CN115809749B - Method for establishing comprehensive online prediction model of sewage treatment and prediction and early warning method - Google Patents
Method for establishing comprehensive online prediction model of sewage treatment and prediction and early warning method Download PDFInfo
- Publication number
- CN115809749B CN115809749B CN202310086581.4A CN202310086581A CN115809749B CN 115809749 B CN115809749 B CN 115809749B CN 202310086581 A CN202310086581 A CN 202310086581A CN 115809749 B CN115809749 B CN 115809749B
- Authority
- CN
- China
- Prior art keywords
- data
- sewage treatment
- prediction model
- online
- calculation
- 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.)
- Active
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A20/00—Water conservation; Efficient water supply; Efficient water use
- Y02A20/152—Water filtration
Abstract
The invention discloses a method for establishing a comprehensive online prediction model for sewage treatment and a prediction early warning method, firstly, in the modeling process, an offline model is established and calibrated by utilizing historical data and supplementary experimental data of sewage treatment facilities; the offline model is communicated with the sewage treatment informationized platform, so that the offline model is an online prediction model connected with the sewage treatment informationized platform, wherein the sewage treatment informationized platform acquires first-class data including operation parameters of sewage treatment facilities and water quality of sewage inflow water; the online prediction model is used for predicting and obtaining second class data containing the effluent quality, comparing the second class data with third class data containing the actual effluent quality, correcting the online prediction model and verifying the model accuracy, and can be used for predicting and acting the potential exceeding risk of the sewage treatment facility, so that the prediction accuracy and the early warning effectiveness are greatly improved.
Description
Technical Field
The invention relates to the field of sewage treatment, in particular to a method for establishing a comprehensive online prediction model for sewage treatment and a prediction and early warning method.
Background
The simulation and prediction of sewage biochemical treatment process by using computer modeling is a popular direction of research in the field of sewage treatment in recent years.
The invention patent application with publication number of CN112397137A discloses a prediction model and a prediction method for the concentration change rule of organic micro-pollutants in sewage, wherein the prediction model is an off-line model of water-assisted Activated Sludge (ASM) -OMPs (open-air sludge systems) based on an ASM (open-air sludge systems) model, and is mainly used for predicting the change of the concentration of the organic micro-pollutants (OMPs) in the sewage, inputting the measured concentration of the organic micro-pollutants in the sewage into the off-line model, and describing the growth and the microorganism by adopting Monod dynamics
The metabolism of the organic micro-pollutants is realized, so that the simulation calculation and prediction of the concentration of the organic micro-pollutants are realized, the dynamics and the stoichiometric parameters are only used as influencing factors for the prediction of the concentration of the organic micro-pollutants, and the typical default values of the dynamics and the stoichiometric parameters proposed by the international water cooperation are only input into the off-line model, so that the concentration of the organic micro-pollutants in sewage can be only predicted, and the method has certain limitation.
In addition, the patent application of the invention with publication number of CN112415911A discloses a parameter automatic calibration method based on sensitivity analysis and differential evolution algorithm, which has the basic thought that a group of off-line models based on an international water cooperation Activated Sludge Model (ASM) are established, water quality parameters and operation parameters of sewage treatment facilities are simulated and calculated through the off-line models, related parameters obtained through the simulation and calculation are used as future predictions of the sewage treatment facilities and sewage quality, and compared with actual measurement values, parameter correction is continuously carried out after the simulation deviation is found, so that the accuracy of the off-line model prediction is improved.
In the two groups of patent documents, an off-line model is established based on an ASM (anaerobic sludge management) active sludge model, the past operation working conditions of the sewage treatment facility are simulated and reproduced on the off-line model by combining the operation parameters of the sewage treatment facility and the historical data of the sewage quality, and various data such as the operation parameters of the sewage facility in an active sludge treatment section, the component concentration in the inlet water, the dynamic parameters and the like cannot be fully considered and utilized, and the accuracy of the off-line model in future operation of the sewage treatment facility and the prediction of the future sewage quality working conditions is directly influenced by the application of the various data.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a method for establishing a comprehensive online prediction model for sewage treatment and a prediction early warning method, which are used for establishing the model for sewage treatment and based on the model, communicating the model with sewage treatment informationized platforms in the existing sewage treatment facilities (including sewage treatment plants and other facilities related to sewage treatment), and carrying out quantitative analysis according to the model and different types of data so as to assist technicians and managers of the sewage treatment facilities to better guide the actual operation of the sewage treatment.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
on one hand, the invention provides a method for establishing a comprehensive online prediction model for sewage treatment, which comprises the following steps:
s1, establishing an accurate offline model of the sewage treatment facility and in the sewage treatment process by utilizing the operation parameters of the sewage treatment facility, the historical data of the sewage quality and the data collected by the supplementary experiment;
s2, communicating data between the offline model and the sewage treatment informatization platform to enable the offline model to be an online prediction model connected with the sewage treatment informatization platform, wherein the sewage treatment informatization platform acquires first type data including operation parameters of sewage treatment facilities and water quality of sewage inflow water;
s3, the online prediction model simulates and calculates the first type of data, generates second type of data comprising calculation results and prediction information, and shares and displays the second type of data to the sewage treatment informatization platform.
In the step S2, defining that the first type of data includes first machine inspection data collected by the sewage treatment facility on line and first human inspection data collected and detected manually and daily;
the first machine inspection data comprise data which are automatically acquired by various instruments and meters in a sewage treatment facility, sensors, monitoring facilities and other equipment and instruments communicated with a sewage treatment informatization platform, and comprise water quality data and flow data of COD (chemical oxygen demand) at a sewage water inlet position, NH3-N, TP, TN, TSS and the like, and data of temperature, dissolved oxygen concentration and the like of a biochemical reaction tank, wherein the COD is chemical oxygen demand and is used for defining the concentration of organic pollutants in a water sample; NH3-N refers to ammonia nitrogen concentration in a water sample; TN refers to the total nitrogen concentration in a water sample; TP refers to the total phosphorus concentration in a water sample; PO4-Ps refer to orthophosphates in a water sample; TSS refers to the total solids suspension concentration in a water sample.
The first human detection data are manually detected and recorded data, and comprise data obtained by manually reading data of various detection instruments, sample experiments at a sewage inlet position, manually read data of operation parameters of sewage treatment facilities and the like, wherein the data comprise indexes such as inlet water quality indexes COD, TN, NH-N, TP, PO4-P, TSS and the like, and operation parameter data such as sludge reflux flow, mixed liquid reflux flow, sludge discharge and the like which are adjusted and manually recorded according to daily operation requirements; the first personnel inspection data are uploaded to the sewage treatment informatization platform through manual work.
Therefore, the first machine inspection data and the first person inspection data have more repeated inspection items, can be used for rechecking the data, and are also convenient for timely finding out the abnormality of the inspection item indexes.
In addition, in step S2, the sewage treatment informatization platform further obtains third class data including COD, TN, NH3-N, TP, TSS, etc. of operation parameters of the sewage treatment facility and effluent quality, wherein the third class data includes third machine inspection data collected by the sewage treatment facility on line and third human inspection data collected and detected by manual daily collection according to different data sources.
The third type of data comprises information such as MLSS, MLVSS and the like and information such as COD, TN, NH-N, TP, TSS and the like of effluent, wherein MLSS refers to the total suspended solid particle concentration of the mixed liquor, and MLVSS refers to the volatile suspended solid particle concentration of the mixed liquor; and uploading the third human detection data to the informationized platform in a report form.
On the other hand, the invention also provides a prediction and early-warning method adopting the comprehensive online prediction model for sewage treatment, which comprises the following sub-steps:
s3-0, steady state calculation: simulating and calculating the mean value of the first person detection data in the past days by using an online prediction model to perform steady-state calculation;
s3-1, dynamic calculation: and using an online prediction model to simulate and calculate first machine inspection data in the past hours, and simulating and calculating and outputting second class data in the future hours.
Specifically, further, the step S3-0 includes:
s3-0-0, comparing the second class data output by steady state calculation with the average value of third human detection data in the past days, and judging whether the second class data is consistent with the average value of the third human detection data;
s3-0-1, if the online prediction model is effective, performing the next step of dynamic calculation;
if not, the online model is calibrated for the first time: correcting the parameter input value of the online prediction model according to the average value of the first human detection data in the past several days so as to enable the comparison result to be consistent;
furthermore, in some preferred embodiments, the step S3-1 is followed by:
s3-2, carrying out dynamic calculation simulation calculation in the step S3-1, outputting an output value of the second class data in a plurality of hours in the future and third machine-check data in a plurality of hours in the future, judging whether the output value is consistent with the third machine-check data in a plurality of hours in the future, and verifying the validity of the online prediction model again.
Preferably, when the third machine inspection data of the next several hours in the step S3-2 and the output value of the second class data in the next several hours are simulated, calculated and output;
if yes, the online prediction model is effective;
if not, the following steps are carried out: s3-2-0, comparing the first machine inspection data with the first person inspection data, and verifying the validity of the first machine inspection data;
further, in some preferred embodiments, in step S3-2-0, if the first machine check data and the first human check data do not match, maintaining the sewage treatment facility that collected the first machine check data based on the first human check data;
if the first machine inspection data and the first person inspection data are consistent, and the second type data which are calculated and output by the online prediction model through dynamic calculation in the past days are inconsistent with the first type data, the parameter correction of the online prediction model is required to be carried out by a professional.
In the implementation step, when the steady state calculation is performed, the steady state calculation is performed once a day in the online prediction model, and in step S3-0, the steady state calculation is performed once a day, and in each step, the online prediction model is used to simulate the first human detection data within the past 30 days of the current day, particularly considering that the sludge retention time of the healthy operation of most sewage treatment facilities is 15 days to 20 days, and in step S3-0, it is reasonable to simulate the first human detection data within the past 30 days of the current day.
Further, in the daily steady-state calculation and the dynamic calculation in the above steps S3-0 to S3-1, the result of the simulation calculation of the daily steady-state calculation is taken as the initial value of the first dynamic calculation on the same day, and the result of the last dynamic calculation for each remaining dynamic calculation is taken as the initial value.
Compared with the prior art, the invention has the following beneficial effects:
(1) In the prediction model and the prediction early warning method, firstly, the operation parameters of sewage treatment facilities and the historical data of sewage quality are adopted for modeling, and the data intercommunication is realized between the offline model and the sewage treatment informatization platform in the existing sewage treatment plant so as to form an online prediction model, so that the timeliness and convenience of the online prediction model for acquiring the first kind of data are greatly improved, and meanwhile, the simulation calculation result and the prediction information can be effectively displayed and early warned on the sewage treatment informatization platform in time.
(2) In addition, in order to ensure the effectiveness of the online model, the method comprises the steps of establishing and calibrating an offline model, performing online operation steady-state calculation and performing online operation dynamic calculation, and defining and distinguishing the names and actions of various data required by model calibration, dividing first data acquired by a sewage treatment informatization platform into first machine detection data and first person detection data, clearly dividing the purposes of the various data in each step of establishing and operating the model, and compared with the machine detection data, the method fully utilizes the accuracy of the human detection data as a data source for performing steady-state calculation and model calibration by an online prediction model, so that the online prediction model can be more fit with the actual conditions of sewage treatment facilities and sewage quality, and greatly improves the prediction accuracy of the online prediction model;
(3) In addition, in the step of predicting the early warning scheme, a plurality of substeps are arranged to realize the comparison between the machine inspection data and the human inspection data, and the effectiveness of the input value calculated by the online prediction model is determined; and setting sub-steps to compare the predicted value of the online model with the measured value of the third type of data, and determining the effectiveness of the online prediction model.
Drawings
FIG. 1 is a flow chart of a method for establishing a comprehensive online prediction model for sewage treatment in one embodiment of the invention;
FIG. 2 is a block diagram of a predictive early warning method of a comprehensive online prediction model for sewage treatment according to an embodiment of the present invention;
fig. 3 is a second flow chart of a predictive early warning method of a comprehensive online prediction model for sewage treatment according to an embodiment of the invention.
Detailed Description
The invention is further illustrated by the following examples in conjunction with the accompanying drawings:
the specific embodiment of the invention is divided into two parts of a method for establishing the comprehensive online sewage treatment prediction model and a method for predicting the comprehensive online sewage treatment prediction model.
Embodiment one:
referring to fig. 1, the embodiment discloses a method for establishing a comprehensive online prediction model for sewage treatment, which comprises the following steps:
s1, establishing an accurate offline model of the sewage treatment facility and in the sewage treatment process by utilizing the operation parameters of the sewage treatment facility and the historical data of the sewage quality and the supplementary experimental data; the offline model is a mechanism type mathematical model describing biochemical reaction, in this step, the built offline model can be based on activated sludge models such as ASM1, ASM2d, ASM3 and the like released by international water cooperation, or based on a model secondarily developed by the activated sludge models, or other commercial models or models independently developed in the market, in short, the offline model in this step can be built on various existing activated sludge models;
in the step, for the historical data of the operation parameters of the sewage treatment facilities and the sewage quality, the historical data including the operation parameters of the sewage treatment facilities, the water quality of inflow water, the operation state, the water quality of outflow water and the like in the past year can be collected; in addition, for the supplementary experiment data, when offline modeling is performed, a supplementary experiment plan for two weeks can be executed, and the supplementary experiment data is utilized to determine the water quality component of the sewage inflow water;
in addition, it is easy to understand that the off-line model is used for simulating and reproducing the historical conditions of the sewage treatment facilities and the sewage quality, and does not have a direct prediction function for future conditions, so that after the off-line model is built, the off-line model is required to be calibrated by means of the real data of the sewage treatment facilities and the sewage quality, so that the off-line model can simulate and reproduce the real running conditions of the sewage treatment facilities and the real sewage quality conditions;
s2, communicating data between the offline model and the sewage treatment informatization platform to enable the offline model to be an online prediction model connected with the sewage treatment informatization platform; in the step, the sewage treatment informatization platform refers to a system platform existing in the existing sewage treatment facility, such as a sewage treatment real-time monitoring system platform in a sewage treatment plant and various intelligent water service informatization system platforms, and has perfect sewage treatment facility operation parameter condition monitoring and water quality condition monitoring functions;
the calibrated offline model is directly embedded into the sewage treatment informatization platform in an interface and wireless receiving and transmitting mode, so that data intercommunication between the offline model and the sewage treatment informatization platform can be realized, and further, online deployment of the offline model is realized.
In addition, since the acquired data of the existing sewage treatment informatization platform contains the state parameters of the sewage treatment facilities and the sewage quality at the sewage inlet/outlet, the following definitions are made for the data in order to distinguish the use occasions and the uses of the data information:
the first type of data comprises operation parameters of the sewage treatment facility and sewage quality information sampled at the sewage inlet, and is divided into first machine inspection data which are collected by the sewage treatment facility on line and first human inspection data which are collected and detected by manual daily according to different sampling modes.
And the second type of data is an output value obtained after analog calculation by taking the first type of data as an input value for the online prediction model.
And the third type of data comprises sewage quality information sampled at a sewage outlet, wherein the third type of data is divided into third machine inspection data which are acquired on line by sewage treatment facilities and third human inspection data which are acquired and detected by manual daily utilization according to different sampling modes.
S3, simulating and calculating the first type of data by using an online prediction model, generating second type of data comprising calculation results and prediction information, and displaying the second type of data through a sewage treatment informatization platform; in the step, the online prediction model acquires first-class data, performs simulation calculation, and sends the generated calculation result and prediction information to the sewage treatment informatization platform, so that the function establishment of the online prediction model is completed.
In the steps S1 to S3, the first machine inspection data and the third machine inspection data include data automatically obtained on line by equipment and instruments, such as various instruments, sensors, monitoring facilities and the like in the sewage treatment facility, which are communicated with the sewage treatment informatization platform, such as obtaining water quality data, flow and dissolved oxygen concentration data, such as COD (chemical oxygen demand), ammonia nitrogen, total phosphorus and the like of the inlet water and the outlet water, specifically:
the first machine inspection data comprise water quality data and flow data such as COD (chemical oxygen demand), NH3-N, TP and the like of the sewage water inlet part, and data such as the temperature and the dissolved oxygen concentration of the biochemical reaction tank.
COD is chemical oxygen demand and is used for defining the concentration of organic pollutants in a water sample;
TP refers to the total phosphorus concentration in a water sample;
TN refers to the total nitrogen concentration in a water sample;
TSS refers to the total solids suspension concentration in a water sample.
In addition, the third machine inspection data comprise COD, TN, NH-N, TP, TSS and other data at the outlet water, and the individual sewage plants also comprise NO3-N and NH3-N measured at Cheng Mou of the biochemical reaction tank edge;
NH3-N refers to ammonia nitrogen concentration in a water sample;
NO3-N refers to the nitrate nitrogen (i.e., nitrate nitrogen) concentration in the water sample.
In the steps S1 to S3, the first human detection data and the third human detection data are manually detected and recorded data, including manually read data of various detection instruments, manually read data obtained by a sewage sample experiment, manually read operation parameter data of sewage treatment facilities, specifically including inflow water quality indexes COD, TN, NH-N, TP, PO4-P, TSS and the like, and MLSS and MLVSS in a biochemical reaction tank, wherein MLSS refers to total suspended solid particle concentration of mixed liquid; MLVSS refers to the mixed liquor volatile suspended solids concentration. And the system comprises operation parameter data information such as sludge reflux flow, mixed liquor reflux flow, sludge discharge amount and the like which are adjusted and manually recorded according to daily operation requirements. The first human detection data and the third human detection data are manually transmitted to the sewage treatment informatization platform.
In conclusion, both the sewage water inlet and the sewage water outlet are provided with two data forms of human detection data and machine detection data, so that the data can be checked conveniently, and the abnormality of the detection item index can be found conveniently in time.
In addition, because the first personnel inspection data and the third personnel inspection data are obtained by adopting a manual inspection experiment, generally, the water sample used for manual inspection is mostly a mixed water sample for 24 hours in the whole day, compared with the mode that the first personnel inspection data and the third personnel inspection data are obtained on line through instruments, the manual inspection has the advantages of more comprehensive sewage sampling sample range and more accurate water quality index, and has the timeliness of data sources like the personnel inspection data, so that the first personnel inspection data and the third personnel inspection data are more suitable for steady-state calculation.
Embodiment two:
referring to fig. 2 to 3, as a second class of parts of the present invention, the present invention provides a predictive early-warning method using the comprehensive online prediction model for sewage treatment according to the first embodiment, wherein the predictive early-warning method includes the following sub-steps:
s3-0, steady state calculation: simulating and calculating the mean value of the first person detection data in the past days by using an online prediction model to perform steady-state calculation;
s3-1, dynamic calculation: and using an online prediction model to simulate and calculate first machine inspection data in the past hours, and simulating and calculating and outputting second class data in the future hours.
Next, the steady state calculation section in step S3-0 will be described.
In step S3-0, the result of the steady-state calculation generally reflects the state and treatment effect of the sewage treatment facility and the sewage quality for a period of time, the purpose of the steady-state calculation is to anchor the average value of the existing states of the sewage treatment facility and the sewage quality by using relatively reliable water quality data, and the first human detection data uses relatively reliable manual detection (report) data, so that the output value obtained by performing the steady-state calculation by using the first human detection data is closer to the average state of the actual sewage treatment facility and the sewage quality than the data obtained by any other detection method, and is a more reasonable initial input value of the dynamic calculation, and therefore, the average value of the first human detection data for a period of time is necessary to be used as the input value of the steady-state calculation by using the online prediction model.
Currently, in the existing sewage treatment facilities, considering that the sludge retention time for healthy operation is about 15 days to 20 days, it is necessary to select a time period not less than the above sludge retention time as the time days for steady state calculation.
Thus, referring to FIG. 2, in one embodiment, the average of the first human data over the past 30 days is selected as the input value for the steady state calculation, which is performed once a day, and it is noted that the input values for the daily steady state calculation are each the average of the first human data over the past 30 days of the day of the steady state calculation.
In addition, referring to fig. 2, in order to make the work of the online prediction model more intelligent, so that a worker can intuitively obtain the current working state of the online prediction model, in one embodiment, a data checking step is further added in the prediction early warning method before the daily steady-state calculation is performed, and the completeness check can be performed on the first human detection data in the past 30 days obtained by the online prediction model; that is, for the average value of the first human detection data in the past 30 days required for the steady-state calculation, it is first checked whether or not each parameter and index data of the first human detection data in the past 30 days required for the past is sound, before the steady-state calculation, and there are cases where:
for the case of a complete missing data, an alert I (a) may be generated and executed: "first human check data over the last 30 days is completely missing, the system will use the result of the last steady state calculation last as the initial value for the next dynamic calculation. ", i.e., the system may use the last steady state calculation result in the past 30 days as the initial value of the dynamic calculation in step S3-1, e.g., the steady state calculation on the current day cannot acquire the first person detection data in the past 30 days, and may use the steady state calculation result on the previous day as the initial value of the dynamic calculation on the current day.
For the case of a partial absence of data, a warning I (b) may be generated and executed that "detect data incomplete, please upload the first human check data, the system will use the average of the uploaded first human check data over the last 30 days as the input value for steady state calculation. At this point, since only the first human check data for several days over the last 30 days is complete, e.g., only 15 or 20 days, the system will take the average of these 15 or 20 days as the input value for the steady state calculation.
Because the data intercommunication between the online prediction model and the sewage treatment informationized platform, the warning I (a) and the warning I (b) and other warning information described below can be displayed in the sewage treatment informationized platform, so that the staff can conveniently learn information and perform maintenance operation.
After the data checking step is passed, steady state calculation is performed.
Furthermore, referring to fig. 2, in one embodiment, after the steady state calculation is completed and before the dynamic calculation, a verification and correction step of the steady state calculation result is further introduced into the online prediction model, that is, the step S3-0 further includes:
s3-0-0, comparing the result of steady state calculation by using the average value of the first human detection data in the past 30 days (namely the second class data calculated by the online prediction model) with the average value of the third human detection data in the past 30 days, and judging whether the results are consistent; ( It should be noted that: the judging standard of whether the average value of all detection indexes in the two kinds of data is consistent is that whether the average value of all detection indexes in the two kinds of data is different by more than 20%, for example, whether the average value of the MLSS indexes predicted in the second kind of data is different from the MSLL indexes in the third person detection data by more than 20%, if the average value of all detection indexes exceeds 20%, the average value of all detection indexes in the two kinds of data is inconsistent. )
S3-0-1, if yes, the online prediction model is effective, and step S3-1 is performed, if not, the input parameters of the online prediction model are wrong, and parameter correction is performed on the online prediction model according to the average value of first human detection data within the past 30 days, so that the comparison results of the two are consistent.
Further, referring to fig. 2, in step S3-0-0, when the average value of the third human detection data in the past 30 days is introduced, it is determined whether the data is sound, and if not, a warning ii is generated and executed: "third person detection data within the past 30 days is not sound, verification of the current steady state calculation result cannot be performed, and the current steady state calculation result is directly used for dynamic calculation. And (2) when warning II is prompted, the online prediction model performs steady state calculation once every day, and the online prediction model directly defaults that the current steady state calculation result is effective and directly enters the next dynamic calculation step in consideration of the fact that the difference between two adjacent steady state calculations is smaller.
Next, the dynamic calculation section will be described.
After the steady state calculation is completed, the online prediction model already obtains the average value of the sewage treatment facility and the sewage quality in the existing state, and the average value can be used for predicting the future water outlet condition, so that in the dynamic calculation of the step 3-1, the average value of the first machine detection data in the past several hours is selected as the input value of the dynamic calculation, and the water outlet change in the unit of hours, which is caused by the future water outlet quantity, the water quality and the running condition fluctuation of the sewage, is reflected through the online prediction model calculation.
In addition, during dynamic calculation, the first machine-check data is derived from the sewage treatment facility to periodically sample on line and detect the water quality input value obtained by the on-line prediction model, so, since the frequency of the on-line sampling detection of the sewage treatment facility is limited by the detection method in the current industry, the longest detection step generally comprises sampling, digestion and analysis, and may be as long as 2 hours, and thus, in one embodiment, the average value of the first machine-check data extracted for the past 2 hours is selected.
After the average value of the first machine-check data of the past 2 hours is extracted, before dynamic calculation, whether the first machine-check data of the past 2 hours is sound or not is checked, and the dynamic detection step is carried out after the sound is confirmed, if the data is not sound, a warning IV is generated and executed: "the first machine check data of the past 2 hours is not sound, this dynamic calculation is terminated. ", after the dynamic calculation is terminated, the next first machine check data after 2 hours is waited.
In the dynamic calculation, the online prediction model calculates and predicts second class data in the future 2 hours according to the average value of the first machine-check data in the past 2 hours as an input value.
Therefore, after the dynamic calculation in step S3-1, step S3-2 is further provided, and the output value (average value of the second class data in 2 hours in the future) of the dynamic calculation is compared with the average value of the third machine inspection data obtained at the water outlet after 2 hours, specifically:
referring to fig. 3, which is a table diagram of the subsequent flow of the dynamic calculation in fig. 2, in step 3-2, the online prediction model performs simulation calculation according to the average value of the first machine-check data for the past 2 hours and outputs the second class data as an output value, and the output value is compared with the third machine-check data acquired by the next water outlet to determine whether the output value is consistent with the third machine-check data; it should be noted that, because the current environmental protection monitoring unit will set up structures such as a water outlet detection sensing probe at the sewage outlet, the acquired data is generally more accurate, so the source of the third machine detection data may be that the environmental protection monitoring unit will set up a water outlet detection sensing probe at the sewage outlet.
In the steps S3-0 to S3-2, the dynamic calculation predicts the water quality of the sewage outlet for a plurality of hours in the future, so that the prediction of the water quality of the sewage outlet for a long time in the future can be performed through continuous dynamic calculation for prediction and early warning; then, in addition to the warning program caused by the above-mentioned data soundness, under normal conditions, the output value of the steady-state calculation performed daily will be used as the initial value of the first dynamic calculation in the day, and the input value of each dynamic calculation remaining in the day adopts the output value of the last dynamic calculation, so that the continuous prediction of the running condition of the sewage treatment facility and the sewage quality condition in the future can be realized.
Referring to fig. 3, it is disclosed whether the output values of the third machine-check data (i.e., the third machine-check data obtained by the next water outlet in this embodiment) and the online prediction model for the next several hours in the future in the determining step S3-2 are matched;
if the online prediction model is valid, the online prediction model and the dynamic calculation are proved to be valid, and the next dynamic calculation can be directly executed;
if not, the following step S3-2-0 is carried out:
s3-2-0, comparing the first machine inspection data with the first person inspection data, and verifying the validity of the first machine inspection data; if the first machine inspection data and the first person inspection data are not consistent, maintaining a sewage treatment facility and related probes for collecting the first machine inspection data by taking the first person inspection data as a reference; if the first machine inspection data and the first person inspection data are consistent, and the second class data which is calculated and output by the online prediction model through dynamic calculation in the past days is inconsistent with the third machine inspection data, the dynamic calculation is invalid, and the next dynamic calculation is performed after the professional is required to correct the parameters of the online prediction model, specifically:
referring to FIG. 3, in one embodiment, in step S3-2-0, it is first determined whether the online prediction model can obtain the first human data of the previous day, and if not, a warning VI is generated: "dynamic calculation result error due to check is performed, first person check data of yesterday is missing, and current data comparison cannot be performed. ", and directly enter the next dynamic calculation.
In step S3-2-0, if the online prediction model can obtain sound first human detection data of the previous day, the first human detection data of the previous day is compared with the first machine detection data, and whether the difference between the two values exceeds 20%, for example, whether the difference between the contents of common indexes such as MSLL, COD, TN, TP and TDS is 20% is determined.
If the first human check data of the previous day differs from the first machine check data by more than 20%, generating and executing a warning V: the first human detection data and the first machine detection data have excessively large phase difference, so that the output value of dynamic calculation is inconsistent with the result of the third machine detection data, and the next round of dynamic calculation is directly carried out.
If the difference between the first person detection data and the first machine detection data in the previous day is not more than 20%, further judging whether 40% of dynamic calculation output values in a plurality of dynamic calculations in the past 5 days are consistent with the result of the third machine detection data, if so, directly entering the next dynamic calculation; if not, a warning VII is generated: the first human detection data is consistent with the first machine detection data, but the dynamic calculation result is inconsistent with the third machine detection data for many days, and the online prediction model needs to be maintained. And after maintenance is completed, entering the next dynamic calculation.
In the whole logic flow of the online prediction model, the method has the steps of carrying out model validity test for a plurality of times, judging the validity of the online prediction model, timely sending out a model maintenance prompt, preventing from entering a technical error zone for once and for all after online deployment of the model, giving eight different warning contents according to different conditions, and effectively prompting the model state and the data state; on the premise that the data are effective and the model is effective, the dynamic operation effect in the sewage treatment facility is predicted and early-warned, technical management personnel of the sewage treatment facility are helped to judge the excessive risk and the actual operation state, and preventive measures can be taken before the problem occurs.
In addition, in the logic flow, the judgment of soundness of manually detecting the water inflow data, the operation parameter data of the sewage treatment facilities, the on-line monitoring data and the like is related to the situation that the validity of the data is not checked after the data is collected is prevented, and truly effective data resources are accumulated for the sewage treatment process.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the scope of the present invention. It will be apparent that the described embodiments are merely some, but not all, embodiments of the invention. Based on these embodiments, all other embodiments that may be obtained by one of ordinary skill in the art without inventive effort are within the scope of the invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art may still combine, add or delete features of the embodiments of the present invention or make other adjustments according to circumstances without any conflict, so as to obtain different technical solutions without substantially departing from the spirit of the present invention, which also falls within the scope of the present invention.
Claims (4)
1. The prediction early warning method of the comprehensive online sewage treatment prediction model is characterized by comprising the following steps of:
s1, establishing an accurate offline model in the sewage treatment facility and the sewage treatment process;
s2, communicating data between the off-line model and the sewage treatment informatization platform to form an on-line prediction model connected with the sewage treatment informatization platform; the sewage treatment informatization platform acquires first-class data including operation parameters of sewage treatment facilities and water quality of sewage inflow water; the first type of data comprises first machine inspection data which are collected by a sewage treatment facility on line and detected by utilizing manual daily collection;
s3, simulating and calculating first class data by using an online prediction model, generating second class data comprising calculation results and prediction early warning information, and sharing the second class data to a sewage treatment informatization platform;
in step S2, the sewage treatment informatization platform further obtains third class data including the quality of the sewage effluent, wherein the third class data includes third machine inspection data collected by the sewage treatment facility on line and third human inspection data collected and detected by using manual work every day;
the prediction and early warning method adopting the comprehensive online prediction model for sewage treatment comprises the following sub-steps:
s3-0, steady state calculation: the online prediction model simulates and calculates the average value of first person detection data in the past several days so as to perform steady-state calculation;
s3-1, dynamic calculation: using an online prediction model to simulate and calculate first machine inspection data in the past hours, and simulating and calculating and outputting second class data in the future hours;
step S3-0 includes:
s3-0-0, comparing the second class data output by steady state calculation with the average value of third human detection data in the past days, and judging whether the second class data is consistent with the average value of the third human detection data;
s3-0-1, if the online prediction model is effective, performing the next step of dynamic calculation;
if not, the online model is calibrated for the first time: correcting the parameter input value of the online prediction model according to the average value of the first human detection data in the past several days so as to enable the comparison result to be consistent;
the step S3-1 is further followed by the steps of:
s3-2, carrying out dynamic calculation simulation calculation in the step S3-1, outputting an output value of the second class data in a plurality of hours in the future and third machine-check data in a plurality of hours in the future, judging whether the output value is consistent with the third machine-check data in a plurality of hours in the future, and verifying the validity of the online prediction model again;
when judging whether the output values of the third machine inspection data in the next several hours in the step S3-2 are consistent with the output values of the second class data in the next several hours through analog calculation and output of the online prediction model;
if yes, the online prediction model is effective;
if not, the following steps are carried out:
s3-2-0, comparing the first machine inspection data with the first person inspection data, and verifying the validity of the first machine inspection data;
in step S3-2-0:
if the first machine inspection data and the first person inspection data are not consistent, maintaining a sewage treatment facility for collecting the first machine inspection data by taking the first person inspection data as a reference;
if the first machine inspection data and the first person inspection data are consistent, and the second type data which are calculated and output by the online prediction model through dynamic calculation in the past days are inconsistent with the first type data, the parameter correction of the online prediction model is required to be carried out by a professional.
2. The predictive and early-warning method of the comprehensive online sewage treatment predictive model according to claim 1, wherein the method is characterized by comprising the following steps of: in the online prediction model, steady state calculations are performed once a day, and in step S3-0, the steady state calculations for each day each simulate first human detection data over a period of time since the day using the online prediction model.
3. The predictive early warning method of the comprehensive online sewage treatment predictive model according to claim 2, which is characterized by comprising the following steps: in the daily steady-state calculation and the dynamic calculation of step S3-0 to step S3-1, the result of the simulation calculation of the daily steady-state calculation is taken as the input value of the first dynamic calculation on the same day, and the remaining results of the last dynamic calculation are taken as the starting point for each dynamic calculation.
4. The predictive early warning method of the comprehensive online sewage treatment predictive model according to claim 2 is characterized by comprising the following steps: in the daily steady state calculation of step S3-0, the first human data within the past 30 days of the deadline day are simulated using the online predictive model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310086581.4A CN115809749B (en) | 2023-02-09 | 2023-02-09 | Method for establishing comprehensive online prediction model of sewage treatment and prediction and early warning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310086581.4A CN115809749B (en) | 2023-02-09 | 2023-02-09 | Method for establishing comprehensive online prediction model of sewage treatment and prediction and early warning method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115809749A CN115809749A (en) | 2023-03-17 |
| CN115809749B true CN115809749B (en) | 2023-09-12 |
Family
ID=85487705
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310086581.4A Active CN115809749B (en) | 2023-02-09 | 2023-02-09 | Method for establishing comprehensive online prediction model of sewage treatment and prediction and early warning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115809749B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110470810A (en) * | 2019-07-09 | 2019-11-19 | 中电建水环境治理技术有限公司 | Monitoring water quality method and apparatus |
| CN111125964A (en) * | 2019-12-27 | 2020-05-08 | 华东理工大学 | Sewage treatment process proxy model construction method based on kriging interpolation method |
| CN111487950A (en) * | 2020-04-24 | 2020-08-04 | 西安交通大学 | 'prediction-verification-feedback-optimization' closed-loop system for online early warning and offline diagnosis |
| CN111639881A (en) * | 2020-06-10 | 2020-09-08 | 浙江和达科技股份有限公司 | Sewage discharge treatment method and system |
| CN113033092A (en) * | 2021-03-24 | 2021-06-25 | 石化盈科信息技术有限责任公司 | Sewage treatment data processing method and device, computer equipment and storage medium |
| CN113947034A (en) * | 2021-08-26 | 2022-01-18 | 深圳市广汇源环境水务有限公司 | Smart city wading link overall process coupling simulation method |
-
2023
- 2023-02-09 CN CN202310086581.4A patent/CN115809749B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110470810A (en) * | 2019-07-09 | 2019-11-19 | 中电建水环境治理技术有限公司 | Monitoring water quality method and apparatus |
| CN111125964A (en) * | 2019-12-27 | 2020-05-08 | 华东理工大学 | Sewage treatment process proxy model construction method based on kriging interpolation method |
| CN111487950A (en) * | 2020-04-24 | 2020-08-04 | 西安交通大学 | 'prediction-verification-feedback-optimization' closed-loop system for online early warning and offline diagnosis |
| CN111639881A (en) * | 2020-06-10 | 2020-09-08 | 浙江和达科技股份有限公司 | Sewage discharge treatment method and system |
| CN113033092A (en) * | 2021-03-24 | 2021-06-25 | 石化盈科信息技术有限责任公司 | Sewage treatment data processing method and device, computer equipment and storage medium |
| CN113947034A (en) * | 2021-08-26 | 2022-01-18 | 深圳市广汇源环境水务有限公司 | Smart city wading link overall process coupling simulation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115809749A (en) | 2023-03-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Rieger et al. | Data Reconciliation for Wastewater Treatment Plant Simulation Studies—Planning for High‐Quality Data and Typical Sources of Errors | |
| Langergraber et al. | A guideline for simulation studies of wastewater treatment plants | |
| Rieger et al. | Progress in sensor technology-progress in process control? Part I: Sensor property investigation and classification | |
| CN110110466B (en) | Sewage treatment target parameter prediction method for machine learning and process simulation | |
| CN101839817A (en) | Experimental method for intelligent detection of poisonous and harmful gas alarm equipment and device thereof | |
| CN110015754B (en) | Online monitoring system and method for distributed type sewage effluent quality standard exceeding early warning | |
| CN109879475A (en) | Dynamic adjustment type sewage operating condition processing method | |
| CN115685792B (en) | Wastewater intermittent drain outlet flow triggering method and device based on flow threshold | |
| CN114297954A (en) | Big data intelligent analysis digital management and control platform based on sewage treatment plant | |
| CN110458529A (en) | A kind of water pollution prediction method based on big data | |
| CN201680985U (en) | Intelligent detecting and testing device for poisonous and harmful gas alarm | |
| KR100920522B1 (en) | Device and method for diagnosing and controlling wastewater treatment process | |
| CN110132629B (en) | Method for predicting operation effectiveness of rural domestic sewage treatment facility by using support vector machine | |
| CN115343419A (en) | Sewage treatment plant carbon emission metering method and device based on greenhouse gas detection | |
| CN115809749B (en) | Method for establishing comprehensive online prediction model of sewage treatment and prediction and early warning method | |
| CN108760664B (en) | Water quality multi-parameter automatic detection and analysis method | |
| CN113533672A (en) | Water quality on-line monitoring method and device | |
| Pan et al. | A new approach to estimating oxygen off-gas fraction and dynamic alpha factor in aeration systems using hybrid machine learning and mechanistic models | |
| CN104502560B (en) | The chemical examination chamber system checked and accepted for Coal Quality | |
| EP3570025A2 (en) | Cylinder management system, cylinder management program and gas leak detection system | |
| Rieger et al. | Computer-aided monitoring and operation of continuous measuring devices | |
| CN116263850A (en) | Online sewage water quality early warning method combining offline simulation data | |
| Villez et al. | Structural observability and redundancy classification for sensor networks in wastewater systems | |
| CN112946227B (en) | Water quality monitoring system and water quality monitoring method | |
| Dold et al. | Practical guidance for WWTP model calibration and associated data gathering requirements |
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 |