CN101684008B - Inorganic waste water reclamation movement expert system - Google Patents
Inorganic waste water reclamation movement expert system Download PDFInfo
- Publication number
- CN101684008B CN101684008B CN2008102006632A CN200810200663A CN101684008B CN 101684008 B CN101684008 B CN 101684008B CN 2008102006632 A CN2008102006632 A CN 2008102006632A CN 200810200663 A CN200810200663 A CN 200810200663A CN 101684008 B CN101684008 B CN 101684008B
- Authority
- CN
- China
- Prior art keywords
- waste water
- scheme
- expert
- water
- quality
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Separation Using Semi-Permeable Membranes (AREA)
Abstract
The invention relates to an inorganic waste water reclamation movement expert system which can enter an enterprise to provide a waste water reclamation scheme on field and perform experiments so as to help the enterprise to make a decision on the waste water reclamation scheme. The system comprises a trolley body, a testing device, an automatic control module and an expert diagnosis decision module, wherein, the trolley body comprises a compartment body arranged on a transportation chassis; the testing device is suitable for forming various device combinations, and each device combination treats waste water in the compartment body according to at least one waste water processing technology; the automatic control module controls the testing device according to the waste water processing scheme to form corresponding device combination and cause the device combination to perform the waste water processing technology; and the expert diagnosis decision module determines and outputs at least one waste water processing scheme to the automatic control module according to a waste water quality index and a reuse water quality index.
Description
Technical field
The present invention relates to inorganic wastewater resource utilization scheme determination method, relate in particular to a kind of inorganic waste water reclamation movement expert system that inorganic wastewater resource utilization scheme is provided.
Background technology
Trade effluent is converted into the universal demand that can reusable new resources have become enterprise, yet waste water is of a great variety, enterprise has nothing in common with each other to the requirement of resource product, is difficult to solve with single scheme.Therefore need to confirm corresponding wastewater treatment and response scheme to different enterprise wastewater sources.
At present,, rely on general knowledge and experience to come initial setting mostly, verify through manufacturer's artificial lab scale of taking a sample again, confirm concrete waste water treatment process and equipment at last for wastewater treatment and reuse scheme determination.If institute's water sampling representativeness very little is not strong, the restriction of empirical deficiency of designer or laboratory test condition tends to cause the failure of lab scale or test that deviation takes place, so that the final plan failure.And a large amount of waste water to transport the laboratory to be unpractical in actually operating, therefore existing mode is difficult to obtain reliable waste water reclaiming scheme.
In addition, existing traditional technology is through after the above requisite operation in each road, and the test period of technology is very long, confirmed process program after the Equipment Design cycle also quite long, and design back equipment also has certain manufacturing cycle.General whole process need some months even more than half a year.After equipment making was installed, treatment process at the scene also differed and satisfies the requirement of user's wastewater treatment surely fully.Can come back after a vain attempt after concerning the user, possibly dropping into great amount of time energy.Making an experiment because technological design is the sample that is directed against waste water, confirm the concrete process program of a cover to this water quality then, is man-to-man working method.Back water quality can not reach requirement if go into operation, and equipment just can not be obtained the result of use of expection.
Summary of the invention
In view of above problem, technical problem to be solved by this invention provides a kind of inorganic waste water reclamation movement expert system, can on-the-spot the resource utilization scheme of waste water be provided and make an experiment, and carries out the waste water reclaiming Scheme Decision-making to help enterprise.
The present invention solves the problems of the technologies described above the technical scheme that adopts to provide a kind of inorganic waste water reclamation movement expert system, comprises locomotive body, testing installation, automatic control module and expert diagnosis decision-making module.The locomotive body comprises the envelope that places on the transportation chassis.Testing installation is suitable for forming the plurality of devices combination, and wherein each equipment combination is carried out wastewater treatment according at least a waste water treatment process in envelope.Automatic control module forms the corresponding apparatus combination and makes the equipment combination carry out waste water treatment process according to wastewater treatment scheme control test equipment.The expert diagnosis decision-making module is confirmed according to waste water quality index and quality of reused water index and is exported at least one wastewater treatment scheme to automatic control module.
In above-mentioned inorganic waste water reclamation movement expert system, above-mentioned testing installation comprises at least a portion with lower component: WQM instrument and instrument, pH regulator equipment, pre-processing device, ion exchange treatment equipment, membrane separation plant, medicine machine.
In above-mentioned inorganic waste water reclamation movement expert system, above-mentioned automatic control module switches said testing installation through self-acting valve and forms said equipment combination.
In above-mentioned inorganic waste water reclamation movement expert system, above-mentioned expert diagnosis decision-making module can comprise input layer, pretreatment layer, hidden layer and output layer.The primary vector that the input layer input comprises said waste water quality index and said quality of reused water index is to pretreatment layer.Pretreatment layer carries out the processing of branch section to waste water quality index in the primary vector and quality of reused water index, and the output secondary vector.Hidden layer is according to each element sum of the trained first weighted value matrix computations secondary vector, and calculates function a plurality of and under a non-linear excitation respectively, export a plurality of and the 3rd vector of function.Output layer is according to each element sum of the trained second weighted value matrix computations the 3rd vector; And calculate function a plurality of and under threshold-type excitation respectively; Output comprise a plurality of and the four-way amount of function, its intermediate value is the wastewater treatment scheme that 1 element representation is selected.
In above-mentioned inorganic waste water reclamation movement expert system; Also comprise an expert knowledge library; Make up input training battle array in order to expert's scheme rule list and train, and utilize the δ learning algorithm of error back propagation to train said first weighted value matrix and the said second weighted value matrix with output training battle array according to prevision.
In above-mentioned inorganic waste water reclamation movement expert system, also comprise data analysis module, water quality and the quality of reused water index after comparison is handled also provides analytical results.
In above-mentioned inorganic waste water reclamation movement expert system, above-mentioned automatic control module and above-mentioned expert diagnosis decision-making module are incorporated in the computingmachine.
Inorganic waste water reclamation movement expert system of the present invention has changed the obtain manner of traditional waste water resource recovery technology scheme; No longer formulate and implement scheme, so scheme determination cycle and safety all there is obvious lifting with personal experience or the little test result in laboratory that relies on a small amount of waste water.And the perfect nigration car of hardware and software can not limited by the region, and getting into the production scene easily is the test that object carries out pilot scale with the actual waste water, finally for enterprise valuable waste water reclaiming technology and economic feasibility scheme is provided.
Description of drawings
For let above-mentioned purpose of the present invention, feature and advantage can be more obviously understandable, elaborate below in conjunction with the accompanying drawing specific embodiments of the invention, wherein:
Fig. 1 moves the expert system structure block diagram according to an embodiment of the invention.
Fig. 2 is a testing installation structure iron according to an embodiment of the invention.
Fig. 3-Fig. 7 is the instance that testing installation of the present invention makes up according to the wastewater treatment scheme.
Fig. 8 is the neural network structure figure of expert diagnosis decision-making module according to an embodiment of the invention.
Fig. 9 is the neural network structure figure that has the expert diagnosis decision-making module of a plurality of sub-networks according to an embodiment of the invention.
Embodiment
Inorganic trade effluent of a great variety; Be difficult to apply with the fixed technical scheme; Must formulate special waste water reclamation scheme to the character of different industry, different waste water and the water quality standard of reuse water, and drafting of scheme must be the basis with the scientific experimentation.Laboratory scale test can only obtain rough scheme owing to receive condition restriction (like wastewater flow rate and change of water quality etc.), let enterprise accept the technical scheme of waste water reclaiming, and it is not enough only depending on laboratory test data, hopes to obtain the data of pilot scale at least.
And a large amount of waste water transports the laboratory to is unpractical in operation; Therefore; Design of the present invention is to design the mobile inorganic waste water reclamation movement expert system of a cover, is equivalent to the mobile laboratory, in the mobile laboratory, settles various WQM instruments, separating device, chemical treatment and medicine machine as required; The testing installation commonly used of Industrial Wastewater Treatment such as film treatment facility and reuse, the enterprise that can get into dispersed placement carries out that the scene is detected, analyzed, test and assessment.Characteristics of the present invention are that testing installation can make up according to different wastewater treatment schemes at any time, so that carry out corresponding waste water treatment process.
Another characteristics of the present invention are practical situation such as to require according to the waste water quality of enterprise and reuse; Robotization and proposition promptly are fit to waste water recycling and the resource recovery process and the technical scheme of enterprise practical situation; Scheme Selection has tens kinds; And can constantly expand, the choice of enterprise is increased greatly.If the water quality after the scheme of selecting is handled can not reach requirement of client, handover scheme continues test at once, till the effluent quality of handling satisfies customer requirement fully.
Fig. 1 moves the expert system structure block diagram according to an embodiment of the invention.Locomotive body 100, testing installation 110, robotization control module 130, expert diagnosis decision-making module 140, expert knowledge library 150 and data analysis module 160.
These self-acting valves in the testing installation 110 link to each other with robotization control module 130, are controlled by robotization control module 130.Robotization control module 130 comprises the auto-control software that is installed in the computingmachine; It is according to switching different equipment combinations from the wastewater treatment scheme control test equipment of expert diagnosis decision-making module 140 110, carry out the conversion of multiple different waste water treatment process.
The operation of the armamentarium in the locomotive can realize full automatic control.Automatically controlled master touch-screen cashier's office in a shop can start or close all sewage treatment equipments in the car.Operation behind the device start is full-automatic, need not manual operation.
Expert diagnosis decision-making module 140 is decision-making parts of whole locomotive.After the water quality of waste water detects through water instrument 128; Waste water quality index and quality of reused water index are input to expert diagnosis decision-making module 140; Expert diagnosis decision-making module 140 confirms in view of the above and exports preliminary wastewater treatment scheme that scheme can be one or more.
Among the embodiment below, expert diagnosis decision-making module 140 utilizes neural network structure to make up.Utilize the outstanding self study adaptive ability of artificial neural network technology, network distributes and stores knowledge, and concurrent operation characteristic and superior non-linear mapping capability realize the automatic decision of wastewater treatment scheme.
Get back to shown in Figure 1ly, expert diagnosis decision-making module 140 connects expert knowledge libraries 150.In one embodiment, expert knowledge library 150 also can connect a data analysis module 160.
To the character of the inorganic wastewater of different industry, different process and the water quality standard of different treatment water,, can formulate sophisticated wastewater treatment scheme in advance according to existing wastewater recycling process.These ripe schemes and Expert Rules scheme table are stored in the expert knowledge library 150, supply expert diagnosis decision-making module 140 to call.Expert diagnosis decision-making module 140 can pass through these scheme knowledge of training study.Use the knowledge of having learnt, expert diagnosis decision-making module 140 can be confirmed preliminary inorganic wastewater processing scheme according to the inorganic wastewater water-quality guideline and the desired quality of reused water index of input from existing wastewater treatment scheme.
The structure of at first describing the expert diagnosis decision-making module of one embodiment of the invention below realizes.
Expert diagnosis decision-making module 140 is made up of neural network structure as shown in Figure 8, and the neural network of this expert diagnosis decision-making module 140 adopts feedforward network, can be divided into four layers, is respectively input layer 141, pretreatment layer 142, hidden layer 143 and output layer 144.These layers can be realized by computer executable program.
In one embodiment, each layer neuron number confirmed as follows: the water-quality guideline that inorganic wastewater is set has the R item, and then the neuron number of input layer each item water-quality guideline of equaling Inlet and outlet water (waste water and reuse water) is counted sum, is 2R; The neurone of pretreatment layer is to divide and decide (following detailed description) at the different condition section in when decision-making by the Inlet and outlet water index; And two times of the desirable pretreatment layer node of the neuron number of hidden layer number; It is S that the neuron number of output layer equals expert's scheme number.
Because expert's scheme of difference in functionality is only relevant with different piece Inlet and outlet water index; In one embodiment; For the pace of learning and simplification network structure that improves neural network; Can a macroreticular be divided into a plurality of sub-networks according to the Inlet and outlet water index classification of association, thus the expert diagnosis decision-making module that to generate a scale be the plurality of subnets network.
For instance; When 13 schemes are only relevant with 2 Inlet and outlet water indexs; Shown in table one, an input layer then can setting up is 2 neurones (specific conductivity of promptly intaking with go out water conductivity), and output layer is four layers of forward direction sub-network of 13 neurones (i.e. 13 schemes).Because the different range of the corresponding different Inlet and outlet water indexs of different processes scheme; Water inlet specific conductivity from 0 to 2999 μ m/cm in the table one is divided into 0-199,200-399,400-999,1000-2999 totally 4 sections, goes out water conductivity from 0.2 to 19 and be divided into 0.2-0.9,1-9,10-19 totally 3 sections.When if sub-network is only got into the water specific conductivity and gone out water conductivity and be 2 neurones of input layer 141, then to get the section sum be 7 to pretreatment layer 142 neuron numbers, and the neuron number of getting hidden layer 143 is 14.Other sub-network design by that analogy.
Table one expert scheme rule list
Accept above-mentionedly, when there was a plurality of sub-network in module, mode that can similar parallel connection connected.Fig. 9 illustrates the neural network structure figure of the expert diagnosis decision-making module with a plurality of sub-networks.The input of total input layer 211 of this expert diagnosis decision-making module 210 is whole Inlet and outlet water index parameter, and total input layer 211 each neuronic input connect each Inlet and outlet water index parameter respectively.The input of each child network 2121-212n connects again and comprises the neurone output of the input layer 211 of relevant Inlet and outlet water index parameter with it.Each child network can have similar structure shown in Figure 2, comprises input layer, pretreatment layer, hidden layer and output layer.The work output of the total output layer 213 of module is whole schemes number, and total output layer 213 each neuronic output correspond to scenarios number.And the output of each child network connects the scheme relevant with these sub-networks number neuronic input of pairing those total output layers.Thereby realize the expert diagnosis decision-making module that a scale is the plurality of subnets network.
Be example with single network shown in Figure 8 below, describe the inference mechanism of expert diagnosis decision-making module 140.
Expert diagnosis decision-making module 140 adopts the forward reasoning methods, by the input as system of each item water-quality guideline of Inlet and outlet water (waste water and reuse water), carries out reasoning through the neural network propagated forward, and the network output of calculating is decision scheme number.Concrete forward reasoning method is following:
A, b are threshold value in the formula, are decided by concrete scheme rule.
So each water-quality guideline x
iAll can be divided into one or more normalizing value y
jThereby, obtain secondary vector Y={y
1..., y
j..., y
T, j=1,2 ..., T.T is the neuronal quantity of pretreatment layer 142, and it depends on the processing of branch section.
The neurone of hidden layer 143 is input as the weighting sum of all pretreatment layer outputs:
W wherein
JkBe first weighted value, all w
JkForm the first weighted value matrix W
1w
JkInitial value can be set arbitrarily.After the study of the neural network that warp is stated later, w
JkTo level off to accurately.k=1,2,...,2T。2T is the neuronal quantity of hidden layer.The neuronal quantity that it is pointed out that hidden layer 143 is not defined as 2 times of pretreatment layer.
Nonlinear function is adopted in the neurone output of hidden layer 143, Sigmoid excitation function (also claiming the S type function) for example commonly used:
In another example, the output of the neurone of hidden layer 143 also can be adopted RBF.
The neuronic of output layer 144 is output as:
W wherein
KlBe second weighted value, all w
KlForm the second weighted value matrix W
2w
KlInitial value can be set arbitrarily.After the study of the neural network that warp is stated later, w
KlTo level off to accurately.
But passing threshold type excitation function further makes the neuronic of output layer 144 be output as:
C is a threshold value in the formula, between the desirable 0.8-0.9.Four-way amount A is formed in these outputs
; 1=1 wherein; 2 ..., S; S is the neuronal quantity of output layer, also is expert's scheme quantity of module.For each neurone output
1=1,2 ..., S, if
Corresponding expert's scheme number has been selected in expression.
approaches 1 more, can think that corresponding expert's scheme is number reliable more.
Be understood that easily for the module with a plurality of sub-networks as shown in Figure 9, the inference mechanism of its sub-network is similar.Difference only is that just from total input layer of network, get relevant with it part water-quality guideline carries out above-mentioned reasoning to each sub-network, exports simultaneously on the neurone relevant in corresponding expert's reuse scheme to total output layer of network.
To describe the foundation of the expert knowledge library of neural network expert diagnosis decision-making module below, the foundation of expert knowledge library comprises knowledge acquisition and two processes of knowledge store.
1) knowledge obtains
Obtaining of knowledge shows as obtaining and selecting of learning sample.According to existing expert decision-making result, list the into corresponding relation (being the Expert Rules scheme) of water index and effluent index and water technology scheme, can generate the input training battle array P and an output training battle array T of neural network.
If Expert Rules scheme table shown in aforementioned table one, has two condition projects in the table: water inlet specific conductivity and water outlet conductivity indices, be divided into 4 sections and 3 sections respectively, the water technology scheme has 13.If in the Expert Rules scheme table 13 rules are arranged, wherein rule 1 is: water inlet conductivity indices condition 400-2999 drops on the 3rd subregion, and water outlet conductivity indices condition 0.2-9 strides the 1st, 2 two subregion, and corresponding scheme as a result is No. 1.The raw column data of then corresponding neural network training battle array P and T battle array is shown in following table two.
Table two
The P battle array | The T battle array | |
Water inlet specific conductivity subregion | Go out the water conductivity subregion | Scheme number as a result |
1?2?3?4 | 1?2 | 1?2?3?4?5?6?7?8?9?10?11?12?13 |
0?0?1?0 | 1?0 | 1?0?0?0?0?0?0?0?0?0 0 0 0 |
0?0?1?0 | 0?1 | 1?0?0?0?0?0?0?0?0?00?0 0?0 |
...... | ...... | ...... |
Other all possible situation is all carried out similar processing, note guaranteeing the completeness and the expandability of Expert Rules, thereby generate the P battle array and the T battle array of training usefulness.
2) storage of knowledge
The expert decision-making knowledge store of neural network expert diagnosis decision-making module is implicitly to disperse to be stored in each neurone of neural network to connect in weights and the threshold value.The storage process of knowledge is exactly the learning process of neural network.
According to the neural network structure that design generates, the parameter that the neural network of expert decision-making can be learnt to adjust is the weight w of hidden layer
JkWeight w with output layer
KlAdopt the δ learning algorithm of error back propagation, adjust the weights of each interlayer, the learning algorithm that can derive neural network is following:
The error performance target function of p sample is:
Wherein N is the neuron number of network output layer.
According to the gradient descent method, the learning algorithm of weights is following:
The connection weights learning algorithm of output layer and hidden layer is:
The t+1 weights of network constantly is:
w
kl(t+1)=w
kl(t)+Δw
kl(t+1)?(9)
Hidden layer and pretreatment layer connect the weights learning algorithm:
Wherein
The k+1 weights of network constantly is:
w
jk(k+1)=w
jk(k)+Δw
jk(t+1) (12)
If consider the influence that last time, weights changed these weights, add factor of momentum, the weights of this moment are:
w
kl(k+1)=w
kl(k)+Δw
kl(t+1)+α(w
kl(k)-w
kl(k-1)) (13)
w
jk(t+1)=w
jk(t)+Δw
jk(t+1)+α(w
jk(t)-w
jk(t-1)) (14)
Wherein η is a learning rate, and α is a factor of momentum.Get α ∈ [0,1] η ∈ [0,1].
By the training battle array P and the T battle array that generate according to the Expert Rules table, neural network is through 5000 training studies, the error criterion function can reach E 0.02, accomplish the foundation of neural network expert knowledge library at this moment.Via the reasoning decision-making of training back neural network, the requirement of its input/output relation and Expert Rules scheme table reaches in full accord.To train the back neural network to embed in the expert diagnosis decision-making module, operation result can reach consistent with expectation.
Therefore, above-mentioned neural network expert diagnosis decision-making module can carry out computing according to input water-quality guideline (being the waste water quality index) and output water-quality guideline (being the quality of reused water index), output wastewater treatment scheme.
Above-mentioned preliminary wastewater treatment scheme can be exported to automatic control module 130, forms the corresponding apparatus combination by automatic control module 130 control test equipment 110, so that carry out waste water treatment process, and obtains the actual water-quality guideline of reuse water.The water-quality guideline that requires of the actual water-quality guideline of reuse water and reuse water is understood input data analysis module 160, carries out the analysis of actual processing effect and the assessment of result.If treatment effect can reach the expection standard, the scheme that then obtains according to expert diagnosis decision-making module 140 is the solution that system draws.If treatment effect is undesirable, other scheme that then obtains at expert diagnosis decision-making module 140 makes an experiment again, and compares and draw rational preferred plan.Carry out so repeatedly, till obtaining satisfied solution.In addition, can also make amendment to the wastewater treatment scheme in the expert knowledge library 120, produce the wastewater treatment scheme scheme of optimizing, and be stored in the expert knowledge library, offer enterprise's reference simultaneously as new scheme according to analysis and assessment result.
Rule of thumb data in the process of the test all runs up in the expert knowledge library 150, and 150 pairs of expert diagnosis decision-making modules 140 of expert knowledge library provide the prioritization scheme data resource, and are perfect so that it is replenished, and constantly increases the safety of expert diagnosis decision-making module.
In practical application, automatic control module 130, expert diagnosis decision-making module 140 all can be the software that is installed on same computingmachine, can realize linking of expert diagnosis decision-making module 140 and automatic control module 130, further improve the efficient of test in place.
Waste water reclaiming nigration car of the present invention can be those has waste water reclaiming to recycle wish; But for want of the technology enterprise of hesitating about what move to make with economic feasibility foundation provides the test in place service, for the investment decision of these enterprises provides quite valuable feasible scheme.Lifting some application examples below describes:
Application examples one:
The part Cleaning Wastewater that produces in certain alumina production, the waste water slant acidity about specific conductivity 800 μ S/cm, at present as wastewater treatment, is not recycled, and therefore hopes to recycle this waste water, requires the reuse water specific conductivity to be superior to the level of tap water.Requirement in view of the above at first with expert diagnosis decision-making module 140 interfaces in the data input computingmachines such as waste water quality parameter and reuse water-quality guideline, obtains " electrodialysis " and " r-o-" two testing programs of output.The waste water that produces in actual production in the production scene according to testing program nigration car has carried out two groups of experiments, and its experiment process is respectively:
The electrodialysis scheme is: former water → pH regulator → multi-medium filtering → activated carbon filtration → secondary filter → electrodialysis → reuse.The equipment combination synoptic diagram of this scheme is please with reference to Fig. 3.
The r-o-scheme is: former water → pH regulator → multi-medium filtering → activated carbon filtration → dosing → secondary filter → ultra-filtration equipment → first-stage reverse osmosis → reuse.The equipment combination synoptic diagram of this scheme is please with reference to Fig. 4.
Table three illustrates the test-results of two kinds of schemes.The result of two groups of tests shows that the r-o-scheme obviously is superior to the electrodialysis scheme aspect technical indicator, and economic target aspect difference of them is little.It in the form comparison of testing two kinds of scheme critical technical parameters that draw.Take all factors into consideration technology and economic feasibility, adopting the r-o-scheme is better selection.
Table three experimental technique data relatively
<tables num="0002"><table ><tgroup cols="4"><colspec colname="c001" colwidth="27%" /><colspec colname="c002" colwidth="13%" /><colspec colname="c003" colwidth="14%" /><colspec colname="c004" colwidth="46%" /><tbody ><row ><entry morerows="1">Scheme</entry><entry morerows="1">Scheme one (electrodialysis)</entry><entry morerows="1">Scheme two (r-o-)</entry><entry morerows="1">Remarks</entry></row><row ><entry morerows="1">Ratio of desalinization (%)</entry><entry morerows="1">68.7</entry><entry morerows="1">98.46</entry><entry morerows="1" /></row><row ><entry morerows="1">The recovery (%)</entry><entry morerows="1"> 41 </entry><entry morerows="1"> 50 </entry><entry morerows="1">Dense water cycle is adopted in electrodialysis, and the waste water reclamation rate can be brought up to about 65%.</entry></row></tbody></tgroup></table></tables>Production-scale r-o-can reach 70% the recovery.</entry></row><row ><entry morerows=" 1 ">Reuse water quality (μ S/cm)</entry><entry morerows=" 1 ">223</entry><entry morerows=" 1 ">9</entry><entry morerows="1" /></row></tbody></tgroup></table></tables>
Application examples two:
Adopt r-o-(RO) technology to produce in the technology of pure water, the higher dense water of a large amount of saltiness is arranged every day, be water saving, this part water of expectation reuse as discharge of wastewater.Mobile expert systems has been carried out test in place.Through computer control valve is automatically switched to the two-pass reverse osmosis experiment process:
Dense water → the multi-medium filtering of RO → activated carbon filtration → dosing → secondary filter → ultrafiltration → first-stage reverse osmosis → two-pass reverse osmosis → pure water reuse.The equipment combination structure synoptic diagram of this scheme sees also Fig. 5.
Testing data shows that specific conductivity surpasses and goes out water conductivity after the dense water of the RO of 1000 μ S/cm is handled through above experiment process and can reach about 5 μ S/cm; Meet the pure water standard, adopt this scheme not only to can be pure water manufacturing enterprise saving tap water but also can enlarge the pure water production capacity.
Application examples three:
Certain waste water reaches emission standard after treatment, but saltiness is higher, and specific conductivity surpasses 5000 μ S/cm, can not repeat reuse.According to the water quality requirement of reuse water, the testing program of expert diagnosis decision-making module 140 output first-stage reverse osmosis, automatic control module 130 according to scheme transfer test flow process is:
Waste water → multi-medium filtering → activated carbon filtration → secondary filter → ultrafiltration → first-stage reverse osmosis → reuse.The equipment combination structure synoptic diagram of this scheme sees also Fig. 6.
The factory effluent that test-results shows specific conductivity 5000 μ S/cm specific conductivity after treatment can drop to 180 μ S/cm, and ratio of desalinization reaches more than 96%, and water quality is better than tap water.
Application examples four:
Move expert systems and carried out test in place for certain acid copper-plating waste water recycling.This kind waste water ph about 2 contains cupric ion, and it mainly is the copper ion concentration that reduces in the waste water that waste water recycling requires, and keeps reuse water pH closely neutral.The fundamental test flow process that moves expert systems does not comprise this type of waste water.Because the extendability of waste water reclamation scope has been considered in the design of testing installation, therefore after having done minor modifications, developed new experiment process.Experiment process is:
Acid copper-plating waste water → pH regulator → multi-medium filtering → activated carbon filtration → Copper Ion Exchange → Copper Ion Exchange → reuse
The change part of basic test flow process is that the strongly basic anion exchange resin with the strongly acidic cationic exchange resin of former cation exchange bed 122 and anion exchange bed 123 changes into can absorbing copper ionic resin.Get into exchange resin bed 122,123, the 99% above cupric ions of having changed resin by flow process behind the acid copper-containing wastewater process pH regulator device and be adsorbed, go out water concentration, return the production line reuse less than 0.5mg/l.
In sum, inorganic waste water reclamation movement expert system of the present invention has changed the obtain manner of traditional waste water resource recovery technology scheme, no longer formulates and implements scheme with personal experience or the little test result in laboratory that relies on a small amount of waste water.The perfect nigration car of hardware and software can not limited by the region, and getting into the production scene easily is that object is optimized test with the actual waste water, finally for enterprise valuable waste water reclaiming technology and economic feasibility scheme is provided.
In addition, the present invention utilizes the outstanding self study adaptive ability of artificial neural network technology, and network distributes and stores knowledge, and concurrent operation characteristic and superior non-linear mapping capability realize the automatic decision of wastewater treatment scheme.Knowledge acquisition problem, the learning capacity of normal expert system are relatively poor because neural network expert system has overcome to a certain extent, " narrow step effect " knowledge is the bottleneck problems such as contradiction of poor fault tolerance and knowledge storage capacity and travelling speed; Thereby further improve expert systems learning capacity and the ability of handling large complicated problem; Make knowledge base have good expandability, the operation of system has higher safety.
Though the present invention discloses as above with preferred embodiment; Right its is not that any those skilled in the art are not breaking away from the spirit and scope of the present invention in order to qualification the present invention; When can doing a little modification and perfect, so protection scope of the present invention is when being as the criterion with what claims defined.
Claims (5)
1. inorganic waste water reclamation movement expert system comprises:
The locomotive body comprises the envelope that places on the transportation chassis;
Testing installation is suitable for forming the plurality of devices combination, and wherein each equipment combination is carried out wastewater treatment according at least a waste water treatment process in said envelope;
Automatic control module; Control said testing installation according to the wastewater treatment scheme and form the corresponding apparatus combination and make said equipment combination carry out waste water treatment process, wherein said automatic control module is to switch said testing installation through self-acting valve to form said equipment combination; And
The expert diagnosis decision-making module is confirmed and is exported at least one wastewater treatment scheme to said automatic control module according to waste water quality index and quality of reused water index,
Wherein said expert diagnosis decision-making module comprises:
Input layer, input comprises primary vector to a pretreatment layer of said waste water quality index and said quality of reused water index;
Pretreatment layer carries out the processing of branch section to said waste water quality index in the said primary vector and said quality of reused water index, and the output secondary vector;
Hidden layer according to each element sum of the said secondary vector of the trained first weighted value matrix computations, and calculates a plurality of said and functions under a non-linear excitation respectively, export comprise a plurality of said and the 3rd vector of function; And
Output layer; Each element sum according to said the 3rd vector of the trained second weighted value matrix computations; And calculate a plurality of said and functions under threshold-type excitation respectively; Output comprise a plurality of said and the four-way amount of function, its intermediate value is the wastewater treatment scheme that 1 element representation is selected.
2. inorganic waste water reclamation movement expert system as claimed in claim 1; It is characterized in that said testing installation comprises at least a portion with lower component: WQM instrument and instrument, pH regulator equipment, pre-processing device, ion exchange treatment equipment, membrane separation plant, medicine machine.
3. inorganic waste water reclamation movement expert system as claimed in claim 1; It is characterized in that; Also comprise an expert knowledge library; Make up input training battle array in order to expert's scheme rule list and train, and utilize the δ learning algorithm of error back propagation to train said first weighted value matrix and the said second weighted value matrix with output training battle array according to prevision.
4. inorganic waste water reclamation movement expert system as claimed in claim 1 is characterized in that, also comprises data analysis module, and water quality and the quality of reused water index after comparison is handled also provides analytical results.
5. inorganic waste water reclamation movement expert system as claimed in claim 1 is characterized in that, said automatic control module and said expert diagnosis decision-making module are incorporated in the computingmachine.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2008102006632A CN101684008B (en) | 2008-09-27 | 2008-09-27 | Inorganic waste water reclamation movement expert system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2008102006632A CN101684008B (en) | 2008-09-27 | 2008-09-27 | Inorganic waste water reclamation movement expert system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101684008A CN101684008A (en) | 2010-03-31 |
CN101684008B true CN101684008B (en) | 2012-06-06 |
Family
ID=42047433
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2008102006632A Expired - Fee Related CN101684008B (en) | 2008-09-27 | 2008-09-27 | Inorganic waste water reclamation movement expert system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101684008B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104150663A (en) * | 2014-08-31 | 2014-11-19 | 晨氏(天津)科技有限公司 | Ultrapure water producing system |
CN104556590B (en) * | 2015-01-30 | 2016-06-01 | 广州粤康环保工程有限公司 | A kind of environment protection sewage treatment unit and sewage water treatment method |
CN105700384A (en) * | 2016-03-14 | 2016-06-22 | 东北大学 | Method of building wastewater pollutant discharge Petri net simulation model |
WO2019079883A1 (en) * | 2017-10-24 | 2019-05-02 | Waterloo Controls Inc. | Systems and methods for detecting waste receptacles using convolutional neural networks |
CN109901542B (en) * | 2019-04-10 | 2020-08-07 | 上海上实龙创智能科技股份有限公司 | Sewage treatment method based on Internet of things, cloud platform and system |
CN112624451A (en) * | 2020-12-04 | 2021-04-09 | 沈阳化工大学 | Photocatalytic multistage membrane separation coupling sewage treatment system |
CN112765951B (en) * | 2021-01-20 | 2023-04-18 | 宁夏计量质量检验检测研究院 | System for formulating environment and new energy field standard |
CN115159769B (en) * | 2022-06-14 | 2023-09-26 | 哈尔滨泽能环保科技有限公司 | Medical sewage treatment system and process thereof |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002186956A (en) * | 2000-12-19 | 2002-07-02 | Sanyo Electric Co Ltd | Wastewater treatment system |
CN2659880Y (en) * | 2003-11-19 | 2004-12-01 | 韩克江 | Mobile acidized waste fluid treatment plant for oil well |
JP2007000722A (en) * | 2005-06-22 | 2007-01-11 | Seiko Epson Corp | Chemical substance treatment system, and chemical substance treatment method |
CN1994928A (en) * | 2006-12-01 | 2007-07-11 | 华中科技大学 | Movable and integral drinking water purifier |
CN101084328A (en) * | 2004-11-02 | 2007-12-05 | 水处理有限责任公司 | Electronic components associated and apparatus for deionization and electrochemical purification and regeneration of electrodes |
CN101229949A (en) * | 2007-12-19 | 2008-07-30 | 核工业理化工程研究院华核新技术开发公司 | Mobile radioactive liquid waste treatment equipment |
CN201309838Y (en) * | 2008-09-27 | 2009-09-16 | 上海轻工业研究所有限公司 | Inorganic wastewater recycling mobile expert system |
-
2008
- 2008-09-27 CN CN2008102006632A patent/CN101684008B/en not_active Expired - Fee Related
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002186956A (en) * | 2000-12-19 | 2002-07-02 | Sanyo Electric Co Ltd | Wastewater treatment system |
CN2659880Y (en) * | 2003-11-19 | 2004-12-01 | 韩克江 | Mobile acidized waste fluid treatment plant for oil well |
CN101084328A (en) * | 2004-11-02 | 2007-12-05 | 水处理有限责任公司 | Electronic components associated and apparatus for deionization and electrochemical purification and regeneration of electrodes |
JP2007000722A (en) * | 2005-06-22 | 2007-01-11 | Seiko Epson Corp | Chemical substance treatment system, and chemical substance treatment method |
CN1994928A (en) * | 2006-12-01 | 2007-07-11 | 华中科技大学 | Movable and integral drinking water purifier |
CN101229949A (en) * | 2007-12-19 | 2008-07-30 | 核工业理化工程研究院华核新技术开发公司 | Mobile radioactive liquid waste treatment equipment |
CN201309838Y (en) * | 2008-09-27 | 2009-09-16 | 上海轻工业研究所有限公司 | Inorganic wastewater recycling mobile expert system |
Non-Patent Citations (1)
Title |
---|
彭芳.污水智能控制***的研究与实现.《中国优秀博硕士学位论文全文数据库(硕士)工程科技I辑》.2005, * |
Also Published As
Publication number | Publication date |
---|---|
CN101684008A (en) | 2010-03-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101684008B (en) | Inorganic waste water reclamation movement expert system | |
Kamali et al. | Artificial intelligence as a sustainable tool in wastewater treatment using membrane bioreactors | |
Bagheri et al. | Modeling and optimization of activated sludge bulking for a real wastewater treatment plant using hybrid artificial neural networks-genetic algorithm approach | |
de Canete et al. | Soft-sensing estimation of plant effluent concentrations in a biological wastewater treatment plant using an optimal neural network | |
Teodosiu et al. | Neural network models for ultrafiltration and backwashing | |
Design | Water treatment plant design | |
Ray et al. | A holistic review on how artificial intelligence has redefined water treatment and seawater desalination processes | |
Okhovat et al. | Modeling of arsenic, chromium and cadmium removal by nanofiltration process using genetic programming | |
Baxter et al. | Development of a full-scale artificial neural network model for the removal of natural organic matter by enhanced coagulation | |
CN110322933A (en) | A kind of polypropylene melt index hybrid modeling method based on dynamic error compensation mechanism | |
Al-Shayji | Modeling, simulation, and optimization of large-scale commercial desalination plants | |
CN201309838Y (en) | Inorganic wastewater recycling mobile expert system | |
AU2018418038B2 (en) | System and method for predicting a parameter associated with a wastewater treatment process | |
Ammi et al. | Stacked neural networks for predicting the membranes performance by treating the pharmaceutical active compounds | |
Baig et al. | Ensemble hybrid machine learning to simulate dye/divalent salt fractionation using a loose nanofiltration membrane | |
Bui et al. | Application of artificial neural networks on water and wastewater prediction: A review | |
CN101685506B (en) | Expert diagnosis decision method of inorganic waste water processing scheme | |
Muniappan et al. | Deep convolutional neural network with sine cosine algorithm based wastewater treatment systems | |
Dologlu et al. | Data driven identification of industrial reverse osmosis membrane process | |
Vyas et al. | Artificial neural network based model in effluent treatment process | |
Buaisha et al. | ANN-assisted forecasting of adsorption efficiency to remove heavy metals | |
Stanley | Process modeling and control of enhanced coagulation | |
Strugholtz et al. | Neural networks and genetic algorithms in membrane technology modelling | |
Han et al. | Knowledge-based fuzzy broad learning algorithm for warning membrane fouling | |
Amini | The sustainability of ion exchange water treatment technology |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120606 Termination date: 20170927 |
|
CF01 | Termination of patent right due to non-payment of annual fee |