confirm that certain abnormal temperature whether there is, or the abnormal temperature information for obtaining and abnormal temperature hypothesis node is related is come
Source, it is the main of acquisition greenhouse abnormal temperature information that can represent external observation phenomenon, result of test operation, observation running node
Channel.
Probability Estimation module is used for for observation running node, and the estimation to its probability is exactly to estimate when given observation
When the father node of running node is closed, the probability of running node value result is observed, for judging various observation operating results.Diagnosis
Running node expresses the Different Results of diagnostic operation, the Different Results of diagnostic operation be typically operation can determine it is specific certain
Abnormal temperature.Diagnostic operation node is expressed as the child node that abnormal temperature assumes variable, diagnostic operation node in diagnostic cast
Estimate that mode is consistent with the conditional probability of observation running node.For the estimation of diagnostic operation node condition probability is exactly to estimate
Count when the father node of given running node is closed, the probability P of the abnormal temperature can be diagnosed to be, wherein taking Succeed represents different
Normal temperature f is diagnosed generation really, takes fail and represents that abnormal temperature f does not occur, under this conditional probability P, P
(correct) the correct probability for performing of operation is represented, P (req) represents the probability that operation is smoothly performed as required.
In probability Estimation, estimate that 3 probits on the right of equation are easier than direct estimation respectively.Different nodes have
Different attributes, in order to need the diagnosis cost problem of consideration when expressing diagnosis decision-making, to observer nodes, diagnostic operation node all
Cost attribute is given, therefore this two classes node can be referred to as running node, the cost of running node referred to as operates cost.
Analyzing and diagnosing module, we introduce conditional independence assumption, abnormal temperature diagnostic field personnel in diagnostic cast
Thinking when diagnosing to greenhouse often according to greenhouse abnormal temperature symptom, seeks most possible abnormal temperature former
Cause, then carries out diagnostic operation test, and if not the abnormal temperature reason abnormal temperature possibility is then excluded, and continues to look for it
Its possible abnormal temperature reason, until eventually finding abnormal temperature reason.And. often exist between abnormal temperature related
Relation, this dependency relation will be helpful to abnormal temperature diagnosis, and in addition analyzing and diagnosing module carries out probability Estimation, especially suitable
In the situation that the prior probability assumed abnormal temperature is not clear.If using cause and effect direction it is assumed that wherein directed edge side
To expressing a kind of direct cause effect relation (or dependence).Assume many in the abnormal temperature for causing certain abnormal temperature symptom
When, the model structure in cause and effect direction can be estimated to make troubles to the conditional probability of node, such as the section for having 4 father nodes
Point, in each father node two-value state node is, then when estimating node condition probability, need the conditional probability value estimated to reach
24-1=15.Therefore, in the conditional probability of cause and effect direction model structure is estimated, it usually needs hypothesis causes abnormal temperature disease
Each abnormal temperature of shape assumes independent of one another, i.e. cause and effect independence assumption, and each abnormal temperature assumes that state is independent of each other, because
The model structure in fruit direction is suitable for obtaining the situation of the prior probability of each abnormal temperature reason greenhouse.Analyzing and diagnosing
Module needs conditional probability table of each node with the state value of its father node as condition, and making diagnostic cast should complete as follows
Work:Determine greenhouse node variable.Comprising following content:Determine that greenhouse abnormal temperature symptom, determination have with abnormal temperature symptom
The possibility abnormal temperature reason of the pass key element relevant with abnormal temperature reason, and they are expressed as into node variable；It is determined that all
The value set of node variable, the set should be included should be had between the value that is possible to of node variable, and different values
Obvious repellence (objectionable intermingling)；Determine running node (running node containing observation and diagnostic operation node) and operation cost.
Set up the directed acyclic graph for representing mutual relation between node variable.Diagnosis decision-making is a causal reasoning process, it is believed that
Directed edge in diagnostic cast between connecting node expresses a kind of cause effect relation.Conditional probability is estimated.It is every in for model
One node, it is necessary to give corresponding conditional probability, the node without father node then needs given prior probability.Abnormal temperature
The process that diagnostic cast is built is an abnormal temperature analysis process first, obtains greenhouse abnormal temperature pattern and its relevant abnormalities
Temperature information, next to that the conditional probability of the dependence set up between node and node is estimated.The operation of analyzing and diagnosing module
Need the determination to node variable and its relation.The factor relevant with abnormal temperature is numerous in greenhouse, especially in complicated greenhouse
The relation of these factors is intricate, and abnormal temperature pattern is obscured, it is difficult to determine the mutual relation between various factors completely.And
And to estimate conditional probability in advance.Many greenhouse abnormal temperature sample acquisitions are difficult, or even cannot obtain complete exception at all
Temperature samples, cause conditional probability to be estimated difficult.The work of analyzing and diagnosing module is probably in practice to intersect to carry out repeatedly, and with
The change for practical application is updated and perfect, may be obtained by all means in the practice of greenhouse abnormal temperature management
Substantial amounts of original observed data or historical accumulation abnormal temperature information, these abnormal temperatures diagnosis abnormal temperature information have fuzzy
Property, inexactness, reflect that the state behavior of complex diagnostics object is possibly incomplete with single abnormal temperature information.In a large number
Practice also indicate that the abnormal temperature feature abnormalities temperature information for obtaining is more, more accurate, and abnormal temperature diagnosis capability is got over
By force.Therefore, it is to improve abnormal temperature diagnosis capability, should under the limited conditions as much as possible using the various exception temperature in greenhouse
Degree information.On the other hand, in the diagnosis of greenhouse abnormal temperature, available abnormal temperature information is a lot, due to needing monitoring
Object and need abnormal temperature to be processed are more, and system needs abnormal temperature quantity of information to be processed very huge, it is generally the case that
We cannot obtain all one abnormal temperature information related to greenhouse abnormal temperature, all of different collecting in practice
Normal temperature information, in the face of numerous association abnormal temperature information, using abnormal temperature sort module abnormal temperature classification is carried out：Point
Generic task is the content according to data exception temperature information to be sorted come the classification belonging to determining the data, classification process point
Solve as two aspects of training and classify, in the training stage, need to learn the data exception temperature information of known class, root
The corresponding characteristic attribute of classification extraction belonged to according to them, obtains grader, in sorting phase, is classified using the training stage
Device is analyzed and calculates to data exception temperature information to be sorted, obtains final classification results.
Embodiment two：As shown in Figure 1 in practice, the complexity, the feature of erratic behavior due to greenhouse itself, greenhouse exception
Temperature classifications processing procedure is related to many comprehensive knowledges, wants to obtain good classifying quality, and training sample is uneven,
Sorting technique first has to be trained study, and its training process needs a number of training sample.And wrap in Greenhouse System
Contain substantial amounts of uncertain factor, nonlinear mapping relation is there may be between abnormal temperature symptom and abnormal temperature source, it is same
Planting abnormal temperature often has different performances, often same symptom and the coefficient result of several abnormal temperatures, Duo Gexiang
Closing abnormal temperature may occur simultaneously.Therefore, more training samples are selected, the representativeness of sample is stronger, training effect's just meeting
Better, classification performance could be higher.Generally, people become more readily available the sample class of some abnormal temperatures substantially
Not, however these training samples concentrate that the quality of samples is often uneven, the Performance simulation of some samples is all right one way or the other, some
Even wrong, if directly being learnt on these samples, the effect for making classification is substantially reduced.It is envisioned that to existing
A large amount of inaccurate classification samples ground study are to be difficult to obtain good classifying quality.And in actual use, in a large number
It is extremely difficult that high-quality training sample set ground is obtained, and it is also not to carry out screening to training sample by artificial method
Reality.As a example by support machine sort, support that machine sort is widely used sorting technique, energy preferably treatment classification is asked
Topic.But, support that machine sort has the weak defect of noise resisting ability, although supporting vector quantity is few, contain classification
Required all abnormal temperature information.Due to supporting that the classifying quality of machine sort is finally determined by the supporting vector of minority in sample
Fixed, most of training sample is not supporting vector, therefore removes or reduce the grader of part sample Liu one and do not affect.If
There is certain noise sample in the classifying space of the influence factor for holding machine sort abnormal temperature and be calculated as supporting vector, it is clear that
The nicety of grading for supporting abnormal temperature machine sort can be severely impacted.Therefore to realize good classifying quality, need to existing
Some training samples carry out necessary pretreatment, concentrate from initial training sample and remove noise sample, choose high-quality sample
This structural classification device.In addition, the key of structural classification device is to select suitable feature constitutive characteristic vector.It is abnormal in Greenhouse System
Greenhouse performance when temperature occurs has very big difference, and the feature for representing different abnormal temperatures is also not quite similar, and what is selected
The feature of sample carrys out constitutive characteristic vector, and the length of characteristic vector takes much, all affects to a certain extent under greenhouse management platform
The precision of abnormal temperature classification.By taking simple bayesian classifier method as an example, it is understood that Nave Bayesian Classifier method have it is simple and
The characteristics of nicety of grading is high, but if arbitrarily use arbitrary unit in training sample not only full as the one-dimensional of characteristic vector
Foot not requirement of the Nave Bayesian Classifier method to feature independence, and the moon. can greatly abnormal temperature nicety of grading.Therefore, I
Different characteristic vectors are chosen to different abnormal temperature type, constituting the characteristic component of abnormal temperature characteristic vector to the greatest extent will may be used
The feature of energy ground reflection abnormal temperature.In practical application, the frequency that we can occur according to feature in the sample of abnormal temperature
Rate chooses the spy of constitutive characteristic vector calculating P finally by calculating per class abnormal temperature to the posteriority probability of single feature
Levy component.Because different types of abnormal temperature has different features, it is necessary to define theirs for each class abnormal temperature
Characteristic vector.To support that machine sort is applied in abnormal temperature classification, can be made up in Bayesian Classification Arithmetic first in addition
In the case of testing knowledge deficiency, support that abnormal temperature grader still has higher classification accuracy rate, it is ensured that abnormal temperature diagnosis system
System has preferable performance.For classification problem, support that sample of the machine sort in classifying space calculates the classifying space
Decision-making curved surface, the classification of sample is determined by the curved surface.Process to simplify classification, be between conventional method hypothesis characteristic attribute
Consider the characteristic attribute being independent of each other when being independent of each other, that is, choose abnormal temperature eigenvalue as far as possible, only use phase
Mutual unrelated characteristic attribute, or assume that each characteristic attribute is independent of each other.Need to choose substantial amounts of spy in actually used
Attribute is levied, it is difficult to ensure that being unrelated between characteristic attribute.If using bayes method, abnormal temperature characteristic attribute ground is chosen
It is very difficult, it is separate between each attribute to be very easy in reality be run counter to, hinder abnormal temperature grader precision ground
Further improve.Determine feature of which attribute more suitable for abnormal expression temperature, not only need to count single candidate feature pair
In the impact of classifying quality, while also needing to consider the interactive abnormal temperature information between candidate feature.So can cause a large amount of
The loss of useful abnormal temperature information.Support machine sort be avoided that such case, it can effective process characteristic attribute it is unrelated
Sample, also can the related sample of processing feature attribute.In addition traditional abnormal temperature diagnostic method is based on empirical risk minimization
The principle of change, so easily lead to study, i.e. training error too small and cause the decline of generalization ability.Produced study existing
As the reason for, one is that sample is insufficient, and two is that the design of abnormal temperature disaggregated model is unreasonable, and the two problems are interrelated
's.For two class classification problems, according to Statistical Learning Theory, for the abnormal temperature disaggregated model for obtaining, traditional exception
Temperature diagnostic method selects the process that examination is gathered for abnormal temperature disaggregated model and algorithm, and the process can be understood as adjustment
The process of fiducial range.If being relatively adapted to existing sample, preferable effect can be obtained.But refer in default of theory
Lead, this selection can only rely on priori and skill, cause the dependence to user skill.However, due to without system
The method of property, even if the brilliant designer of a skill, the parameter that cannot guarantee that can obtain abnormal temperature every time all connects
Nearly optimum abnormal temperature diagnosis scheme.From the above analysis, traditional abnormal temperature sorting technique learnt based on data
Limitation root be that it is the theory based on empirical risk minimization, and the requirement of big-sample data is in reality in the theory
More difficult satisfaction in the use of border.Support that abnormal temperature adopts structural risk minimization principle, take into account training error and generalization ability,
The advantage for having uniqueness in Small Sample Database collection and nonlinear problem is solved, is particularly suitable for the classification of abnormal temperature.Improve and prop up
The effect that machine sort is classified to abnormal temperature is held, is needed to play as much as possible and is supported machine sort inherent advantages, made up and prop up
Machine sort is held for the deficiency in processing in classification.Support that machine sort does not have the standard of Feature Selection, and what is selected
Feature constituting the every one-dimensional of abnormal temperature characteristic vector, characteristic vector space dimension number, all strong influences
Hold the classifying quality of abnormal temperature grader.We are selected for classification most using Feature Selection strategy from candidate feature attribute
Important attribute, with this classifying space for supporting abnormal temperature is constituted.With being a difference in that for Bayes's classification, in selected characteristic
Bayes's classification needs to consider the essence that between characteristic attribute be separate, will likely otherwise affect abnormal temperature to classify during attribute
Degree.
The present invention can have the specific embodiment of various multi-forms, combine accompanying drawing by taking Fig. 1 as an example above to the present invention's
Technical scheme explanation for example, the present invention will be described in detail.It should be noted that specific embodiment described herein
Only to explain the present invention, it is not intended to limit the present invention, the product that can realize said function belongs to equivalent and improvement,
It is included within protection scope of the present invention.
The invention has the beneficial effects as follows：Existing temperature can be avoided ceaselessly by this greenhouse temperature data test device
The shortcoming that abnormal temperature information produces abnormal temperature information redundancy is sent, greenhouse thermometric time error rate is reduced, and is passed through
This greenhouse temperature data test device removes greenhouse temperature using more intelligent method of testing to realize.Therefore, it can succeed in reaching an agreement
Cross it is this be achieved in that it is necessary.
Claims (1)
1. a kind of greenhouse abnormal temperature diagnostic equipment, it is characterised in that：Including：Abnormal temperature compilation device, management module, observation
Operation module, probability Estimation module, analyzing and diagnosing module；
Reason abnormal temperature at the abnormal temperature compilation device and cause symptom event；First the symptom event is generated one
The individual Quick Response Code that can identify abnormal temperature, the process of the correlation analysiss of the abnormal temperature compilation device is to the symptom thing
The process that part is interpreted, final those abnormal temperatures determining Quick Response Code mark, that symptom event occur；
The mode of operation of the abnormal temperature compilation device is divided into two stages；In the choice phase of two-dimentional code book, select to intercept
Event subset, the result of this process is to generate the two-dimentional code book, and the two-dimentional code book is optimum event subset, wherein
It is included in noise error tolerance and distinguishes the symptom event that must be monitored；In decoding stage, in monitoring the two-dimentional code book
The symptom event, analyzed in real time, by by the symptom event with observe the symptom event do recently
Like matching, final abnormal temperature reason is found；
The basic module being connected with the abnormal temperature compilation device is the management module, and the management module adapts to temperature automatically
The change of chamber system topological structure, the scope in greenhouse and complexity, event is managed in automatic local area network, determines event
Mutual relation, calls the predefined problem of management personnel, finds the abnormal temperature for causing the symptom event, the management mould
Block collects the abnormal temperature information in the management module in other the described abnormal temperature compilation devices cooperated therewith, reaches
The purpose of whole greenhouse LAN temperature is safeguarded jointly；
Using temperature as a temperature nodes variable, the temperature nodes variable can be divided into following a few classes to the management module:
Abnormal temperature symptom node set, abnormal temperature assume node set, observation running node set, diagnostic operation node set and
Other nodes；Described other nodes refer to the set more than in addition to node；The value of the temperature nodes variable is limited
In discrete range, there is significantly difference between different values, it is impossible to ambiguous；Represent normal with state 0；Abnormal temperature is assumed
Node value is normal (state 0) and abnormal temperature (state 1) two states；The abnormal temperature symptom node refers to that greenhouse is different
The node of normal temperature symptom, according to different abnormal temperature symptoms corresponding diagnostic method is set up；The management module is in detection
To after the abnormal temperature symptom, carry out diagnostic reasoning, and by different abnormal temperature symptoms carry out it is integrated, so as to different to find
The reason for normal temperature occurs；The abnormal temperature assumes a variety of abnormal temperature patterns in node table temperature indicating room, for temperature
The abnormal temperature pattern of room managed object, is that it defines corresponding diagnostic reasoning, if it is not final abnormal temperature reason,
The abnormal temperature pattern is considered into profound abnormal temperature symptom；
The observation operation module is used for the Different Results of expression observation operation, and observation operation module acts on some operation section
Point is used to confirm that certain abnormal temperature whether there is, or the abnormal temperature information for obtaining and abnormal temperature hypothesis node is related is come
Source, it is the main canal for obtaining greenhouse abnormal temperature information to represent external observation phenomenon, result of test operation, observation running node
Road；
The probability Estimation module acts on observation running node, observes the probability of running node value result, each for judging
Plant observation operating result；
The analyzing and diagnosing module, up to the Different Results of diagnostic operation, is tied in diagnostic operation node table by the difference of diagnostic operation
Can fruit operation determine specific certain abnormal temperature；And by the probability of diagnostic operation node and observation running node whether
Cause, so as to be diagnosed to be the probability of abnormal temperature；
Analyzing and diagnosing module in diagnosis when diagnostic reasoning is carried out according to greenhouse abnormal temperature symptom, seek most possible
Abnormal temperature reason, is then operated, and if not the abnormal temperature reason abnormal temperature possibility is then excluded, and is continued to look for
Other possible abnormal temperature reasons, until eventually finding abnormal temperature reason.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611132830.5A CN106599568B (en) | 2016-12-10 | 2016-12-10 | A kind of greenhouse abnormal temperature diagnostic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611132830.5A CN106599568B (en) | 2016-12-10 | 2016-12-10 | A kind of greenhouse abnormal temperature diagnostic device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106599568A true CN106599568A (en) | 2017-04-26 |
| CN106599568B CN106599568B (en) | 2019-06-21 |
Family
ID=58598761
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611132830.5A Active CN106599568B (en) | 2016-12-10 | 2016-12-10 | A kind of greenhouse abnormal temperature diagnostic device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106599568B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115053735A (en) * | 2022-07-13 | 2022-09-16 | 黄山四月乡村农艺场有限公司 | Constant temperature greenhouse temperature control system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN203337277U (en) * | 2013-09-04 | 2013-12-11 | 天津科技大学 | ZigBee-based greenhouse wireless temperature monitoring system |
| JP2015188421A (en) * | 2014-03-28 | 2015-11-02 | ダイキン工業株式会社 | remote monitoring system |
-
2016
- 2016-12-10 CN CN201611132830.5A patent/CN106599568B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN203337277U (en) * | 2013-09-04 | 2013-12-11 | 天津科技大学 | ZigBee-based greenhouse wireless temperature monitoring system |
| JP2015188421A (en) * | 2014-03-28 | 2015-11-02 | ダイキン工業株式会社 | remote monitoring system |
Non-Patent Citations (3)
| Title |
|---|
| SUN J, ET AL.: "The System Design of Monitoring abnormal Environment Parameter of Vegetable Greenhouse based on Short Message", 《THE SYSTEM DESIGN OF MONITORING ABNORMAL ENVIRONMENT PARAMETER OF VEGETABLE GREENHOUSE BASED ON SHORT MESSAGE》 * |
| 刘锦，等；: "基于物联网架构的温室环境监测系统", 《河北农业大学学报》 * |
| 赵方，等；: "基于MSP430的温室大棚温度远程监控系统", 《农机化研究》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115053735A (en) * | 2022-07-13 | 2022-09-16 | 黄山四月乡村农艺场有限公司 | Constant temperature greenhouse temperature control system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106599568B (en) | 2019-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hu et al. | Detection of frequent alarm patterns in industrial alarm floods using itemset mining methods | |
| CN106844194B (en) | A kind of construction method of multi-level software fault diagnosis expert system | |
| KR101872342B1 (en) | Method and device for intelligent fault diagnosis using improved rtc(real-time contrasts) method | |
| KR101948604B1 (en) | Method and device for equipment health monitoring based on sensor clustering | |
| JP3651693B2 (en) | Plant monitoring diagnosis apparatus and method | |
| KR102440335B1 (en) | A method and apparatus for detecting and managing a fault | |
| CN107391369A (en) | A kind of spanned item mesh failure prediction method based on data screening and data oversampling | |
| CN107273924A (en) | The Fault Analysis of Power Plants method of multi-data fusion based on fuzzy cluster analysis | |
| CN110738255A (en) | device state monitoring method based on clustering algorithm | |
| US20220004163A1 (en) | Apparatus for predicting equipment damage | |
| CN108268893A (en) | A kind of chemical industrial park method for early warning and device based on machine learning | |
| US20230221684A1 (en) | Explaining Machine Learning Output in Industrial Applications | |
| JP2002082713A (en) | Method and system for monitoring industrial process | |
| CN110455397A (en) | A kind of shop equipment failure optical fiber sensing method and system based on LSTM | |
| CN109634233A (en) | Industrial big data intellectual analysis decision-making technique, readable storage medium storing program for executing and terminal | |
| JP6917805B2 (en) | Data filtering device and method | |
| CN116842423A (en) | Aeroengine fault diagnosis method and system based on multi-mode deep learning | |
| CN109614451A (en) | Industrial big data intellectual analysis decision making device | |
| JPH0793018A (en) | Method and system for diagnosing operating state | |
| CN115237086A (en) | Decision support in industrial plants | |
| CN106599568A (en) | Abnormal temperature diagnosis device for greenhouse | |
| KR102366787B1 (en) | Real-time sliding window based anomaly detection system for multivariate data generated by manufacturing equipment | |
| KR102110319B1 (en) | System for generating learning data | |
| CN110837953A (en) | Automatic abnormal entity positioning analysis method | |
| CN111353160B (en) | Software bug abnormity intelligent detection system and method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |