CN115196730A - Intelligent sodium hypochlorite adding system for water plant - Google Patents
Intelligent sodium hypochlorite adding system for water plant Download PDFInfo
- Publication number
- CN115196730A CN115196730A CN202210847866.0A CN202210847866A CN115196730A CN 115196730 A CN115196730 A CN 115196730A CN 202210847866 A CN202210847866 A CN 202210847866A CN 115196730 A CN115196730 A CN 115196730A
- Authority
- CN
- China
- Prior art keywords
- chlorine
- chlorination
- model
- water plant
- water
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 201
- 239000005708 Sodium hypochlorite Substances 0.000 title claims abstract description 62
- SUKJFIGYRHOWBL-UHFFFAOYSA-N sodium hypochlorite Chemical compound [Na+].Cl[O-] SUKJFIGYRHOWBL-UHFFFAOYSA-N 0.000 title claims abstract description 62
- 239000000460 chlorine Substances 0.000 claims abstract description 135
- 229910052801 chlorine Inorganic materials 0.000 claims abstract description 135
- ZAMOUSCENKQFHK-UHFFFAOYSA-N Chlorine atom Chemical compound [Cl] ZAMOUSCENKQFHK-UHFFFAOYSA-N 0.000 claims abstract description 131
- 238000005660 chlorination reaction Methods 0.000 claims abstract description 98
- 238000000034 method Methods 0.000 claims abstract description 27
- 230000001502 supplementing effect Effects 0.000 claims abstract description 23
- 238000012549 training Methods 0.000 claims description 21
- 238000004062 sedimentation Methods 0.000 claims description 13
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 12
- 239000001301 oxygen Substances 0.000 claims description 12
- 229910052760 oxygen Inorganic materials 0.000 claims description 12
- 238000012935 Averaging Methods 0.000 claims description 5
- 239000000126 substance Substances 0.000 claims description 5
- 238000003062 neural network model Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 7
- 230000000875 corresponding effect Effects 0.000 description 15
- 230000001276 controlling effect Effects 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 5
- 238000004659 sterilization and disinfection Methods 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 3
- 230000035772 mutation Effects 0.000 description 3
- 241000195493 Cryptophyta Species 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012797 qualification Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- VHUUQVKOLVNVRT-UHFFFAOYSA-N Ammonium hydroxide Chemical compound [NH4+].[OH-] VHUUQVKOLVNVRT-UHFFFAOYSA-N 0.000 description 1
- 241000894006 Bacteria Species 0.000 description 1
- 241000700605 Viruses Species 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 239000003513 alkali Substances 0.000 description 1
- XKMRRTOUMJRJIA-UHFFFAOYSA-N ammonia nh3 Chemical compound N.N XKMRRTOUMJRJIA-UHFFFAOYSA-N 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000006227 byproduct Substances 0.000 description 1
- 239000000084 colloidal system Substances 0.000 description 1
- -1 contact pool PH Chemical compound 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- QWPPOHNGKGFGJK-UHFFFAOYSA-N hypochlorous acid Chemical compound ClO QWPPOHNGKGFGJK-UHFFFAOYSA-N 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- IJGRMHOSHXDMSA-UHFFFAOYSA-N nitrogen Substances N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 1
- 229910052757 nitrogen Inorganic materials 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000007800 oxidant agent Substances 0.000 description 1
- 230000001590 oxidative effect Effects 0.000 description 1
- 150000004965 peroxy acids Chemical class 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000029219 regulation of pH Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001954 sterilising effect Effects 0.000 description 1
- 230000009469 supplementation Effects 0.000 description 1
Images
Classifications
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F1/00—Treatment of water, waste water, or sewage
- C02F1/50—Treatment of water, waste water, or sewage by addition or application of a germicide or by oligodynamic treatment
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F1/00—Treatment of water, waste water, or sewage
- C02F1/008—Control or steering systems not provided for elsewhere in subclass C02F
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F1/00—Treatment of water, waste water, or sewage
- C02F1/72—Treatment of water, waste water, or sewage by oxidation
- C02F1/76—Treatment of water, waste water, or sewage by oxidation with halogens or compounds of halogens
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D21/00—Control of chemical or physico-chemical variables, e.g. pH value
- G05D21/02—Control of chemical or physico-chemical variables, e.g. pH value characterised by the use of electric means
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2201/00—Apparatus for treatment of water, waste water or sewage
- C02F2201/002—Construction details of the apparatus
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Hydrology & Water Resources (AREA)
- Environmental & Geological Engineering (AREA)
- Water Supply & Treatment (AREA)
- Chemical & Material Sciences (AREA)
- Organic Chemistry (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Treatment Of Water By Oxidation Or Reduction (AREA)
Abstract
The invention provides an intelligent sodium hypochlorite adding system for a water plant, which relates to the technical field of intelligent control of the water plant and comprises the following components: intelligently controlling the pre-chlorination amount of the sodium hypochlorite at each first feeding point according to a first proportional integral differential control model, and collecting corresponding pre-chlorination residual chlorine; inputting a water quality parameter set containing each front chlorine adding amount and the corresponding front chlorine adding residual chlorine into a rear chlorine adding model to obtain a rear chlorine adding amount, intelligently controlling sodium hypochlorite added at each second adding point and the rear chlorine adding amount, and collecting the corresponding rear chlorine adding residual chlorine; and intelligently controlling the chlorine supplementing amount of the sodium hypochlorite at each third feeding point according to a second proportional-integral-derivative control model, wherein the second proportional-integral-derivative control model takes the front chlorine supplementing amount, the front chlorine supplementing residual amount, the rear chlorine supplementing amount and the rear chlorine supplementing residual amount as feedforward parameters. The method has the advantages that the post-chlorination adding amount can be dynamically adjusted according to the pre-chlorination effect, the chlorine adding amount can be dynamically adjusted and supplemented according to the pre-chlorination and post-chlorination adding effects, and the three adding points are linked to accurately add chlorine.
Description
Technical Field
The invention relates to the technical field of intelligent control of water plants, in particular to an intelligent sodium hypochlorite adding system of a water plant.
Background
The conventional chlorination link of the water plant is generally divided into pre-chlorination, post-chlorination and supplementary chlorination. The chlorine is added at the raw water inlet part mainly according to the raw water flow proportion, the raw water flow and the recovery flow are comprehensively considered, and the sodium hypochlorite adding amount is in direct proportion to the water flow. The chlorine is added in the front mainly to play the roles of disinfection, algae removal, PH regulation, water colloid destruction and the like. The traditional chlorine adding method is characterized in that a flowmeter is arranged, the flow is converted into a current signal and the current signal is input into a PLC (programmable logic controller), and the purpose of chlorine adding before flow proportion control is achieved. The traditional chlorination method is relatively extensive, and the PLC can only generally carry out the chlorination according to flow indexes or 2-3 raw water indexes including flow. The actual sodium hypochlorite dosage may be related to a plurality of complex water quality parameters such as raw water flow, raw water turbidity, raw water temperature, raw water dissolved oxygen, raw water ammonia nitrogen and the like. Limited to software and hardware, the traditional chlorination process cannot comprehensively consider the parameters.
Then adding chlorine before the clean water tank, and using strong oxidant (hypochlorous acid) produced by interaction of chlorine and water to kill residual bacteria and virus in water. The chlorine is added later, so that the main link of ensuring the residual chlorine of the factory water is directly related to the qualification rate and the stability of the residual chlorine of the factory water. In a post-chlorination process, conventional processes typically use a PID complex loop control of a flow proportional feed forward signal and a residual chlorine feedback feed back signal. The flow rate proportion feedforward signal is similar to the pre-chlorination, namely, the flow rate signal is converted into an electric signal through a flow meter and is input into the PLC. The residual chlorine feedback signal is that a residual chlorine meter is installed at the position 5-10 minutes after the post-chlorination feeding point, and the residual chlorine meter is used for measuring the residual chlorine amount and then is used as a post-feedback signal to be combined with the flow rate to adjust the chlorination amount so as to ensure the post-chlorination effect.
Two main problems exist in the traditional post-chlorination process. First, as with the pre-chlorination, the conventional post-chlorination only takes into account a few feedback signals, including flow rate and residual chlorine, due to limitations of the PLC equipment. However, under extreme conditions such as rainstorm and debris flow weather, the water quality indexes such as turbidity, PH and the like are changed sharply along with the flow change. These drastic water quality indicators are not in the conventional process of considering post-chlorination, but greatly affect the dosage of sodium hypochlorite. Secondly, the post-chlorination control process is used as an independent control unit and is completely independent from the pre-chlorination control unit. In practical application, the amount of the added chlorine is strongly correlated with the amount of the added chlorine. For example, the disinfection and sterilization effects of structures such as a sedimentation tank, a filter and the like are ensured under the condition of perfect addition of the pre-chlorination, and meanwhile, chlorination byproducts cannot be generated due to excessive addition. But it is difficult to achieve this effect in actual production.
The supplementary chlorine is usually positioned before leaving factory, and the supplementary chlorine is used as the last gate before chlorine disinfection leaving factory, which is directly related to the qualification rate and stability of the residual chlorine leaving factory. The traditional chlorine supplementing control usually uses the factory water flow as feedforward and the factory residual chlorine as feedback, and controls the chlorine adding amount through PID adjustment, so that the factory water residual chlorine is kept in a preset range, and the aim of disinfection is fulfilled. The current PID control chlorine adding process is the same as the post-chlorination process, and cannot be adjusted in association with the pre-chlorination step and the post-chlorination step. The independent and separated chlorine supplementing process cannot effectively feed back and adjust the dosage of the pre-added chlorine and the post-added chlorine with more or less chlorine.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an intelligent sodium hypochlorite adding system for water plants, wherein each water plant is provided with at least one first adding point arranged at the raw water inlet of the water plant, at least one second adding point arranged in front of a clean water tank of the water plant and at least one third adding point arranged behind the clean water tank;
then water plant sodium hypochlorite intelligence dosing system includes an at least water plant customer end, the water plant customer end with the water plant one-to-one, the water plant customer end includes:
the front chlorine adding control module is used for intelligently controlling a front chlorine adding amount of the sodium hypochlorite at each first adding point of the corresponding water plant according to a preset first proportional-integral-derivative control model and collecting front chlorine adding residual chlorine after the sodium hypochlorite is added at each first adding point;
the post-chlorination control module is connected with the pre-chlorination control module and used for inputting a water quality parameter set containing each pre-chlorination amount and the corresponding pre-chlorination residual chlorine into a post-chlorination model which is updated in a self-adaptive mode to obtain a post-chlorination amount, intelligently controlling the corresponding water plant to add the sodium hypochlorite with the post-chlorination amount at each second adding point, and collecting the post-chlorination residual chlorine after the sodium hypochlorite is added at each second adding point;
and the chlorine supplementing control module is respectively connected with the front chlorine adding control module and the rear chlorine adding control module and is used for intelligently controlling the chlorine supplementing quantity of the sodium hypochlorite at each third feeding point of the water plant according to a preset second proportional-integral-derivative control model, and the second proportional-integral-derivative control model takes the front chlorine adding quantity, the front chlorine adding residual quantity, the rear chlorine adding quantity and the rear chlorine adding residual quantity as feedforward parameters.
Preferably, still include a high in the clouds server, connect each the water plant client, the water plant client still includes a model update module, connects the back chlorination control module, the model update module includes:
the local training unit is used for continuously acquiring a plurality of groups of the front chlorine adding amount, the front chlorine adding allowance and the corresponding rear chlorine adding amount to form a training set, and training and updating the rear chlorine adding model according to the training set to obtain corresponding local model parameters;
the parameter uploading unit is connected with the local training unit and is used for uploading the local model parameters obtained by each training to the cloud server;
the cloud server is used for averaging the local model parameters to obtain a latest model parameter when the local model parameters of the water plants are received in each round, and sending the latest model parameter to all the model updating clients;
the model update module further comprises:
the updating unit is used for updating the post-chlorination model according to the latest model parameters when receiving the latest model parameters;
and the post-chlorination control module predicts and obtains the post-chlorination amount according to the updated post-chlorination model.
Preferably, the cloud server includes:
the delay module is used for outputting an end signal representing that the local model parameters in the current round are completely received when the local model parameters are not received again within a preset time period after the local model parameters are received for the last time in the process of receiving the local model parameters in each round;
and the computing module is connected with the delay module and used for averaging all the local model parameters received in the current round according to the ending signal to obtain the latest model parameter and sending the latest model parameter to all the water plant clients.
Preferably, a model network is pre-configured in the cloud server; the cloud server further comprises a configuration module, and the configuration module is used for configuring the latest model parameters and the model network into the newly accessed water plant client to form the post-chlorination model when detecting that the new water plant client is accessed.
Preferably, the feedforward parameters of the first proportional integral derivative control model include a total intake instantaneous flow of the water plant, intake instantaneous flows of sedimentation tanks of the water plant, a PH value of raw water, a temperature of the raw water, a turbidity of the raw water, a dissolved oxygen content of the raw water, and a chemical oxygen demand of the raw water, and the feedback parameters of the first proportional integral derivative control model are the residual amount of the front chlorine.
Preferably, the first adding points are arranged in the sedimentation tanks of the water plant, the front chlorine adding amount correspondingly comprises the adding amount of the sodium hypochlorite in the sedimentation tanks respectively, and the front chlorine adding surplus correspondingly comprises the residual chlorine after the sodium hypochlorite is added in the sedimentation tanks respectively.
Preferably, the water quality parameter set further includes turbidity of the raw water, turbidity of each contact tank of the water plant, and PH of each contact tank.
Preferably, the second adding points are arranged in the contact tanks, the post-chlorine adding amount respectively comprises the adding amount of the sodium hypochlorite in the contact tanks, and the post-chlorine adding allowance respectively comprises the residual chlorine after the sodium hypochlorite is added in the contact tanks.
Preferably, the post-chlorination model is a deep neural network model.
The technical scheme has the following advantages or beneficial effects:
1) The front chlorine adding, the rear chlorine adding and the chlorine supplementing full-process sodium hypochlorite adding are comprehensively considered, the rear chlorine adding amount can be dynamically adjusted according to the effect of the front chlorine adding, the chlorine supplementing amount can be dynamically adjusted according to the adding effects of the front chlorine adding and the rear chlorine adding, the three adding points are linked for accurate adding, the full-process accurate adding is really realized, and the medicament is saved;
2) The updated chlorination model is configured based on distributed deep learning, the post-chlorination model can be used by the newly-built water plant in a mode that the tested water plant with historical data shares model parameters with the newly-built water plant, and the model parameters are updated in the iteration process along with the accumulation of data of the newly-built water plant, so that the post-chlorination model is more accurate;
3) On the basis of distributed deep learning, each water plant client only uses the chlorine model after data training locally without uploading the data, and only the trained local model parameters are uploaded, so that on one hand, the model can absorb the experience of intelligent chlorine addition of each water plant, the model is more accurate, on the other hand, the original sensitive data can be ensured to be deployed locally, and the data safety is effectively protected;
4) The feedforward parameters of the first proportional integral derivative control model comprehensively consider multiple water quality parameters, and the accuracy of pre-chlorination is effectively improved.
Drawings
Fig. 1 is a schematic structural diagram of an intelligent sodium hypochlorite dosing system of a water plant in a preferred embodiment of the invention.
Detailed Description
The invention is described in detail below with reference to the figures and the specific embodiments. The present invention is not limited to the embodiment, and other embodiments may be made within the scope of the present invention as long as the gist of the present invention is satisfied.
In a preferred embodiment of the present invention, based on the above problems in the prior art, an intelligent sodium hypochlorite adding system for water plants is provided, in which each water plant is configured with at least one first adding point disposed at the raw water inlet of the water plant, at least one second adding point disposed in front of the clean water tank of the water plant, and at least one third adding point disposed behind the clean water tank;
as shown in fig. 1, the intelligent sodium hypochlorite dosing system of the water plant comprises at least one water plant client 1, the water plant client 1 corresponds to the water plant one by one, and the water plant client 1 comprises:
the front chlorine adding control module 11 is used for intelligently controlling the front chlorine adding amount of the sodium hypochlorite at each first adding point of the corresponding waterworks according to a preset first proportional-integral-differential control model, and collecting the front chlorine adding residual chlorine after the sodium hypochlorite is added at each first adding point;
the post-chlorination control module 12 is connected with the pre-chlorination control module 11 and used for inputting a water quality parameter set containing each pre-chlorination amount and corresponding pre-chlorination residual chlorine into a self-adaptively updated post-chlorination model to obtain a post-chlorination amount, intelligently controlling the corresponding waterworks to add sodium hypochlorite with the post-chlorination amount at each second adding point, and collecting the post-chlorination residual chlorine after the sodium hypochlorite is added at each second adding point;
and the chlorine supplementing control module 13 is respectively connected with the front chlorine supplementing control module 11 and the rear chlorine supplementing control module 12 and is used for intelligently controlling the chlorine supplementing amount of the sodium hypochlorite at each third feeding point of the corresponding waterworks according to a preset second proportional-integral-differential control model, and the second proportional-integral-differential control model takes the front chlorine supplementing amount, the front chlorine supplementing residual amount, the rear chlorine supplementing amount and the rear chlorine supplementing residual amount as feedforward parameters.
Specifically, considering that chlorination processes adopted by different water plants have extremely high similarity, the chlorination process is basically divided into three parts, namely front chlorination, rear chlorination and chlorine supplementation, and in the embodiment, model building is carried out based on the three parts. For the front chlorine adding part, a first proportional integral derivative control model can be configured in server hardware, and compared with the traditional front chlorine adding PID scheme which is limited by PLC hardware conditions and only can consider flow factors, the first proportional integral derivative control model comprehensively considers flow, PH value, temperature, turbidity, dissolved oxygen and chemical oxygen demand as feedforward parameters of the first proportional integral derivative control model, and takes front chlorine residual as feedback parameters, so that the sodium hypochlorite adding accuracy of the front chlorine adding part is effectively improved. The flow rate comprises the total water inlet instantaneous flow rate of the water plant and the water inlet instantaneous flow rate of each sedimentation tank of the water plant, the pH value is the raw water pH value, the temperature is the raw water temperature, the turbidity is the raw water turbidity, the dissolved oxygen is the raw water dissolved oxygen content, and the chemical oxygen demand is the raw water chemical oxygen demand.
And for the post-chlorination part, collecting a distributed deep learning mode and taking the pre-chlorination allowance and the predicted pre-chlorination amount as input parameters of post-chlorination. And for each water plant client, the corresponding local water quality data of the water plant and a local post-chlorination model are configured in the client.
In a preferred embodiment of the present invention, the present invention further includes a cloud server 2 connected to each water plant client 1, the water plant client 1 further includes a model updating module 14 connected to the post-chlorination control module 12, and the model updating module 14 includes:
the local training module 141 is configured to continuously obtain a plurality of groups of front chlorine adding amounts, front chlorine adding allowances and corresponding rear chlorine adding amounts to form a training set, and train and update the rear chlorine adding model according to the training set to obtain corresponding local model parameters;
the parameter uploading module 142 is connected to the local training module 141 and is configured to upload the local model parameters obtained in each training to the cloud server;
the cloud server 2 is used for averaging the local model parameters to obtain a latest model parameter when the local model parameters of each water plant are received in each round, and sending the latest model parameter to all model updating clients;
the model update module 14 further comprises:
the updating unit 143 is configured to update the chlorination model according to the latest model parameters when the latest model parameters are received;
and the post-chlorination control module 12 predicts and obtains the post-chlorination amount according to the updated post-chlorination model.
Specifically, in this embodiment, it is preferable that the cloud server 2 initializes a corresponding initial model and initial model parameters, the initial model and the initial model parameters may be allocated by an algorithm engineer, when the system is just started to use, the cloud server 2 may issue the initial model and the initial model parameters to each connected water plant client 1, each water plant client 1 may perform post-chlorination control according to an initial post-chlorination model formed by the initial model and the initial model parameters, and in a subsequent use process, as the data volume of local water quality parameter data is continuously abundant, each water plant client 1 may continuously train and update the initial model based on local water quality parameter data of a corresponding water plant to continuously improve the model accuracy.
Further, considering the self condition of each water plant client, hardware computing power, data volume difference and the like may exist, and the accuracy of the latest model parameters generated by local training is affected. Based on this, after each water plant client 1 completes the post-chlorination model based on the local water quality parameter data training, the updated latest model parameters corresponding to the post-chlorination model are uploaded to the cloud server 2, the cloud server 2 aggregates and receives the latest model parameters of each water plant client 1 and averages the latest model parameters, and then the latest model parameters are sent to all the water plant clients 1 to be used, so that the latest model parameters of each water plant client 1 are balanced, and the overall model accuracy of each water plant client 1 is improved. The requirements that individual water plants lack historical experience data or insufficient historical data are effectively met. Meanwhile, interaction data between each water plant client 1 and the cloud server 2 are model parameters, and water quality parameter data for local training are not contained, so that the confidentiality requirement of sensitive data of the water plant is effectively met.
In a preferred embodiment of the present invention, the cloud server 2 includes:
the delay module 21 is configured to, in each round of receiving the local model parameter, output an end signal indicating that the local model parameter is completely received in the round when the local model parameter is not received again within a preset time period in which the local model parameter is received for the last time;
and the calculating module 22 is connected with the delay module 21, and is configured to average the local model parameters received in this round according to the ending signal to obtain the latest model parameter, and send the latest model parameter to all the water plant clients 1.
Specifically, in this embodiment, in consideration of influences of other factors such as unstable network environment, one or more water plant clients 1 may not upload the latest model parameters of their own in time, so that all the water plant clients 1 may not update the model parameters in time, when receiving the local model parameters in each round, the cloud server 2 does not need to wait for receiving the local model parameters of all the water plant clients 1, and in the preset time period when receiving the local model parameters at the last time, if the local model parameters are not received again within 500ms, the average value calculation is performed, and even if the local model parameters are received again, the average value calculation is not performed again until the next round of receiving the local model parameters is started.
In a preferred embodiment of the present invention, a model network is pre-configured in the cloud server 2; the cloud server 2 further includes a configuration module 23, configured to configure the latest model parameters and the model network formed adding chlorine model in the newly accessed water plant client 1 when detecting that the new water plant client 1 is accessed.
In a preferred embodiment of the invention, the first adding point is arranged in each sedimentation tank of the water plant, the front chlorine adding amount respectively comprises the adding amount of sodium hypochlorite in each sedimentation tank, and the front chlorine adding allowance respectively comprises residual chlorine after the sodium hypochlorite is added in each sedimentation tank.
In a preferred embodiment of the present invention, the water quality parameter set further includes turbidity of raw water, turbidity of each contact tank of the water plant, and PH of each contact tank.
In a preferred embodiment of the present invention, the second adding point is disposed in each contact tank, the post-chlorine adding amount respectively includes the adding amount of sodium hypochlorite in each contact tank, and the post-chlorine adding residual amount respectively includes residual chlorine after sodium hypochlorite is added in each contact tank.
In a preferred embodiment of the present invention, the post-chlorination model is a deep neural network model.
As a preferred embodiment, before the intelligent chlorination control, the method further comprises a parameter screening process:
obtaining water quality measuring and calculating indexes of water inflow, water inflow PH, raw water temperature, raw water turbidity, sodium hypochlorite pre-adding amount, sodium hypochlorite pre-adding residual chlorine, contact pool PH, sodium hypochlorite post-adding amount, sodium hypochlorite post-adding residual chlorine, sodium hypochlorite supplementing amount, sodium hypochlorite supplementing residual chlorine, water outflow, water PH, effluent dissolved oxygen, water outflow pressure, water outflow turbidity, water outflow residual chlorine and the like of any water plant. These water quality data are time intervals of 5 minutes, with a total sampling span of about one year.
And then, carrying out data cleaning on the obtained data, wherein the general principle of the data cleaning is to keep water quality mutation data as much as possible and remove abnormal data which are not in accordance with the conventional law. The water quality mutation data is retained because water sources of a water plant may change greatly due to various reasons such as seasons, weather and the like. For example, the ammonia nitrogen content of raw water is overproof occasionally in a certain water plant, for example, rainstorm or water quality mutation caused by debris flow may occur in the geographical position of the certain water plant, for example, the algae content is overproof in the certain water plant, and the like. The above situations are possibly encountered in the daily operation process of the water plant, and may be shown in that the water quality data is greatly different from the ordinary data, but the data is precious burst data and is kept as much as possible.
For another part of the abnormal data we use the Median Absolute Difference (MAD) to cull the abnormal data. The abnormal data mainly refers to data which cannot be reached under the conditions of improper water quality, normal water quality or sudden water quality change. For example, the turbidity of raw water is data describing the turbidity degree of raw water, and its value must not be negative. The PH value of the nature is usually between 5 and 10, peracid or over-alkali data outside the range are removed, the raw water temperature is usually between 0 and 35 ℃, the residual chlorine value is usually positive, and data which do not meet the indexes need to be removed. And removing initial unstable stage data, wherein the data include but are not limited to water inflow rate of 2000-4000 liters per hour, and removing data of flow rate lower than 200 when a valve is just opened.
The basic principle of the above-described MAD method for anomaly detection is as follows. Assuming the data obeyed a normal distribution, let the outliers fall in the 50% area on both sides, and let the normal values fall in the middle 50% region:
wherein,
since, phi (-a) = 1-phi (a)
Then
The table look-up can show that,
therefore, the value of σ estimated from the MAD is
Namely MAD 1.4826.
Based on this, the values greater than MAD 1.4826 or less than MAD 1.4826 may be all within 50% of the area on both sides of the distribution, and determined as abnormal data, and the elimination process may be performed.
Furthermore, considering that the raw water quality parameters comprise hundreds of different indexes for measurement, more than ten water quality parameters which have great influence on sodium hypochlorite addition can be selected by using a characteristic screening method in combination with operator experience, and the sodium hypochlorite addition amount can be accurately predicted in real time. In the feature selection method, a Minimum Redundancy Maximum correlation method (mRMR) may be selected.
The method of the mRMR takes the correlation and redundancy among a plurality of features into consideration when performing feature selection, and particularly punishment is performed on redundant features which have higher correlation with the selected features. mRMR may use various measures of correlation, such as mutual information, correlation coefficients, and other distance or similarity scores. The mRMR feature selection method based on the mutual information approach is preferably used. The correlation between the feature set S and the target variable c can be defined as all the individual feature variables f in the feature set i And mutual information value I (f) of target variable c i (ii) a c) The average value of (a) is specifically as follows:
the redundancy of all the features in S is the mutual information I (f) between all the feature variables i ；f j ) The average value of (a) is specifically as follows:
the mRMR criteria are:
and solving the optimization problem to obtain a feature subset, wherein the feature subset comprises parameters used in the first proportional-integral-derivative control model, the post-chlorination model and the second proportional-integral-derivative control model.
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.
Claims (9)
1. An intelligent sodium hypochlorite adding system for water plants is characterized in that each water plant is provided with at least one first adding point arranged at a raw water inlet of the water plant, at least one second adding point arranged in front of a clean water tank of the water plant and at least one third adding point arranged behind the clean water tank;
then water plant sodium hypochlorite intelligence dosing system includes an at least water plant customer end, the water plant customer end with the water plant one-to-one, the water plant customer end includes:
the front chlorine adding control module is used for intelligently controlling a front chlorine adding amount of the sodium hypochlorite of the corresponding water plant at each first adding point according to a preset first proportional-integral-derivative control model and collecting front chlorine adding residual chlorine after the sodium hypochlorite is added at each first adding point;
the post-chlorination control module is connected with the pre-chlorination control module and used for inputting a water quality parameter set containing each pre-chlorination amount and the corresponding pre-chlorination residual chlorine into a post-chlorination model which is updated in a self-adaptive mode to obtain a post-chlorination amount, intelligently controlling the corresponding water plant to add the sodium hypochlorite with the post-chlorination amount at each second adding point, and collecting the post-chlorination residual chlorine after the sodium hypochlorite is added at each second adding point;
and the chlorine supplementing control module is respectively connected with the front chlorine adding control module and the rear chlorine adding control module and is used for intelligently controlling the chlorine supplementing quantity of the sodium hypochlorite at each third feeding point of the water plant according to a preset second proportional-integral-derivative control model, and the second proportional-integral-derivative control model takes the front chlorine adding quantity, the front chlorine adding residual quantity, the rear chlorine adding quantity and the rear chlorine adding residual quantity as feedforward parameters.
2. The intelligent sodium hypochlorite dosing system for the water plant as claimed in claim 1, further comprising a cloud server connected to each water plant client, wherein the water plant client further comprises a model updating module connected to the post-chlorination control module, and the model updating module comprises:
the local training unit is used for continuously acquiring a plurality of groups of the front chlorination amount, the front chlorination allowance and the corresponding rear chlorination amount to form a training set, and training and updating the rear chlorination model according to the training set to obtain corresponding local model parameters;
the parameter uploading unit is connected with the local training unit and is used for uploading the local model parameters obtained by each training to the cloud server;
the cloud server is used for averaging the local model parameters of the water plants to obtain a latest model parameter when the local model parameters of the water plants are received in each round, and sending the latest model parameter to all the model updating clients;
the model update module further comprises:
the updating unit is used for updating the post-chlorination model according to the latest model parameters when receiving the latest model parameters;
and the post-chlorination control module predicts and obtains the post-chlorination amount according to the updated post-chlorination model.
3. The intelligent sodium hypochlorite dosing system for water plants as claimed in claim 2, wherein the cloud server comprises:
the delay module is used for outputting an end signal representing that the local model parameters in the current round are completely received when the local model parameters are not received again within a preset time period after the local model parameters are received for the last time in the process of receiving the local model parameters in each round;
and the computing module is connected with the delay module and used for averaging all the local model parameters received in the current round according to the ending signal to obtain the latest model parameters and sending the latest model parameters to all the water plant clients.
4. The intelligent sodium hypochlorite dosing system for the water plant as claimed in claim 3, wherein a model network is pre-configured in the cloud server; the cloud server further comprises a configuration module, and the configuration module is used for configuring the latest model parameters and the model network into the newly accessed water plant client to form the post-chlorination model when detecting that the new water plant client is accessed.
5. The intelligent sodium hypochlorite dosing system for the water plant as claimed in claim 1, wherein the feed-forward parameters of the first proportional integral differential control model comprise a total instantaneous water inlet flow of the water plant, instantaneous water inlet flows of sedimentation tanks of the water plant, a raw water pH value, a raw water temperature, a raw water turbidity, a raw water dissolved oxygen content and a raw water chemical oxygen demand, and the feed-back parameters of the first proportional integral differential control model are the front chlorine adding allowance.
6. The intelligent sodium hypochlorite adding system for the water plant as claimed in claim 5, wherein the first adding points are arranged in the sedimentation tanks of the water plant, the pre-chlorine adding amount correspondingly and respectively comprises the adding amount of the sodium hypochlorite in the sedimentation tanks, and the pre-chlorine adding allowance correspondingly and respectively comprises residual chlorine after the sodium hypochlorite is added in the sedimentation tanks.
7. The intelligent sodium hypochlorite dosing system for the water plant as claimed in claim 1, wherein the set of water quality parameters further comprises turbidity of the raw water, turbidity of each contact tank of the water plant and PH of each contact tank.
8. The intelligent sodium hypochlorite adding system for the water plant as claimed in claim 7, wherein the second adding point is arranged in each contact tank, the post-chlorination amount correspondingly and respectively comprises the adding amount of the sodium hypochlorite in each contact tank, and the post-chlorination surplus correspondingly and respectively comprises residual chlorine after the sodium hypochlorite is added in each contact tank.
9. The intelligent sodium hypochlorite dosing system for the water plant as claimed in claim 1, wherein the post-chlorination model is a deep neural network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210847866.0A CN115196730B (en) | 2022-07-19 | 2022-07-19 | Intelligent sodium hypochlorite adding system for water plant |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210847866.0A CN115196730B (en) | 2022-07-19 | 2022-07-19 | Intelligent sodium hypochlorite adding system for water plant |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115196730A true CN115196730A (en) | 2022-10-18 |
| CN115196730B CN115196730B (en) | 2024-06-14 |
Family
ID=83581746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210847866.0A Active CN115196730B (en) | 2022-07-19 | 2022-07-19 | Intelligent sodium hypochlorite adding system for water plant |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115196730B (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07132293A (en) * | 1993-05-10 | 1995-05-23 | Toshiba Corp | Chlorine injection controller for filtration plant |
| JP2005329358A (en) * | 2004-05-21 | 2005-12-02 | Ffc Ltd | System for controlling injection of chlorine into water purification plant |
| KR20100012535A (en) * | 2008-07-29 | 2010-02-08 | 주식회사 한힘테크놀러지 | Chlorine automatic system |
| CN201914961U (en) * | 2010-12-27 | 2011-08-03 | 天津创业环保集团股份有限公司 | On-demand chlorine adding control system |
| CN110442457A (en) * | 2019-08-12 | 2019-11-12 | 北京大学深圳研究生院 | Model training method, device and server based on federation's study |
| CN111807558A (en) * | 2020-07-23 | 2020-10-23 | 上海城市水资源开发利用国家工程中心有限公司 | Multi-point chlorination process for drinking water |
| CN112759118A (en) * | 2020-12-21 | 2021-05-07 | 上海熊猫机械(集团)有限公司 | Water plant dosing integrated system and dosing method |
| KR102290331B1 (en) * | 2021-04-09 | 2021-08-19 | 이성갑 | IOT chlorine feeder management system reflecting water quality analysis data |
| CN113537513A (en) * | 2021-07-15 | 2021-10-22 | 青岛海尔工业智能研究院有限公司 | Model training method, device, system, equipment and medium based on federal learning |
| CN113651402A (en) * | 2021-08-31 | 2021-11-16 | 哈尔滨天达控制股份有限公司 | Water treatment chlorination control method and device |
| CN114064623A (en) * | 2021-11-01 | 2022-02-18 | 上海威派格智慧水务股份有限公司 | Generation system and method of intelligent adding model |
| WO2022121804A1 (en) * | 2020-12-10 | 2022-06-16 | 华为技术有限公司 | Method for semi-asynchronous federated learning and communication apparatus |
-
2022
- 2022-07-19 CN CN202210847866.0A patent/CN115196730B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07132293A (en) * | 1993-05-10 | 1995-05-23 | Toshiba Corp | Chlorine injection controller for filtration plant |
| JP2005329358A (en) * | 2004-05-21 | 2005-12-02 | Ffc Ltd | System for controlling injection of chlorine into water purification plant |
| KR20100012535A (en) * | 2008-07-29 | 2010-02-08 | 주식회사 한힘테크놀러지 | Chlorine automatic system |
| CN201914961U (en) * | 2010-12-27 | 2011-08-03 | 天津创业环保集团股份有限公司 | On-demand chlorine adding control system |
| CN110442457A (en) * | 2019-08-12 | 2019-11-12 | 北京大学深圳研究生院 | Model training method, device and server based on federation's study |
| CN111807558A (en) * | 2020-07-23 | 2020-10-23 | 上海城市水资源开发利用国家工程中心有限公司 | Multi-point chlorination process for drinking water |
| WO2022121804A1 (en) * | 2020-12-10 | 2022-06-16 | 华为技术有限公司 | Method for semi-asynchronous federated learning and communication apparatus |
| CN112759118A (en) * | 2020-12-21 | 2021-05-07 | 上海熊猫机械(集团)有限公司 | Water plant dosing integrated system and dosing method |
| KR102290331B1 (en) * | 2021-04-09 | 2021-08-19 | 이성갑 | IOT chlorine feeder management system reflecting water quality analysis data |
| CN113537513A (en) * | 2021-07-15 | 2021-10-22 | 青岛海尔工业智能研究院有限公司 | Model training method, device, system, equipment and medium based on federal learning |
| CN113651402A (en) * | 2021-08-31 | 2021-11-16 | 哈尔滨天达控制股份有限公司 | Water treatment chlorination control method and device |
| CN114064623A (en) * | 2021-11-01 | 2022-02-18 | 上海威派格智慧水务股份有限公司 | Generation system and method of intelligent adding model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115196730B (en) | 2024-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110101106B (en) | Moisture control method and system for dampening and humidifying process based on fuzzy feedforward feedback algorithm | |
| JP6609808B2 (en) | Prediction program using decision tree learning algorithm, apparatus and method | |
| US9895630B2 (en) | Allocation measurement systems and methods | |
| CN110054274B (en) | Water purification flocculation precipitation dosing control method | |
| US20130220902A1 (en) | Plc system for automatically controlling pid for maintaining target water quality value by depositing water treatment chemical | |
| CN107728477A (en) | A kind of industrialized aquiculture water quality dissolved oxygen prediction control method and system | |
| CN106802563B (en) | A kind of sewage procedure optimization control method based on drosophila optimization and LSSVM | |
| CN111320246B (en) | Coagulant intelligent accurate adding control system based on multivariable control | |
| CN112419095A (en) | Accurate dosing method based on historical data and real-time data feedback | |
| US20230257283A1 (en) | Process, system, and computer readable storage medium for determining optimal coagulant dosage | |
| CN108142976B (en) | Cut tobacco drying process parameter optimization method | |
| CN112132700A (en) | Formula decision method for regulating and controlling water quality | |
| CN115196730A (en) | Intelligent sodium hypochlorite adding system for water plant | |
| CN111777138A (en) | Sewage treatment refined control system and control method based on Internet of things | |
| JP2001214470A (en) | Controlling equipment for most adequate water transportation | |
| TWI813437B (en) | Water treatment control system and control method of water treatment device | |
| EP4407481A2 (en) | Novel systems and methods for monitoring and management of a fluid infrastructure network | |
| CN207699239U (en) | A kind of Powdered Activated Carbon Automatic Dosage Control of Additives system for water process | |
| CN115108595A (en) | Intelligent water purifying agent adding method based on machine learning | |
| CN115504557A (en) | Automatic sludge discharge method applied to sedimentation tank | |
| CN115375009A (en) | Method for establishing coagulation intelligent monitoring linkage system | |
| JP2005329358A (en) | System for controlling injection of chlorine into water purification plant | |
| CN110671340B (en) | Flow self-adaptive adjustment method of single-stage full-adjustment pump station | |
| WO2021243458A1 (en) | Apparatus and methods for cleaning reverse osmosis systems | |
| CN105425581B (en) | Steam power plant's steam pipeline oxygenation control method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |