Invention content
The present invention provides a kind of machine room humiture environment for overcoming the above problem or solving the above problems at least partly
Prediction technique and system.
According to an aspect of the present invention, a kind of machine room humiture environmental forecasting method is provided, including:
To each collection point in computer room, the fisrt feature parameter, second feature parameter and third for acquiring the collection point are special
Parameter is levied, and is separately input into temperature prediction model, humidity model and transient temperature prediction model, prediction is obtained
In temperature sequence of the sampled point in first predetermined period, the humidity sequence in second predetermined period and third predetermined period
Temperature sequence, described third predetermined period be much smaller than described first predetermined period;
According to the temperature sequence and second of the position and sampled point of each sampled point of prediction in first predetermined period
Humidity sequence in predetermined period obtains computer room in the temperature sequence and second predetermined period in described first predetermined period
Humidity sequence;According to the temperature existed in temperature sequence of the collection point in third predetermined period more than preset temperature threshold
Degree, is alerted.
Preferably, the fisrt feature parameter include outdoor temperature, air-conditioner temperature, ambient humidity, server contention states,
The position of collection point, heat dissipation capacity, illumination parameter and the first sampling period of the last time interior this of collection point place equipment are adopted
Collect the temperature sequence of point
The second feature parameter includes outside humidity, environment temperature, weather conditions, humidifier working condition, server
Working condition, the position of collection point, illumination parameter and the humidity sequence in the second sampling period of the last time;
The third feature parameter includes outdoor temperature, air-conditioner temperature, ambient humidity, server contention states, collection point
Position, equipment where collection point heat dissipation capacity, illumination parameter and the collection point in the third of the last time sampling period
Temperature sequence.
Preferably, the temperature prediction model is established especially by following steps:
Multigroup first sample data are obtained from computer room database and constitute the first training set, and the first sample data include
Temperature sequence of the fisrt feature parameter of any collection point with the collection point within the first the last sampling period in computer room;
Sample data in first training set is substituted into decision Tree algorithms, is established described in the conduct of the first decision-tree model
Temperature prediction model.
Preferably, the decision-tree model is that gradient promotes tree-model.
Preferably, the humidity model is established especially by following steps:
Multigroup second sample data is obtained from computer room database as the second training set, second sample data includes
Humidity sequence of the second feature parameter of any collection point with the collection point within the second the last sampling period in computer room;
The second sample data in second training set is updated in decision Tree algorithms, the second decision-tree model is established
As the humidity model.
Preferably, the transient temperature prediction model is established especially by following steps:
Multigroup third sample data is obtained from computer room database as third training set, the third sample data includes
Temperature sequence of the third feature parameter of any collection point with the collection point within the last third sampling period in computer room;
Third sample data in the third training set is updated in decision Tree algorithms, third decision-tree model is established
As the transient temperature prediction model.
Preferably, temperature of the position and sampled point of each sampled point according to prediction in first predetermined period
It is pre- to obtain temperature sequence and second of the computer room in described first predetermined period for humidity sequence in sequence and second predetermined period
The step of surveying the humidity sequence in the period, specifically includes:
Temperature sequence of the position and sampled point of each sampled point of prediction in described first predetermined period, establishes machine
Temperature field figure of the room in described first predetermined period;According to the temperature field figure, the computer room is obtained in first predetermined period
Interior temperature sequence;
Humidity sequence of the position and sampled point of each sampled point of prediction in described second predetermined period, establishes machine
Moisture field figure of the room in described second predetermined period;According to the moisture field figure, the computer room is obtained in second predetermined period
Interior humidity sequence;
According to another aspect of the present invention, a kind of machine room humiture environmental forecasting system is also provided, including:
Sampled point prediction module, for each collection point in computer room, acquire the collection point fisrt feature parameter, second
Characteristic parameter and third feature parameter, and it is pre- to be separately input into temperature prediction model, humidity model and transient temperature
It surveys in model, obtains temperature sequence of the sampled point of prediction in first predetermined period, the humidity sequence in second predetermined period
Temperature sequence in row and third predetermined period, described third predetermined period are much smaller than described first predetermined period;
Computer room predicts alarm module, is used for position and sampled point according to each sampled point of prediction in the first prediction week
The humidity sequence in temperature sequence and second predetermined period in phase obtains temperature sequence of the computer room in described first predetermined period
Humidity sequence in row and second predetermined period;It is big according to existing in temperature sequence of the collection point in third predetermined period
In the temperature of preset temperature threshold, alerted.
Preferably, the fisrt feature parameter include outdoor temperature, air-conditioner temperature, ambient humidity, server contention states,
The position of collection point, heat dissipation capacity, illumination parameter and the first sampling period of the last time interior this of collection point place equipment are adopted
Collect the temperature sequence of point;
The second feature parameter includes outside humidity, environment temperature, weather conditions, humidifier working condition, server
Working condition, the position of collection point, illumination parameter and the humidity sequence in the second sampling period of the last time;
The third feature parameter includes outdoor temperature, air-conditioner temperature, ambient humidity, server contention states, collection point
Position, equipment where collection point heat dissipation capacity, illumination parameter and the collection point in the third of the last time sampling period
Temperature sequence.
Preferably, the temperature prediction model is established especially by following steps:
Multigroup first sample data are obtained from computer room database and constitute the first training set, and the first sample data include
Temperature sequence of the fisrt feature parameter of any collection point with the collection point within the first the last sampling period in computer room;
Sample data in first training set is substituted into decision Tree algorithms, is established described in the conduct of the first decision-tree model
Temperature prediction model.
Machine room humiture environmental forecasting method and system proposed by the present invention, using decision Tree algorithms such as gradient boosted trees,
Corresponding temperature prediction model and humidity model are established according to the history environment data of computer room, and sensor is collected
The current epidemic disaster information of respectively acquiring correspond to and input in above-mentioned prediction model, to predict epidemic disaster variation tendency in real time,
Computer room accident rate is reduced, energy use efficiency is improved..
Specific implementation mode
With reference to the accompanying drawings and examples, the specific implementation mode of the present invention is described in further detail.Implement below
Example is not limited to the scope of the present invention for illustrating the present invention.
It is a kind of flow diagram of machine room humiture environmental forecasting method provided in an embodiment of the present invention, such as referring to Fig. 1
Shown in figure, this method includes:
101, to each collection point in computer room, fisrt feature parameter, the second feature parameter and of the collection point are acquired
Three characteristic parameters, and be separately input into temperature prediction model, humidity model and transient temperature prediction model, it obtains pre-
Temperature sequence of the sampled point in first predetermined period, the humidity sequence in second predetermined period and the third prediction week surveyed
Temperature sequence in phase, third predetermined period are much smaller than first predetermined period;
102, the temperature sequence according to the position and sampled point of each sampled point of prediction in first predetermined period and
Humidity sequence in second predetermined period obtains computer room in the temperature sequence and second predetermined period in first predetermined period
Humidity sequence;According in temperature sequence of the collection point in third predetermined period exist more than preset temperature threshold temperature, into
Row alarm.
Specifically, computer room is the place that information equipment is left concentratedly, most important to informatization, good computer room ring
Border is the basis for ensureing equipment high efficiency rate operating in computer room, and building environment plays considerable effect in plant maintenance.
When building environment when something goes wrong, the equipment life in computer room may be caused to shorten, operational reliability reduce so that computer room accident
Incidence rise.The main environmental factor for influencing calculator room equipment performance includes temperature and humidity in computer room.It is with temperature
Example can make steam generate condensation and dew condensation phenomenon, to reduce the insulating properties of circuit board when temperature is too low in computer room so that
Insulating material is harder, more crisp, and electronic component can be caused to get rusty, and accelerates its oxidation.And work as in computer room when the temperature is excessively high, it can make
Device electronics aging accelerates, and causes job insecurity, and it is bad in ventilating and cooling condition when, electronics member device can be caused
Part is burned out.If rate of temperature change is big, equipment is easy fogging, due to the coefficient of expansion difference of electronic component be easy to happen it is disconnected
It splits.
Since equipment now is essentially super large-scale integration form, semiconductor devices such as chip in circuit etc. are all
It can be affected by temperature, will produce heat when equipment is run.If not having good heat dissipation equipment, heat can be made timely
It disperses, thus the crash rate of these caused semiconductor devices can be improved with the raising of temperature, cause equipment fault.Therefore,
It makes prediction to the variation tendency of temperature in computer room, and according to predicted temperature real-time monitoring computer room room temperature, reduces computer room accident
Incidence, be very it is necessary to.
Sensor is set on the equipment component spread in computer room, using the place where sensor as collection point, sensing
Device is used to acquire the environmental information of the point.By all collection points in computer room in the collected fisrt feature parameter input of synchronization
Into temperature prediction model so that temperature prediction model corresponds to the temperature information for exporting each collection point following a period of time.Here
Fisrt feature parameter include outdoor temperature, air-conditioner temperature, ambient humidity, server contention states, collection point position, acquisition
The temperature sequence of the collection point in the heat dissipation capacity of equipment, illumination parameter and the first sampling period of the last time where point.
Wherein, server includes the working conditions such as busy, normal, pause, under different working conditions, the hair of server
Hot degree is different, and the influence to temperature is also different.And since the degree of heat of distinct device at work is also not quite similar, i.e., not
Influence with equipment to temperature under identical room temperature situation is also different, needs the heat dissipation capacity for obtaining the equipment where collection point,
Influence degree for evaluating equipment to sensor position temperature.And illumination parameter is when including mainly intensity of illumination and illumination
Between etc. light conditions, in the case of mutually synthermal, light conditions can also generate the variation tendency of temperature certain influence.For
It realizes the real-time prediction to computer room temperature, needs the temperature of the first sampling period for the acquiring computer room the last time interior collection point
Sequence, above-mentioned first sampling period is 30 minutes desirable, that is, acquires multiple temperature that nearest 30 minutes inner sensors chronologically obtain
Value, you can know the temperature changing trend at collection point in this 30 minutes.
It, can be defeated according to temperature prediction model after the fisrt feature parameter of each collection point is input in temperature prediction model
Predicted temperature sequence of each collection point gone out in first predetermined period, first predetermined period here is 10 minutes desirable, is
Chronologically predict multiple temperature values of each collection point within 10 minutes futures.In conjunction with the predicted temperature sequence of each collection point
With the position of the collection point, temperature field figure of the computer room in first predetermined period can be established, temperature field is for indicating temperature in sky
Between distribution on domain and time-domain, temperature sequence of the computer room in first predetermined period can be known according to temperature field figure, you can know
Temperature changing trend of the computer room within following a period of time.Machine can be reduced by adjusting the operation conditions of air-conditioning in computer room at this time
Room energy consumption improves energy use efficiency, while realizing the real time monitoring to building environment.
The embodiment of the present invention passes through the collected Current Temperatures letter of input pickup into trained temperature prediction model
Breath predicts that the variation tendency of computer room future temperature, the operation conditions of the equipment such as adjustment air-conditioning make corresponding reply in advance in real time
And processing, reduce the incidence of computer room accident.
Temperature prediction model is established especially by following steps:Multigroup sample data is obtained from computer room database as
One training set, sample data include fisrt feature parameter and the collection point of any collection point in computer room the last first
Temperature sequence in predetermined period;Sample data in first training set is updated in decision Tree algorithms, corresponding determine is established
Plan tree algorithm is as temperature prediction model.
Specifically, before establishing temperature prediction model, multigroup sample data is first obtained from computer room database as the first instruction
Practice collection, temperature prediction model is trained according to the first training set.All include the of any collection point in above-mentioned each sample data
Actual temperature sequence of one characteristic parameter with the collection point in first the last predetermined period.By collection point it is current the
One characteristic parameter inputs trained temperature prediction model can acquire the collection point after getting corresponding predicted temperature sequence
Practical future temperature, compare the predicting temperature values of the collection point and actual temperature value to verify the accuracy rate of model, and according to
Above-mentioned data continue to be trained original temperature prediction model, are constantly carried out in practical applications to the temperature prediction model
It is perfect, to improve the accuracy rate of temperature prediction model.
Based on above-described embodiment, as a kind of optional embodiment, decision-tree model is that gradient promotes tree-model.
Specifically, gradient boosted tree (full name in English Gradient Boosting Decison Tree, abbreviation GBDT) is
A kind of decision-tree model, GBDT models can flexibly handle various types of data, including successive value and centrifugal pump.Due to this hair
Temperature prediction model in bright is according to the multiple temperature values chronologically obtained interior for the previous period come when obtaining one section following
Between temperature changing trend, the training of the more convenient this successive value of GBDT models.In addition, GBDT models are when relatively little of tune is joined
Between in the case of, the accuracy rate of prediction can also be relatively high.It should be noted that the loss function used due to GBDT models is more
Stalwartness makes it spatially have stronger robustness for exceptional value in output.
Further include as a kind of optional embodiment based on above-described embodiment:Acquire second of each collection point in computer room
Characteristic parameter, second feature parameter include outside humidity, environment temperature, weather conditions, humidifier working condition, server work
The humidity sequence of the collection point in state, the position of collection point, illumination parameter and the second sampling period of the last time;It will be every
In the second feature parameter input humidity model of one collection point, pre- measuring moisture of the collection point in second predetermined period is obtained
Sequence;In conjunction with the position of pre- the measuring moisture sequence and the collection point of each collection point, computer room is established in second predetermined period
Moisture field figure knows humidity sequence of the computer room in second predetermined period according to moisture field figure.
Specifically, equipment also has strict requirements to humidity in computer room.When humidity is excessively high in computer room, inside equipment
Chip may be rusted so that phenomena such as short circuit, electric leakage or poor contact occurs in equipment, so that the failure of equipment occurs
Rate increase, badly damaged electronic component;When humidity is too low in computer room, electrostatic phenomenon is easy tod produce, easily causes part electricity
Sub- component, such as bipolar circuit, CMOS, MOS circuit breakdown and breaking-up, be unfavorable for the normal work of machine.
All collection points in computer room are input in the collected second feature parameter of synchronization in humidity model,
So that humidity model corresponds to the humidity information for exporting each collection point following a period of time.Here second feature parameter includes
Outside humidity, environment temperature, weather conditions, humidifier working condition, server contention states, the position of collection point, illumination ginseng
The humidity sequence of number and the second the last sampling period interior collection point.Wherein, in order to realize the reality to computer room temperature
When predict, need the humidity sequence for acquiring the collection point in the second sampling period of computer room the last time, above-mentioned second sampling all
Phase is 30 minutes desirable, that is, acquires multiple humidity values that nearest 30 minutes inner sensors chronologically obtain, you can know in this 30 minutes
Humidity variation tendency at collection point.
It, can be defeated according to humidity model after the second feature parameter of each collection point is input in humidity model
Pre- measuring moisture sequence of each collection point gone out in second predetermined period, second predetermined period here is 10 minutes desirable, is
Chronologically predict multiple humidity values of each collection point within 10 minutes futures.In conjunction with the pre- measuring moisture sequence of each collection point
With the position of the collection point, moisture field figure of the computer room in second predetermined period can be established, moisture field is for indicating humidity in sky
Between distribution on domain and time-domain, humidity sequence of the computer room in second predetermined period can be known according to moisture field figure, you can know
Humidity variation tendency of the computer room within following a period of time.It at this time can be by adjusting the operation conditions of humidifier in computer room, wet
Reinforce the real time monitoring to building environment in terms of degree.
Based on above-described embodiment, humidity model is established especially by following steps:It is obtained from computer room database more
For group sample data as the second training set, sample data includes second feature parameter and collection point of any collection point in computer room
Pre- measuring moisture sequence in second the last predetermined period;Sample data in second training set is updated to decision tree
In algorithm, corresponding decision Tree algorithms are established as humidity model.
Specifically, before establishing humidity model, multigroup sample data is first obtained from computer room database as the second instruction
Practice collection, humidity model is trained according to the second training set.All include the of any collection point in above-mentioned each sample data
Actual humidity sequence of two characteristic parameters with the collection point in second the last predetermined period.By collection point it is current the
Two characteristic parameters input trained humidity model, after getting corresponding pre- measuring moisture sequence, can acquire the collection point
The practical following humidity, compare prediction humidity value and the actual humidity value of the collection point to verify the accuracy rate of model, and according to
Above-mentioned data continue to be trained original humidity model, are constantly carried out in practical applications to the humidity model
It is perfect, to improve the accuracy rate of humidity model.It should be noted that the decision-tree model of humidity can also pass through gradient
Tree algorithm is promoted to establish.
Further include as a kind of optional embodiment based on above-described embodiment:Acquire the third of any collection point in computer room
Characteristic parameter, third feature parameter include outdoor temperature, air-conditioner temperature, ambient humidity, server contention states, collection point position
It sets, the temperature of the interior collection point of the heat dissipation capacity of equipment, illumination parameter and the third of the last time sampling period where collection point
Sequence;By in the third feature parameter input transient temperature prediction model of collection point, collection point is obtained in third predetermined period
Temporal prediction temperature sequence, according to any temperature value in the temporal prediction temperature sequence of collection point be more than temperature threshold, it is right
Computer room makes early warning;Wherein, the third sampling period was less than for the first sampling period.
Specifically, it since temperature may there is a situation where the rapid drawdowns that rises sharply, to prevent the occurrence of such, then also needs to predict
Temperature change in computer room instantaneous time need to only ensure the temperature value of any point in computer room all in normal range (NR) at this time.In computer room
In the key positions such as accident-prone site, computer room core position on sensor is set, using the place where sensor as acquisition
Point.By any collection point in computer room, currently collected third feature parameter is input in transient temperature prediction model so that wink
When temperature prediction model correspond to and export temperature information of the collection point in the following very short time.Here fisrt feature parameter packet
The heat dissipation capacity of equipment, illumination ginseng where including outdoor temperature, air-conditioner temperature, server contention states, the position of collection point, collection point
Number and the temperature sequence of the interior collection point of the last third sampling period.
Wherein, in order to realize the real-time temporal prediction to computer room temperature, the third sampling for acquiring computer room the last time is needed
The temperature sequence of the collection point in period.It should be noted that due to being when there is cataclysm for temperature at collection point at this time
Monitoring, therefore the third sampling period is shorter, the above-mentioned third sampling period is 30 seconds desirable, that is, acquires nearest 30 seconds inner sensors on time
Multiple temperature values that sequence obtains, you can know the temperature changing trend at collection point in this 30 seconds.
It, can be according to transient temperature after the third feature parameter of any collection point is input in transient temperature prediction model
Predicted temperature sequence of the collection point of prediction model output in third predetermined period, it should be noted that due to instantaneous temperature
Degree prediction model is monitoring when occurring cataclysm for temperature at collection point, because third predetermined period is much smaller than the first prediction week
Phase.Here third predetermined period is 10 seconds desirable, is the multiple temperature values for chronologically predicting the collection point within 10 seconds futures,
To realize the prediction of temperature transient change at collection point.Any temperature in the temporal prediction temperature sequence of any collection point
Value is more than temperature threshold, it is known that the temperature of the collection point will be more than instantaneously normal range (NR), may cause computer room accident, then immediately
Early warning is made to computer room, reduces the incidence of computer room accident.
Based on above-described embodiment, transient temperature prediction model is established especially by following steps:It is obtained from computer room database
Take multigroup sample data as third training set, sample data includes that the third feature parameter of any collection point in computer room is adopted with this
Temporal prediction temperature sequence of the collection point in the last third predetermined period;Sample data in third training set is substituted into
Into decision Tree algorithms, corresponding decision Tree algorithms are established as transient temperature prediction model.
Specifically, before establishing transient temperature prediction model, multigroup sample data is first obtained from computer room database as the
Three training sets train transient temperature prediction model according to third training set.All include any adopt in above-mentioned each sample data
Collect actual temperature sequence of the third feature parameter of point with the collection point in the last third predetermined period.By collection point
Current third feature parameter inputs trained transient temperature prediction model can after getting corresponding transient temperature sequence
The practical future temperature for acquiring the collection point, the predicting temperature values for comparing the collection point verify the standard of model with actual temperature value
True rate, and continue to be trained original transient temperature prediction model according to above-mentioned data, in practical applications constantly to this
Temperature prediction model carry out it is perfect, to improve the accuracy rate of transient temperature prediction model.It should be noted that instant moisture is predicted
Decision-tree model also can pass through gradient promoted tree algorithm establish.
The embodiment of the present invention is established corresponding using decision Tree algorithms such as gradient boosted trees according to the history environment data of computer room
Temperature prediction model and humidity model, and the current epidemic disaster information of the collected each acquisition of sensor is corresponded to
It inputs in above-mentioned prediction model, to predict epidemic disaster variation tendency in real time, to make corresponding reply and processing in advance,
The operation conditions of the equipment such as air-conditioning, humidifier is adjusted, energy consumption of machine room is reduced, improves energy use efficiency, reduces computer room accident
Incidence.
The present invention also provides a kind of machine room humiture environmental forecasting systems, referring to Fig. 2, including:
Sampled point prediction module 201, for each collection point in computer room, acquire the collection point fisrt feature parameter,
Second feature parameter and third feature parameter, and it is separately input into temperature prediction model, humidity model and instantaneous temperature
It spends in prediction model, obtains temperature sequence of the sampled point of prediction in first predetermined period, wet in second predetermined period
Temperature sequence in degree series and third predetermined period, third predetermined period are much smaller than first predetermined period;
Computer room predicts alarm module 202, is used for pre- first according to the position and sampled point of each sampled point of prediction
The humidity sequence in temperature sequence and second predetermined period in the period is surveyed, temperature sequence of the computer room in first predetermined period is obtained
Humidity sequence in row and second predetermined period;It is more than in advance according to existing in temperature sequence of the collection point in third predetermined period
If the temperature of temperature threshold, is alerted.
In one alternate embodiment, fisrt feature parameter includes outdoor temperature, air-conditioner temperature, ambient humidity, server
The heat dissipation capacity of equipment, illumination parameter and the first sampling week of the last time where working condition, the position of collection point, collection point
The temperature sequence of the collection point in phase;
Second feature parameter includes outside humidity, environment temperature, weather conditions, humidifier working condition, server work
State, the position of collection point, illumination parameter and the humidity sequence in the second sampling period of the last time;
Third feature parameter include outdoor temperature, air-conditioner temperature, ambient humidity, server contention states, collection point position
It sets, the temperature of the interior collection point of the heat dissipation capacity of equipment, illumination parameter and the third of the last time sampling period where collection point
Sequence.
Specifically, computer room is the place that information equipment is left concentratedly, most important to informatization, good computer room ring
Border is the basis for ensureing equipment high efficiency rate operating in computer room, and building environment plays considerable effect in plant maintenance.
When building environment when something goes wrong, the equipment life in computer room may be caused to shorten, operational reliability reduce so that computer room accident
Incidence rise.The main environmental factor for influencing calculator room equipment performance includes temperature and humidity in computer room.It is with temperature
Example can make steam generate condensation and dew condensation phenomenon, to reduce the insulating properties of circuit board when temperature is too low in computer room so that
Insulating material is harder, more crisp, and electronic component can be caused to get rusty, and accelerates its oxidation.And work as in computer room when the temperature is excessively high, it can make
Device electronics aging accelerates, and causes job insecurity, and it is bad in ventilating and cooling condition when, electronics member device can be caused
Part is burned out.If rate of temperature change is big, equipment is easy fogging, due to the coefficient of expansion difference of electronic component be easy to happen it is disconnected
It splits.
Since equipment now is essentially super large-scale integration form, semiconductor devices such as chip in circuit etc. are all
It can be affected by temperature, will produce heat when equipment is run.If not having good heat dissipation equipment, heat can be made timely
It disperses, thus the crash rate of these caused semiconductor devices can be improved with the raising of temperature, cause equipment fault.Therefore,
It makes prediction to the variation tendency of temperature in computer room, and according to predicted temperature real-time monitoring computer room room temperature, reduces computer room accident
Incidence, be very it is necessary to.
Sensor is set on the equipment component spread in computer room, using the place where sensor as collection point, sensing
Device is used to acquire the environmental information of the point.By all collection points in computer room in the collected fisrt feature parameter input of synchronization
Into temperature prediction model so that temperature prediction model corresponds to the temperature information for exporting each collection point following a period of time.Here
Fisrt feature parameter include outdoor temperature, air-conditioner temperature, server contention states, the position of collection point, set where collection point
The temperature sequence of the interior collection point of standby heat dissipation capacity, illumination parameter and the first sampling period of the last time.
Wherein, server includes the working conditions such as busy, normal, pause, under different working conditions, the hair of server
Hot degree is different, and the influence to temperature is also different.And since the degree of heat of distinct device at work is also not quite similar, i.e., not
Influence with equipment to temperature under identical room temperature situation is also different, needs the heat dissipation capacity for obtaining the equipment where collection point,
Influence degree for evaluating equipment to sensor position temperature.And illumination parameter is when including mainly intensity of illumination and illumination
Between etc. light conditions, in the case of mutually synthermal, light conditions can also generate the variation tendency of temperature certain influence.For
It realizes the real-time prediction to computer room temperature, needs the temperature of the first sampling period for the acquiring computer room the last time interior collection point
Sequence, above-mentioned first sampling period is 30 minutes desirable, that is, acquires multiple temperature that nearest 30 minutes inner sensors chronologically obtain
Value, you can know the temperature changing trend at collection point in this 30 minutes.
It, can be defeated according to temperature prediction model after the fisrt feature parameter of each collection point is input in temperature prediction model
Predicted temperature sequence of each collection point gone out in first predetermined period, first predetermined period here is 10 minutes desirable, is
Chronologically predict multiple temperature values of each collection point within 10 minutes futures.In conjunction with the predicted temperature sequence of each collection point
With the position of the collection point, temperature field figure of the computer room in first predetermined period can be established, temperature field is for indicating temperature in sky
Between distribution on domain and time-domain, temperature sequence of the computer room in first predetermined period can be known according to temperature field figure, you can know
Temperature changing trend of the computer room within following a period of time.Machine can be reduced by adjusting the operation conditions of air-conditioning in computer room at this time
Room energy consumption improves energy use efficiency, while realizing the real time monitoring to building environment.
The machine room humiture environmental forecasting system of the embodiment of the present invention into trained temperature prediction model by inputting
The collected Current Temperatures information of sensor predicts the variation tendency of computer room future temperature, the fortune of the equipment such as adjustment air-conditioning in real time
Row situation makes corresponding reply and processing, reduces the incidence of computer room accident in advance.
In one alternate embodiment, temperature prediction model is established especially by following steps:
Multigroup first sample data are obtained from computer room database and constitute the first training set, and first sample data include computer room
In any collection point temperature sequence within the first the last sampling period of fisrt feature parameter and the collection point;
Sample data in first training set is substituted into decision Tree algorithms, establishes the first decision-tree model as temperature prediction
Model.
The apparatus embodiments described above are merely exemplary, wherein can be as the unit that separating component illustrates
Or may not be and be physically separated, the component shown as unit may or may not be physical unit, i.e.,
A place can be located at, or may be distributed over multiple network units.It can select according to the actual needs therein
Some or all of module achieves the purpose of the solution of this embodiment.Those of ordinary skill in the art are not paying creative labor
In the case of dynamic, you can to understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It is realized by the mode of software plus required general hardware platform, naturally it is also possible to pass through hardware.Based on this understanding, on
Stating technical solution, substantially the part that contributes to existing technology can be expressed in the form of software products in other words, should
Computer software product can store in a computer-readable storage medium, such as ROM/RAM, magnetic disc, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
The method of certain parts of example or embodiment.
Finally it should be noted that:The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, it will be understood by those of ordinary skill in the art that:It still may be used
With technical scheme described in the above embodiments is modified or equivalent replacement of some of the technical features;
And these modifications or replacements, various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution spirit and
Range.