The monitoring of aluminum section extruder real time energy consumption and energy consumption abnormality detection system
Technical field
The present invention relates to the monitoring of a kind of aluminum section extruder real time energy consumption and energy consumption abnormality detection technical field.
Background technology
Aluminium section bar manufacturing enterprise belongs to high energy consumption, maximum discharge manufacturing enterprise, and extruding production cost controls to be one of Important Problems of enterprises pay attention.In order to ensure extruding administration measure, run economically, carrying out Real-Time Monitoring and energy consumption abnormality detection to extruder energy consumption, is the development trend realizing manufacturing automation and cleaner production.In addition, to extrude in units of unit, workshop, production Energy-saving Situation is assessed, effectively can reduce the energy consumption abnormal occurrencies such as the energy is revealed, stand-by time is long, squeezing parameter is incorrect.
The mode that energy measurement major part adopts human metering produced by aluminum current section bar, staff is needed to go over each extruding unit recorder data, not only can not gather production data (gathering once for a day) because plant area is excessive in time, can't the energy consumption abnormal problem in production scene be processed timely.
The Evaluation on Energy Saving means of current employing do not adopt the energy consumption of the energy consumption of the unit product of conservation measures and the unit product after adopting conservation measures to contrast.But this method exists serious hysteresis quality, along with difference, the difference of unit staff, the difference of production product of production time, this appraisal procedure can produce larger error.
Summary of the invention
The object of the invention is, a kind of real-time energy consumption monitoring system and the energy consumption abnormality detection system that are applied to aluminum section extruder are proposed, thus the energy consumption data in the production of monitoring extruding in real time, and the energy consumption abnormal occurrence in can producing according to energy consumption prediction discovery extruding.
For reaching this object, first the present invention provides a kind of aluminum section extruder real-time energy consumption monitoring system, comprising: be installed on the intelligent electric meter in ectrusion press electric line; Be installed on the intelligent gas meter on auxiliary oven heat supply pipeline; Described digital instrument is all connected to serial server; Described serial server is connected to switch; Described switch is connected to monitoring server.
Setting like this, can be implemented in the real time energy consumption acquisition and processing in aluminum extrusion process, saves manual operation, raises the efficiency, can energy consumption data in examinations extrusion process of production, the phenomenon of process energy consumption exception in time.
Secondly, the invention provides the monitoring of a kind of aluminum section extruder real time energy consumption and energy consumption abnormality detection system, it comprises: be installed on the intelligent electric meter in ectrusion press electric line; Be installed on the intelligent gas meter on auxiliary oven heat supply pipeline; Described digital instrument is all connected to serial server; Described serial server is connected to switch; Described switch is connected to monitoring server;
Also comprise data acquisition module and server control module;
Wherein said data acquisition module runs on described serial server, gathers the load energy consumption data from intelligent electric meter and intelligent gas meter, and load energy consumption data is passed to server control module;
Described server control module runs on monitoring server, comprises data memory module, training pattern module, the interval module of prediction of energy consumption, energy consumption comparison module;
Described data memory module receives the load energy consumption data that data acquisition module transmits, and is stored in database by load energy consumption data according to acquisition time, and data memory module also stores the production data had an impact to energy consumption simultaneously;
Described training pattern module is based on Support Vector Machines for Regression, the historic load energy consumption data in usage monitoring server database and production data training Support Vector Machines for Regression;
The interval module of described prediction of energy consumption, with the fiducial interval of statistical analysis technique unit of account product energy consumption;
Described energy consumption comparison module, reads real time energy consumption data, compares with energy consumption fiducial interval, show that extruding energy consumption is normal or energy consumption abnormal.
Wherein, described energy consumption comparison module is the actual value X reading out extruding unit energy consumption of unit product from monitoring server database
t, and the predicted value X of the energy consumption of unit product of Support Vector Machines for Regression calculating
p, according to formula
calculate the energy-saving effect η of described unit product.
Wherein, described training pattern module comprises: the pre-service load energy consumption data read out in monitoring server database and production data being converted to the training data of Support Vector Machines for Regression model, regression function f (x) in training Support Vector Machines for Regression model.
Wherein, described input pre-service is training data load energy consumption data and production data being converted to Support Vector Machines for Regression model, namely according to the time gathered, by load energy consumption data { f (x
1), f (x
2) ..., f (x
n) and corresponding production data { x
1, x
2..., x
nas one group of data <f (x
i), x
i>, i=1,2, ..., n, for training regression function f (x)=wx+b, w and b is the lineoid parameter of matching training data respectively, and namely training process is the form by solving equation, with multi-group data <f (x
i), x
i>, i=1,2 ..., n calculates the process of lineoid parameter w and b.
Wherein, the interval module of described prediction of energy consumption also comprises: for analyzing the history energy consumption data in monitoring server database according to Estimating Confidence Interval method, and given degree of confidence 1-α, obtains the normal interval of energy consumption prediction.
Wherein, the prediction power consumption of unit extruded aluminium section is X1, X2 ... Xn obeys sample distribution (μ, σ
2),
and S
2represent sample average and sample variance, the then stochastic variable of prediction power consumption respectively
For given degree of confidence 1-α,
Wherein P represents probability, then predict that the fiducial interval of the average μ of power consumption is
The invention has the beneficial effects as follows: real-time energy consumption monitoring system of the present invention reads the real time data of each digital instrument by serial server, the data (interval <3 second) such as power consumption, voltage, electric current, consumption, temperature, pressure, flow that centralized displaying extruding is produced, by adding up extruding production history data, obtain normal energy consumption fiducial interval, the energy consumption abnormal occurrence in producing extruding can Timeliness coverage make corresponding process.Based on the method for Support Vector Machines for Regression determination energy-saving effect, it carries out energy consumption prediction according to current manufacturing parameter, compares with actual consumption value, effectively can avoid the hysteresis quality of data.
Accompanying drawing explanation
Fig. 1 is aluminum section extruder real-time energy consumption monitoring system structural drawing of the present invention.
Fig. 2 is the realization flow figure of the abnormal energy consumption detection system of the extruding based on Support Vector Machines for Regression of the present invention.
Fig. 3 is the energy-conservation efficiency figure of certain extruding unit that one embodiment of the present of invention provide.
Embodiment
Below in conjunction with drawings and the specific embodiments, the present invention is carried out more in detail and complete explanation.Be understandable that, specific embodiment described herein is only for explaining the present invention, but not limitation of the invention.It also should be noted that, for convenience of description, illustrate only part related to the present invention in accompanying drawing but not full content.
As shown in Figure 1, it is the structural drawing of aluminum section extruder real-time energy consumption monitoring system of the present invention, shows the composition of hardware components, and it comprises digital instrument, serial server, switch and monitoring server four layers from bottom successively to upper strata.Wherein, bottom digital instrument comprises intelligent electric meter and intelligent gas meter, intelligent electric meter is installed in ectrusion press electric line, for gathering the parameters such as power consumption, voltage, electric current, applied power, active power, reactive power, power factor, frequency, this parameter is divided into again A, B, C three-phase and gathers; Intelligence gas meter is installed on auxiliary oven heat supply pipeline, mainly gathers 4 parameters such as consumption, temperature, pressure, flow.
Can be connected by wired or wireless mode between each layer, The present invention gives a kind of wired connected mode, described bottom digital instrument is connected by RS-485 interface with serial server, MODBUS-RTU communications protocol is used to carry out data communication, this serial server adopts MOXANPort5430, serial ports end is connected with instrumentation, network-side is connected with switch, and use ICP/IP protocol to carry out data communication, serial server is installed on workshop, for transmitting data in serial ports and the network equipment, realize the data double-way transparent transmission of RS-485 serial ports to TCP/IP network interface, respective switch is connected with monitoring server by LAN (Local Area Network), and transmitting network data is to monitoring server.And monitoring server is for receiving and record the load energy consumption data of switch transmission and running Support Vector Machines for Regression method.
Software section of the present invention comprises data acquisition module and server control module, described data acquisition module runs on serial server, its action is after receiving configuration parameter from monitoring server, the MODBUS data command frame of generation standard, and send to digital meter's, after receiving the reply data frame that digital meter's returns, be TCP/IP packet used by the content packaging in Frame, be forwarded in switch gateway by network interface.Described server control module runs on monitoring server, the Software Development Platform VC++6.0 of Microsoft's exploitation is adopted to develop, use the Mscomm control of encapsulation to carry out Serial Port Transmission, and the data collected by hardware are transferred in host computer interface and show in real time and be stored in SQL database.It at least comprises database module, training pattern module, prediction of energy consumption interval module, energy consumption comparison module.
As shown in Figure 2, it is structure and the operating diagram of server control module of the present invention.Database is comprised in described database module, for the load energy consumption data that stores data collecting module collected and the production data that can have an impact to energy consumption in process of production, wherein load energy consumption data comprises (power consumption that digital electric meter gathers and the gas consumption that digital gas meter gathers), described production data comprises (extrusion temperature, extrusion speed, cooling velocity, aluminium bar type, alloying component etc.), wherein extrusion temperature directly has influence on mold heated energy consumption and aluminium bar heating energy consumption, extrusion speed then has a direct impact extruding energy consumption, cooling velocity directly affects cooling energy consumption, aluminium bar type and alloying component have remote effect to extruding energy consumption.
Described training pattern module, is characterized in that training the Evaluation on Energy Saving model based on Support Vector Machines for Regression.The load energy consumption data read out in monitoring server database and production data are converted to the pre-service of the training data of Support Vector Machines for Regression model, regression function f (x) in training Support Vector Machines for Regression model.Described input pre-service is training data load energy consumption data and production data being converted to Support Vector Machines for Regression model, namely according to the time gathered, by load energy consumption data { f (x
1), f (x
2) ..., f (x
n) and corresponding production data { x
1, x
2..., x
nas one group of data <f (x
i), x
i>, i=1,2, ..., n, for training regression function f (x)=wx+b, w and b is the lineoid parameter of matching training data respectively, and namely training process is the form by solving equation, with multi-group data <f (x
i), x
i>, i=1,2 ..., n calculates the process of lineoid parameter w and b.
For certain aluminium section bar manufacturing enterprise extruding workshop, original energy consumption time series data, the energy consumption data comprising the different time such as day, Month And Year dimension is pre-stored within the SQL database of monitoring server by energy-consumption monitoring system.From the SQL database of monitoring server, read a certain extruding unit often extrude the power consumption of 1 ton of aluminium section bar and gas quantity as training data in 2013.4-2013.8, the production data reading out this extruding unit production run again from database, as input data (comprising: extrusion temperature, extrusion speed, cooling velocity, aluminium bar type, alloying component etc.), uses MATLAB training based on regression function f (x) of Support Vector Machines for Regression; According to regression function calculate this extruding unit in 2013.9 often extrude the power consumption of 1 ton of aluminium section bar and gas quantity be 373.74 KWhs/ton, 36.68 cubic metres/ton.
The interval module of described prediction of energy consumption, is characterized in that the fiducial interval of Using statistics analytical approach unit of account product energy consumption.If the prediction power consumption of unit extruded aluminium section is X
1, X
2... X
nobey sample distribution (μ, σ
2),
and S
2represent sample average and sample variance, the then stochastic variable of prediction power consumption respectively
For given degree of confidence 1-α,
P is probability, then predict that the fiducial interval of the average μ of power consumption is
For certain aluminium section bar manufacturing enterprise extruding workshop, in 2013.7 months, the power consumption of 5 days is respectively 345.24 KWhs/ton, 343.82 KWhs/ton, 354.05 KWhs/ton, 346.44 KWhs/ton, 353.26 KWhs/ton, to be then the power consumption fiducial interval of 0.99 be degree of confidence:
S
2=22.55,
Then predict that the fiducial interval of power consumption is for [348.56-9.77,348.56+9.77].
Described energy consumption comparison module, is characterized in that the energy consumption predicted value X more often extruding 1 ton of aluminium section bar
poften extrude the actual value X of 1 ton of aluminium section bar energy consumption
t, according to formula
calculate the energy-saving effect η of described unit product.To be degree of confidence shown in Fig. 3 be 99% certain extruding unit day can find out multiple energy consumption abnormal conditions in energy-conservation efficiency figure, figure.
Be only the preferred embodiments of the present invention described in upper, be not limited to the present invention, to those skilled in the art, the present invention can have various change and change.All do within spirit of the present invention and principle any amendment, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.