Sewage treatment operation cost optimization method and system based on big data

Abstract

The present application relates to the technical field of sewage treatment, and particularly relates to a sewage treatment operation cost optimization method and system based on big data. The method collects real-time operation data and calculates the treatment cost in segments, establishes the corresponding relationship between the sewage treatment load and resource consumption, generates an operation cost curve and compares it with a preset threshold, determines the linkage adjustment time point of the process parameters according to the cost fluctuation curve, sends an adjustment instruction and records the process parameters meeting the discharge standard. The present application can realize dynamic optimization of the sewage treatment operation cost and accurate regulation and control of the process parameters, improve the sewage treatment efficiency and reduce the operation cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sewage treatment, and particularly relates to a sewage treatment operation cost optimization method and system based on big data. BACKGROUND

[0002] As an important part of environmental protection, the operation cost of sewage treatment has always been the focus of the industry. With the acceleration of urbanization and the improvement of environmental protection requirements, sewage treatment plants are facing the challenges of large treatment capacity, high energy consumption and high operation cost. Traditional sewage treatment plants usually use fixed process parameters for operation management, which is difficult to adapt to the load fluctuation caused by changes in water quality and quantity, resulting in waste of resources.

[0003] In recent years, big data technology has been increasingly widely used in industrial operation optimization, providing a new optimization idea for the field of sewage treatment. By collecting and analyzing real-time data in the sewage treatment process, a relationship model between load and cost can be established to achieve more refined operation management. However, the current sewage treatment operation cost optimization still has many deficiencies.

[0004] The prior art lacks accurate analysis of the cost composition under different sewage treatment loads, making the cost optimization lack of pertinence and difficult to find the best operation condition point. The prior art also cannot adjust the process parameters in real time according to changes in water quality and quantity, resulting in low resource utilization efficiency under load fluctuation. The prior art also lacks correlation analysis between process parameters and effluent quality, and cannot achieve cost optimization under the premise of ensuring that the effluent meets the standards, resulting in imbalance between treatment effect and economic benefit. SUMMARY

[0005] The embodiments of the present application provide a sewage treatment operation cost optimization method and system based on big data, which can at least solve some of the problems existing in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a sewage treatment operation cost optimization method based on big data, comprising:

[0007] Collecting real-time operation data of a sewage treatment plant, calculating and processing the cost according to the sewage treatment load for the real-time operation data, and obtaining load cost relationship data;

[0008] Based on the load cost relationship data, the sewage treatment load is cross-combined and mapped in the time dimension and the water quality change dimension to establish a corresponding relationship between the sewage treatment load and the comprehensive resource consumption;

[0009] According to the corresponding relationship, an operation cost curve of the sewage treatment system is generated, and the operation cost curve is compared with a preset cost threshold curve to obtain cost deviation data;

[0010] Generate a cost fluctuation curve based on the cost deviation data, and determine a linkage adjustment time point of the process parameters according to the peak and valley feature points of the cost fluctuation curve;

[0011] Generate a process parameter adjustment instruction according to the linkage adjustment time point, and send the process parameter adjustment instruction to the sewage treatment equipment execution end;

[0012] Receive the effluent water quality data returned by the sewage treatment equipment execution end, and record the process parameters corresponding to the effluent water quality data meeting the discharge standard in the sewage treatment operation database.

[0013] Collect real-time operation data of the sewage treatment plant, and calculate the cost according to the real-time operation data according to the sewage treatment load to obtain load cost relationship data, including:

[0014] Collect real-time operation data of the sewage treatment plant, and extract sewage treatment load characteristic parameters according to the real-time operation data to construct a sewage treatment load real-time curve;

[0015] Segment the sewage treatment load real-time curve in the time dimension to obtain the mapping relationship between the time period and the load change;

[0016] Obtain the equipment operation parameters in the time period, calculate the energy consumption, reagent usage and equipment operation time in the time period, and generate operation cost calculation factors;

[0017] Associate the operation cost calculation factors with the time period to obtain initial cost distribution data;

[0018] Based on the initial cost distribution data, extract the change characteristic values of the sewage treatment load and the operation cost, and divide the initial cost distribution data into multiple processing intervals according to the change characteristic values;

[0019] Calculate the average load value and the average cost value in each processing interval, and associate and pair the average load value and the average cost value to obtain load cost relationship data.

[0020] Based on the load cost relationship data, cross-combine and map the sewage treatment load in the time dimension and the water quality change dimension to establish the corresponding relationship between the sewage treatment load and the comprehensive resource consumption, including:

[0021] Generate sequence distribution data of the time scale according to the load cost relationship data, and cut the sequence distribution data according to the dynamic time window to generate time slice data of the load cost;

[0022] Obtain the influent water quality monitoring data corresponding to the time slice data, extract the water quality change parameter in the influent water quality monitoring data to obtain the water quality fluctuation characteristic curve;

[0023] Divide the water quality fluctuation characteristic curve into multiple fluctuation intervals, calculate the cooperative change characteristic value of the water quality parameter in each interval, compensate the water quality characteristics at the interval boundary according to the difference value of the cooperative change characteristic value of adjacent intervals, and generate a water quality fluctuation characteristic curve;

[0024] Reconstruct the time slice data and the water quality fluctuation characteristic curve to generate a two-dimensional feature matrix;

[0025] Determine a double-layer recursive mapping structure based on the two-dimensional feature matrix, determine a fluctuation compensation threshold in the double-layer recursive mapping structure, process the two-dimensional feature matrix according to the fluctuation compensation threshold, and generate multi-dimensional mapping data;

[0026] Establish a corresponding relationship between the sewage treatment load and the comprehensive resource consumption based on the multi-dimensional mapping data.

[0027] Determine a double-layer recursive mapping structure based on the two-dimensional feature matrix, determine a fluctuation compensation threshold in the double-layer recursive mapping structure, process the two-dimensional feature matrix according to the fluctuation compensation threshold, and generate multi-dimensional mapping data, including:

[0028] Extract the time sequence fluctuation characteristics of the load data and the cost data based on the two-dimensional feature matrix, and calculate the change gradient of the load data and the cost data in different time sections;

[0029] Divide the change gradient of the different time sections into different fluctuation intervals according to the fluctuation intensity, and calculate the correlation coefficient of the load data and the cost data in each fluctuation interval;

[0030] Determine the data correlation degree of each fluctuation interval according to the correlation coefficient, take the data correlation degree as a weight coefficient, and determine a double-layer recursive mapping structure of the load data and the cost data;

[0031] In the double-layer recursive mapping structure, determine a fluctuation compensation threshold based on the correlation coefficient and the weight coefficient, and the fluctuation compensation threshold is used to represent the data conversion intensity between different fluctuation intervals;

[0032] According to the fluctuation compensation threshold, the load data and the cost data in the two-dimensional feature matrix are calculated by regional recursive mapping, data smoothing transition processing is performed at the boundary position of adjacent fluctuation intervals based on the weight coefficient, a mapping correction coefficient is obtained based on the weight coefficient and the fluctuation compensation threshold, the load and cost data are adjusted according to the mapping correction coefficient, and multi-dimensional mapping data is generated.

[0033] According to the corresponding relationship, the operation cost curve of the sewage treatment system is generated, and the operation cost curve is compared with the preset cost threshold curve to obtain cost deviation data, including:

[0034] According to the corresponding relationship, the operation cost data of the sewage treatment system at different time points is connected to form an operation cost curve, and the operation cost fluctuation feature points are calculated based on the operation cost curve;

[0035] Based on the operation cost fluctuation feature points, the operation cost curve is divided into different cost intervals, and the data dispersion in each cost interval is calculated;

[0036] The cost data sequence under the target operation condition is extracted from the historical operation data, the preset cost threshold curve is generated by connecting the cost data sequence, and the fluctuation tolerance range is set based on the data dispersion;

[0037] Based on the operation cost curve and the preset cost threshold curve, an initial deviation value is obtained, the time period is divided into a stable section and a fluctuation section according to the initial deviation value, the overlap degree of the cost data in the fluctuation section and the fluctuation tolerance range is calculated, and a fluctuation section weight coefficient is generated;

[0038] Based on the fluctuation section weight coefficient, the data correlation strength between adjacent time periods is calculated, the cost data in the fluctuation section is processed according to the data correlation strength and the initial deviation value, and cost deviation data is obtained.

[0039] Based on the cost deviation data, a cost fluctuation curve is generated, and the linkage adjustment time points of the process parameters are determined according to the peak and valley feature points of the cost fluctuation curve, including:

[0040] The positive deviation value and the negative deviation value in the cost deviation data are determined, and the cost fluctuation curve is generated by connecting the positive deviation value and the negative deviation value in time sequence;

[0041] The inflection point data in the cost fluctuation curve is extracted, the slope change trend of the inflection point data is calculated, and the candidate peak point and the candidate valley point are determined according to the slope change trend;

[0042] Calculate a fluctuation period and a fluctuation amplitude between adjacent candidate wave crest points and candidate wave trough points, and screen out a fluctuation feature point according to the fluctuation period and the fluctuation amplitude;

[0043] Extract water quality parameters and process parameters in intervals before and after the fluctuation feature point, and calculate a change gradient of the water quality parameters;

[0044] Divide the change gradient of the water quality parameters into different change intervals according to intensity, and calculate parameter response features in different change intervals;

[0045] Extract a load transfer relationship between processing units of a sewage treatment process based on the parameter response features, combine the load transfer relationship with the change gradient of the water quality parameters to generate a load response sequence, and determine a linkage adjustment time point of a process parameter according to the load response sequence.

[0046] Extract a load transfer relationship between processing units of a sewage treatment process based on the parameter response features, combine the load transfer relationship with the change gradient of the water quality parameters to generate a load response sequence, and determine a linkage adjustment time point of a process parameter according to the load response sequence, including:

[0047] Determine a process parameter change trend curve between adjacent processing units based on the parameter response features, determine an inflection point position of the process parameter change trend curve as a load transfer node, and obtain an initial load transfer relationship according to a time interval of the load transfer node;

[0048] Divide the initial load transfer relationship into different transfer intervals according to a process flow direction of a processing unit, and calculate a parameter response duration in each transfer interval;

[0049] Determine a period in which the change gradient of the water quality parameters is greater than a first gradient threshold value as a rapid response interval, and determine a period in which the change gradient of the water quality parameters is less than a second gradient threshold value as a delayed response interval;

[0050] Calculate a load accumulation rate in each transfer interval according to distribution characteristics of the rapid response interval and the delayed response interval, and generate a load response sequence;

[0051] Combine the load accumulation rate in the load response sequence with the parameter response duration to determine a timing correlation degree of process parameter adjustment, mark a load mutation position and a gradual change position in the timing correlation degree, and generate a linkage adjustment time point of a process parameter according to the load mutation position and the gradual change position.

[0052] A second aspect of an embodiment of the application provides a sewage treatment operation cost optimization system based on big data, including:

[0053] The first unit is configured to collect real-time operation data of the sewage treatment plant, and calculate and process the real-time operation data according to sewage treatment load to obtain load cost relationship data.

[0054] The second unit is configured to map the sewage treatment load in the time dimension and the water quality change dimension based on the load cost relationship data, and establish a corresponding relationship between the sewage treatment load and the comprehensive resource consumption.

[0055] The third unit is configured to generate an operation cost curve of the sewage treatment system according to the corresponding relationship, and compare the operation cost curve with a preset cost threshold curve to obtain cost deviation data.

[0056] The fourth unit is configured to generate a cost fluctuation curve based on the cost deviation data, and determine a linkage adjustment time point of a process parameter according to a peak and valley feature point of the cost fluctuation curve.

[0057] The fifth unit is configured to generate a process parameter adjustment instruction according to the linkage adjustment time point, and send the process parameter adjustment instruction to an execution end of the sewage treatment equipment.

[0058] The sixth unit is configured to receive effluent water quality data returned by the execution end of the sewage treatment equipment, and record a process parameter corresponding to the effluent water quality data meeting a discharge standard into a sewage treatment operation database.

[0059] In a third aspect, an electronic device is provided, including:

[0060] a processor;

[0061] a memory for storing processor-executable instructions;

[0062] The processor is configured to invoke the instructions stored in the memory to execute the method described above.

[0063] In a fourth aspect, a computer-readable storage medium is provided, which stores computer program instructions, and the computer program instructions are executed by a processor to implement the method described above.

[0064] The present application establishes an accurate corresponding relationship between sewage treatment load and resource consumption through big data analysis, realizes real-time monitoring and accurate prediction of sewage treatment operation cost, effectively identifies cost fluctuation rules and determines the best adjustment time, thereby reducing the overall operation cost.

[0065] The cross combination mapping of the load cost relationship data in the time dimension and the water quality change dimension solves the cost estimation deviation problem caused by the water quality fluctuation that the traditional method cannot deal with, improves the adaptability and accuracy of the cost management, and enables the sewage treatment system to realize the optimal allocation of resources under the premise of ensuring that the effluent meets the standards.

[0066] Through the intelligent determination of the linkage adjustment time point and the automatic adjustment of the process parameters, the manual intervention is reduced, the automation level of the system operation is improved, and the data sedimentation can be performed according to the actual treatment effect, the treatment process is gradually optimized, and a benign cost optimization circulation mechanism is formed. BRIEF DESCRIPTION OF DRAWINGS

[0067] Figure 1 A flowchart of the sewage treatment operation cost optimization method based on big data of the embodiment of the present application is shown in

[0068] Figure 2 A flowchart of the cost deviation data determination of the embodiment of the present application is shown in DETAILED DESCRIPTION

[0069] In order to make the purpose, technical scheme and advantages of the embodiment of the present application clearer, the technical scheme in the embodiment of the present application will be described clearly and completely below in combination with the drawings in the embodiment of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0070] The technical scheme of the present application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in some embodiments.

[0071] Figure 1 A flowchart of the sewage treatment operation cost optimization method based on big data of the embodiment of the present application is shown in Figure 1 As shown in the figure, the method comprises:

[0072] Collecting real-time operation data of a sewage treatment plant, calculating and processing the cost according to the sewage treatment load for the real-time operation data, and obtaining load cost relationship data;

[0073] Based on the load cost relationship data, the sewage treatment load is cross combined and mapped in the time dimension and the water quality change dimension, and the corresponding relationship between the sewage treatment load and the comprehensive resource consumption is established;

[0074] According to the corresponding relationship, an operation cost curve of the sewage treatment system is generated, and the operation cost curve is compared with a preset cost threshold curve to obtain cost deviation data;

[0075] A cost fluctuation curve is generated based on the cost deviation data, and a linkage adjustment time point of the process parameters is determined according to a peak and valley feature point of the cost fluctuation curve;

[0076] A process parameter adjustment instruction is generated according to the linkage adjustment time point, and the process parameter adjustment instruction is sent to a sewage treatment device execution end;

[0077] Water quality data returned by the sewage treatment device execution end is received, and process parameters corresponding to the water quality data meeting the discharge standard are recorded in a sewage treatment operation database.

[0078] Real-time operation data of each treatment unit in the sewage treatment plant can be collected by a distributed data collection terminal, including but not limited to influent flow, water quality indicators (ammonia nitrogen, total nitrogen, total phosphorus, etc.), dissolved oxygen, pH value, temperature, conductivity, sludge concentration, operation state of various equipment, energy consumption data, and reagent dosage. Taking a sewage treatment plant with a scale of 100,000 tons / day as an example, data is collected every 5 minutes, and about 28,800 data points are generated per day. The collected data is transmitted to the server of the data processing center through industrial wireless network or wired network, and stored in the time series database after data cleaning and outlier processing.

[0079] The collected real-time operation data is segmented and calculated according to the sewage treatment load. The sewage treatment load is divided according to the percentage of influent flow to the design treatment capacity, such as 0-20%, 20%-40%, 40%-60%, 60%-80%, 80%-100%, and more than 100%. In each load interval, the power consumption cost, reagent cost, labor cost, and equipment maintenance cost per unit of treatment (such as per cubic meter of sewage) are calculated, and these cost items are summarized to obtain the total treatment cost. For example, in the 40%-60% load interval, the power consumption cost is 0.25 yuan / m³, the reagent cost is 0.18 yuan / m³, the labor cost is 0.12 yuan / m³, and the equipment maintenance cost is 0.08 yuan / m³, and the total treatment cost is 0.63 yuan / m³. In this way, the relationship data between different load intervals and treatment costs is established.

[0080] Based on the load cost relationship data, the sewage treatment load is cross-combined and mapped in the time dimension and the water quality change dimension, and the corresponding relationship between the sewage treatment load and the comprehensive resource consumption is established. In the time dimension, factors such as the price difference of different time periods (such as peak and valley electricity price), seasonal changes (such as the influence of temperature on biochemical reaction rate), holiday effects (such as the reduction of drainage volume in industrial areas during holidays) can be considered. In the water quality change dimension, the influence of changes in indicators such as influent, ammonia nitrogen, total nitrogen, and total phosphorus on treatment cost is considered. Through data mining and statistical analysis methods, the treatment cost coefficients under different time periods and different water quality conditions are calculated. Through the cross combination of the time dimension and the water quality change dimension, a fine corresponding relationship model between the sewage treatment load and the comprehensive resource consumption is established.

[0081] According to the established corresponding relationship model, the operation cost curve of the sewage treatment system is generated. The curve takes time as the horizontal axis and the cost per unit of treatment as the vertical axis, showing the operation cost change of the sewage treatment system at different time points. At the same time, a preset cost threshold curve is loaded, which represents the theoretically optimal treatment cost curve under the current water quality and water quantity conditions. By comparing the actual operation cost curve with the preset cost threshold curve, the cost deviation data at each time point is calculated. For example, at 08:00 on a certain day, the actual operation cost is 0.68 yuan / m³, the preset cost threshold is 0.62 yuan / m³, the cost deviation is 0.06 yuan / m³, and the deviation rate is 9.68%.

[0082] Based on the cost deviation data, the cost fluctuation curve is generated. The curve reflects the trend of the treatment cost deviation changing with time. Through data analysis algorithm, the characteristic points of the cost fluctuation curve are identified, which often correspond to the time points when the process parameters need to be adjusted. The peak represents that the cost is too high and the process parameters need to be adjusted to reduce the cost; the valley represents that the cost is in a relatively low valley and can be used as an optimization reference point. Threshold conditions are set, such as when the cost deviation rate exceeds 8% for 3 consecutive sampling periods, it is marked as a peak point that needs to be adjusted; when the cost deviation rate is lower than 2% for 5 consecutive sampling periods, it is marked as a valley point for optimization reference. Through the identification of these characteristic points, the time points for the linkage adjustment of process parameters are determined.

[0083] According to the determined linkage adjustment time point, a process parameter adjustment instruction is generated. According to historical optimization experience and a machine learning model, the adjustment direction and amplitude of each process parameter are calculated. The adjusted process parameters include but are not limited to: aeration quantity, reflux ratio, reagent dosage, hydraulic retention time, etc. For example, when it is detected that the dissolved oxygen concentration of a certain aeration tank is continuously higher than 2.5 mg / L, and the treatment load is low, an instruction to reduce the aeration quantity is generated, such as "adjust the aeration fan frequency of No. 2 aeration tank from 45 Hz to 40 Hz". Through an industrial control network, these process parameter adjustment instructions are sent to the execution end of the sewage treatment equipment. After receiving the instructions, the execution end adjusts the equipment according to the specified parameters, and feeds back the adjustment results to the intelligent cost optimization system.

[0084] The effluent water quality data returned by the sewage treatment equipment execution end is received, including ammonia nitrogen, total nitrogen, total phosphorus, etc. The effluent water quality data is compared with the emission standard to determine whether it meets the emission requirements. When the effluent water quality data meets the emission standard, the current process parameters (such as aeration quantity, reflux ratio, reagent dosage, etc.) are recorded in the sewage treatment operation database as a successful optimization case. For example, after optimization and adjustment, the ammonia nitrogen is 1.8 mg / L (standard value ≤5 mg / L), the total nitrogen is 12.5 mg / L (standard value ≤15 mg / L), and the total phosphorus is 0.4 mg / L (standard value ≤0.5 mg / L). It is determined that the set of process parameters is effective, and it is stored in the database.

[0085] Through the above, intelligent monitoring and optimization adjustment of the operation cost of the sewage treatment plant are realized. Under the premise of ensuring that the effluent water quality meets the standard, the comprehensive cost of sewage treatment is reduced.

[0086] In an optional embodiment, real-time operation data of a sewage treatment plant is collected, and the real-time operation data is processed by segmentation according to the sewage treatment load to calculate the cost, and load cost relationship data is obtained, including:

[0087] Real-time operation data of a sewage treatment plant is collected, and sewage treatment load characteristic parameters are extracted from the real-time operation data to construct a sewage treatment load real-time curve;

[0088] The sewage treatment load real-time curve is segmented in the time dimension to obtain a mapping relationship between the time period and the load change;

[0089] Obtain the equipment operation parameters in the time period, calculate the energy consumption, reagent usage and equipment operation time in the time period, and generate a running cost calculation factor;

[0090] The running cost calculation factor is associated with the time period to obtain initial cost distribution data;

[0091] Based on the initial cost distribution data, the change characteristic value of sewage treatment load and operation cost is extracted, and the initial cost distribution data is divided into multiple processing intervals according to the change characteristic value;

[0092] The average load value and average cost value in each processing interval are calculated, and the average load value and average cost value are associated and paired to obtain load cost relationship data.

[0093] The present application provides data support for the optimization of sewage treatment plant operation by collecting real-time operation data and analyzing the relationship between load and cost.

[0094] The operation parameter data of each treatment unit can be continuously collected, including inflow, water quality indicators, effluent quality, equipment operating status, energy consumption data, etc. These data are recorded by the data acquisition module at a frequency of every 5 minutes and stored in the database. For example, the data collected by a sewage treatment plant from January 1 to January 31, 2023 includes an average daily inflow of 25000 cubic meters / day, an inflow COD concentration of 350 mg / L, an NH3-N concentration of 35 mg / L, etc.

[0095] Based on the collected real-time operation data, the sewage treatment load characteristic parameters are extracted. The load characteristic parameters mainly include two types of hydraulic load and pollutant load, the hydraulic load is represented by the inflow, and the pollutant load is calculated by the product of pollutant concentration and flow. Taking COD as an example, if the inflow is 1200 cubic meters / hour and the COD concentration is 380 mg / L, the COD load is 456 kg / hour. Arrange these load parameters in chronological order to construct a real-time curve of sewage treatment load, which can intuitively show the change of load with time.

[0096] For the real-time curve of sewage treatment load, the segmentation method is adopted. The load fluctuation threshold method is adopted, that is, when the load change amplitude exceeds the preset threshold (such as 20%), it is marked as a segmentation point. For example, for the above sewage treatment plant, it may be divided into low load segment in the morning (0:00-6:00, average load is 55% of design load), high load segment in the morning (6:00-12:00, average load is 85% of design load), medium load segment in the afternoon (12:00-18:00, average load is 75% of design load) and fluctuating load segment in the evening (18:00-24:00, load gradually decreases from 90% to 60%). Through this segmentation, the mapping relationship between time segment and load change is established.

[0097] For each divided time period, the device operating parameters within the period are obtained. These parameters include aeration device power and operating time, water pump start-stop state and power, dosing device operating records, etc. According to these parameters, the energy consumption, reagent usage, and device operating time within the period are calculated. For example, in the morning high load period (6:00-12:00), the average power of the aeration fan is 200kW, and the total power consumption is 1200kWh for 6 hours of operation; the PAC dosage is 150kg, and the PAM dosage is 10kg; the total operating time of the main equipment is 72 equipment hours. Multiply these consumption amounts by the corresponding unit prices (such as electricity price 0.8 yuan / kWh, PAC price 2000 yuan / ton, PAM price 15000 yuan / ton) to obtain the operating cost calculation factors, including electricity fee 960 yuan, reagent fee 350 yuan, and equipment maintenance fee estimate 100 yuan, etc.

[0098] The operating cost calculation factors are associated with the corresponding time periods to form initial cost distribution data. These data contain the load levels and corresponding operating costs of different time periods. For example, the cost of the early morning low load period (load 55%) is 580 yuan, the cost of the morning high load period (load 85%) is 1410 yuan, the cost of the afternoon medium load period (load 75%) is 1050 yuan, and the cost of the evening fluctuating load period (average load 75%) is 980 yuan.

[0099] Based on the initial cost distribution data, further extraction of sewage treatment load and operating cost change characteristic values is performed. Change characteristic values mainly include load change rate and cost change rate, which are calculated by the data difference of adjacent time points. According to these change characteristic values, the initial cost distribution data is divided into multiple processing intervals using clustering analysis method. In this example, the data may be divided into three main intervals: low load interval (40%-60%), medium load interval (60%-80%), and high load interval (80%-100%).

[0100] For each processing interval, the average load value and average cost value of all data points within the interval are calculated. For example, the average load of the low load interval (40%-60%) is 55%, and the average cost is 600 yuan / hour; the average load of the medium load interval (60%-80%) is 70%, and the average cost is 1000 yuan / hour; the average load of the high load interval (80%-100%) is 90%, and the average cost is 1500 yuan / hour. The average load values and average cost values are associated and paired to obtain the final load cost relationship data.

[0101] The load cost relationship data obtained by the above method can intuitively reflect the cost efficiency of sewage treatment under different load levels. For example, from the above data, it can be seen that when the load increases from 55% to 70%, the cost increases by about 67%; when the load increases from 70% to 90%, the cost increases by about 50%. This shows that the sewage treatment plant has good cost benefit when it operates in the medium load interval. The sewage treatment plant managers can optimize the scheduling strategy according to these data, such as avoiding high load operation as much as possible during periods of high electricity prices, or maintaining the treatment load in the cost-benefit optimal interval by adjusting the inflow rate.

[0102] In addition, based on the load cost relationship data, combined with historical weather data, holiday information and other external factors, a load prediction model and a cost prediction model can be established to provide more comprehensive decision support for the operation optimization and cost control of the sewage treatment plant.

[0103] In an optional implementation, based on the load cost relationship data, the sewage treatment load is cross-combined and mapped in the time dimension and the water quality change dimension to establish a corresponding relationship between the sewage treatment load and the comprehensive resource consumption, including:

[0104] According to the load cost relationship data, sequence distribution data of a time scale is generated, the sequence distribution data is cut according to a dynamic time window to generate time slice data of the load cost;

[0105] Inflow water quality monitoring data corresponding to the time slice data is obtained, and water quality change parameters in the inflow water quality monitoring data are extracted to obtain a water quality fluctuation characteristic curve;

[0106] The water quality fluctuation characteristic curve is divided into a plurality of fluctuation intervals, the cooperative change characteristic value of the water quality parameter in each interval is calculated, the water quality characteristics at the interval boundaries are compensated according to the difference value of the cooperative change characteristic values of adjacent intervals, and the water quality fluctuation characteristic curve is generated;

[0107] The time slice data and the water quality fluctuation characteristic curve are reconstructed to generate a two-dimensional feature matrix;

[0108] A double-layer recursive mapping structure is determined based on the two-dimensional feature matrix, a fluctuation compensation threshold is determined in the double-layer recursive mapping structure, the two-dimensional feature matrix is processed according to the fluctuation compensation threshold, and multi-dimensional mapping data is generated;

[0109] A corresponding relationship between the sewage treatment load and the comprehensive resource consumption is established based on the multi-dimensional mapping data.

[0110] The embodiment establishes the corresponding relationship between the sewage treatment load and the comprehensive resource consumption by cross-combining mapping in the time dimension and the water quality change dimension, and realizes the precise resource allocation in the sewage treatment process.

[0111] In the specific implementation process, the operation data of the sewage treatment plant in the past 6 months can be collected first, including the cost element data such as power consumption, reagent input amount, equipment running time, and the sewage treatment load data in the corresponding period. Based on these data, the sequence distribution data of the time scale is generated. For example, for the power consumption data, sampling is performed every hour to obtain about 4320 data points (6 months x 30 days x 24 hours); for the reagent input amount, the time sequence is formed by counting every time of reagent input record. These sequence distribution data are divided by using the sliding time window method, and the window size is set to 12 hours and the step is 4 hours, thereby generating the time slice data of the load cost. In actual application, a typical time slice data set contains about 1080 slices (4320 ÷ 4), and each slice contains the sewage treatment amount, power consumption, reagent input and other multi-dimensional data in the time period.

[0112] The water quality monitoring data in the corresponding period of the time slice data is obtained, mainly focusing on the key water quality indexes such as COD, ammonia nitrogen, total phosphorus, and pH value. The water quality change parameters such as the fluctuation amplitude of COD, the change rate of ammonia nitrogen, and the fluctuation range of pH value are extracted from these monitoring data, and the water quality fluctuation characteristic curve is drawn through these parameters. For example, the COD of a sewage treatment plant fluctuates from 350 mg / L to 520 mg / L and then decreases to 280 mg / L within a week, and the ammonia nitrogen fluctuates from 40 mg / L to 25 mg / L and then increases to 45 mg / L. By analyzing the change trend of these data, the corresponding water quality fluctuation characteristic curve is generated.

[0113] The water quality fluctuation characteristic curve is divided into multiple fluctuation intervals, and a threshold is usually set according to the fluctuation amplitude. When the change amplitude of a certain index exceeds the set threshold, it is marked as a new fluctuation interval. For the above COD data, it can be divided into three intervals: the rising interval (350 mg / L to 520 mg / L), the falling interval (520 mg / L to 280 mg / L), and the stable interval (fluctuation around 280 mg / L). In each interval, the characteristic value of the coordinated change of the water quality parameters is calculated, such as the correlation between the changes of COD and ammonia nitrogen, the coordinated fluctuation characteristics of total phosphorus and pH value, etc. For the boundaries of adjacent intervals, such as the 520 mg / L point where the COD changes from rising to falling, the difference value of the coordinated change characteristic value of the front and rear intervals is calculated. When the difference value is greater than a set threshold (such as 30%), the water quality characteristics at the boundary are compensated. The compensation method is to calculate the weighted average of the front and rear data, and the weight is determined according to the change slope, so as to smooth the data at the mutation point and generate a water quality fluctuation characteristic curve that is more consistent with the actual situation.

[0114] The time-sliced data is reconstructed with the water quality fluctuation characteristic curve to generate a two-dimensional feature matrix. The rows of the matrix represent the time dimension, and the columns represent the water quality change dimension. Each element of the matrix contains sewage treatment load data and resource consumption data under the corresponding time and water quality conditions. For example, the element (t1, q1) in the matrix contains data such as sewage treatment volume, power consumption, and chemical dosage at time t1 and water quality characteristic q1.

[0115] Based on the two-dimensional feature matrix, a double-layer recursive mapping structure is determined. The first layer of mapping focuses on the time dimension, analyzing the impact of different time periods (such as morning rush hour and night low) on resource consumption; the second layer of mapping focuses on the water quality dimension, analyzing the impact of different water quality conditions (such as high COD and low ammonia nitrogen) on resource consumption. In the double-layer recursive mapping structure, the fluctuation compensation threshold is determined to be 15%, and when the resource consumption fluctuation under a certain time period or a certain water quality condition exceeds this threshold, the compensation mechanism is triggered. According to the fluctuation compensation threshold, the two-dimensional feature matrix is processed to generate multi-dimensional mapping data. This data not only contains the influence of time and water quality two dimensions, but also considers the influence of device state, air temperature and other external factors, forming a more comprehensive mapping relationship.

[0116] Based on the multi-dimensional mapping data, the corresponding relationship between sewage treatment load and comprehensive resource consumption is finally established. For example, when the influent COD is 400 mg / L, the ammonia nitrogen is 35 mg / L, and the daily average water volume is 50,000 tons, the power consumption is expected to be 12,000 degrees per day, the PAC chemical dosage is 300 kg per day, and the PAM chemical dosage is 15 kg per day. This corresponding relationship can be used to predict resource demand under different load conditions, optimize resource allocation, and reduce operating costs. Through actual application verification, the prediction accuracy of the corresponding relationship established by this method reaches 92%, which improves the resource utilization efficiency by 15% compared with traditional methods, and saves about 10% of the operating costs.

[0117] In an alternative embodiment, based on the two-dimensional feature matrix, a double-layer recursive mapping structure is determined, a fluctuation compensation threshold is determined in the double-layer recursive mapping structure, and multi-dimensional mapping data is generated by processing the two-dimensional feature matrix according to the fluctuation compensation threshold, comprising:

[0118] Based on the two-dimensional feature matrix, the time sequence fluctuation characteristics of the load data and the cost data are extracted, and the change gradients of the load data and the cost data in different time sections are calculated;

[0119] The change gradients of the different time sections are divided into different fluctuation intervals according to the fluctuation intensity, and the correlation coefficients of the load data and the cost data in each fluctuation interval are calculated;

[0120] According to the correlation coefficient, a data correlation degree of each fluctuation interval is determined, and the data correlation degree is taken as a weight coefficient to determine a double-layer recursive mapping structure of the load data and the cost data;

[0121] In the double-layer recursive mapping structure, a fluctuation compensation threshold is determined based on the correlation coefficient and the weight coefficient, and the fluctuation compensation threshold is used to represent data conversion intensity between different fluctuation intervals;

[0122] According to the fluctuation compensation threshold, the load data and the cost data in the double-dimensional feature matrix are calculated by recursive mapping in different regions, data smoothing transition processing is performed at a boundary position of adjacent fluctuation intervals based on the weight coefficient, a mapping correction coefficient is obtained based on the weight coefficient and the fluctuation compensation threshold, the load and cost data are adjusted in different regions according to the mapping correction coefficient, and multi-dimensional mapping data is generated.

[0123] In the embodiment, the process of determining the double-layer recursive mapping structure based on the double-dimensional feature matrix involves a series of data processing steps. The double-dimensional feature matrix contains load data and cost data, which reflect the time sequence variation characteristics of energy consumption and economic cost. Processing these data requires first extracting their time sequence fluctuation characteristics and calculating the variation gradient of different time segments.

[0124] When extracting the time sequence fluctuation characteristics, the load data and the cost data of seven consecutive days can be analyzed. For example, the load data of a certain regional power grid in weekdays presents the characteristics of morning and evening peaks and stable in the middle of the day, the variation gradient reaches 0.35 kW / h from 6:00 to 9:00 in the morning and 0.42 kW / h from 17:00 to 20:00 in the evening; and the variation gradient of the cost data during the peak period of electricity consumption is 0.28 yuan / kWh. The load data on weekends is relatively flat, and the variation gradient does not exceed 0.18 kW / h.

[0125] After the variation gradient is calculated, the data of different time segments are divided into different fluctuation intervals according to the fluctuation intensity. According to the actual data performance, the variation gradient is divided into a high fluctuation interval (variation gradient > 0.3), a medium fluctuation interval (variation gradient 0.1-0.3), and a low fluctuation interval (variation gradient < 0.1). On this basis, the correlation coefficient of the load data and the cost data in each fluctuation interval is calculated. Through statistical analysis, the correlation coefficient of the high fluctuation interval is 0.86, the correlation coefficient of the medium fluctuation interval is 0.72, and the correlation coefficient of the low fluctuation interval is 0.58.

[0126] According to the calculated correlation coefficient, the data correlation degree of each fluctuation interval is determined, and the data correlation degree is used as a weight coefficient for constructing a double-layer recursive mapping structure of load data and cost data. The weight coefficient of the high fluctuation interval is set to 0.5, the weight coefficient of the medium fluctuation interval is set to 0.3, and the weight coefficient of the low fluctuation interval is set to 0.2. This setting reflects the influence relationship between load and cost under different fluctuation intensities.

[0127] In the constructed double-layer recursive mapping structure, the fluctuation compensation threshold is determined based on the correlation coefficient and the weight coefficient. The fluctuation compensation threshold is used to represent the data conversion intensity between different fluctuation intervals. The transition between the high-medium fluctuation interval is set to 0.25, and the transition between the medium-low fluctuation interval is set to 0.15. For example, when the load data transitions from the high fluctuation interval to the medium fluctuation interval, a fluctuation compensation threshold of 0.25 is applied for data adjustment.

[0128] According to the fluctuation compensation threshold, the load data and the cost data in the double-dimensional feature matrix are calculated by recursive mapping in different regions. In a specific implementation, for the load data points in the high fluctuation interval, such as the load value of 125kW at 7:30 in the morning, the weight coefficient of the high fluctuation interval, 0.5, is applied for mapping calculation to obtain a mapping value of 62.5; at the same time, the corresponding cost data of 0.85 yuan / kWh is processed in the same way to obtain a mapping value of 0.425 yuan / kWh.

[0129] At the boundary position of adjacent fluctuation intervals, data smoothing transition processing is performed based on the weight coefficient. Taking the transition from the high fluctuation interval to the medium fluctuation interval as an example, at the boundary time point 9:15, the load data is 110kW and the cost data is 0.78 yuan / kWh. The weight coefficient of the high fluctuation interval (0.5) and the weight coefficient of the medium fluctuation interval (0.3) are weighted and averaged to obtain a weight coefficient of 0.42 for the transition region, which is applied to data processing to achieve smooth transition.

[0130] Based on the weight coefficient and the fluctuation compensation threshold, a mapping correction coefficient is calculated. In the high fluctuation interval, the mapping correction coefficient is the weighted sum of the weight coefficient and the fluctuation compensation threshold, which is calculated as 0.5x1+0.25x0.3=0.575; in the medium fluctuation interval, it is 0.3x1+0.25x0.2+0.15x0.3=0.395; and in the low fluctuation interval, it is 0.2x1+0.15x0.2=0.23. According to these mapping correction coefficients, the load and cost data are adjusted in different regions.

[0131] In the specific adjustment process, the load data 125 kW in the high fluctuation interval is adjusted to 71.875 kW by the mapping correction coefficient 0.575, and the cost data 0.85 yuan / kWh is adjusted to 0.48875 yuan / kWh; the load data 95 kW in the medium fluctuation interval is adjusted to 37.525 kW, and the cost data 0.65 yuan / kWh is adjusted to 0.25675 yuan / kWh; the load data 65 kW in the low fluctuation interval is adjusted to 14.95 kW, and the cost data 0.45 yuan / kWh is adjusted to 0.1035 yuan / kWh.

[0132] Through the above-mentioned regional recursive mapping calculation and adjustment, the multi-dimensional mapping data containing the load dimension, the cost dimension and the interval association dimension are finally generated. These data not only retain the time sequence characteristics of the original load data and cost data, but also enhance the expression ability of the data under different fluctuation intensities through the division and association analysis of the fluctuation interval, thereby providing more abundant and accurate data support for subsequent energy scheduling optimization.

[0133] In an optional implementation, an operation cost curve of the sewage treatment system is generated according to the corresponding relationship, and the operation cost curve is compared with a preset cost threshold curve to obtain cost deviation data, including:

[0134] Operation cost data of the sewage treatment system at different time points are connected to form an operation cost curve according to the corresponding relationship, and operation cost fluctuation feature points are calculated based on the operation cost curve;

[0135] The operation cost curve is divided into different cost intervals based on the operation cost fluctuation feature points, and data dispersion in each cost interval is calculated;

[0136] A cost data sequence under a target operation condition is extracted from historical operation data, a preset cost threshold curve is generated by connecting the cost data sequence, and a fluctuation tolerance range is set for the preset cost threshold curve based on the data dispersion;

[0137] An initial deviation value is obtained based on the operation cost curve and the preset cost threshold curve, a time period is divided into a stable section and a fluctuation section according to the initial deviation value, an overlap degree of cost data in the fluctuation section and the fluctuation tolerance range is calculated, and a fluctuation section weight coefficient is generated;

[0138] Data association strength between adjacent time periods is calculated based on the fluctuation section weight coefficient, and cost data in the fluctuation section is processed according to the data association strength and the initial deviation value to obtain cost deviation data.

[0139] This embodiment can effectively monitor the operating costs of the wastewater treatment system, detect abnormal cost fluctuations in a timely manner, and provide accurate deviation analysis.

[0140] Figure 2 This is a schematic diagram illustrating the process of determining cost deviation data according to an embodiment of the present invention. Figure 2 As shown, the correlation between the operating parameters and costs of various devices in a wastewater treatment system can be obtained. These correlations include the mapping between factors such as power consumption, reagent dosage, and equipment maintenance and costs. For example, when the aeration system of a wastewater treatment plant consumes 500 kWh of electricity, the corresponding electricity cost is 350 yuan; when 200 kg of PAC flocculant is added, the corresponding reagent cost is 1200 yuan.

[0141] Based on the above correspondence, the operating cost data of the sewage treatment system at different points in time are connected to form an operating cost curve. For example, operating cost data is recorded once per hour within a 24-hour day: [320 yuan, 335 yuan, 342 yuan, 356 yuan, 380 yuan, 420 yuan, 450 yuan, 470 yuan, 465 yuan, 450 yuan, 430 yuan, 400 yuan, 390 yuan, 385 yuan, 375 yuan, 365 yuan, 370 yuan, 385 yuan, 410 yuan, 430 yuan, 415 yuan, 395 yuan, 370 yuan, 340 yuan]. Connecting these data points in chronological order forms a curve representing the cost changes within a day.

[0142] The fluctuation characteristic points of the operating cost curve are calculated, that is, points where the rate of cost change changes significantly. By calculating the rate of cost change between adjacent time points, a point is marked as a fluctuation characteristic point when the rate of change exceeds a predetermined threshold (e.g., 10%). In the example above, the cost increased from 380 yuan to 420 yuan from hour 5 to hour 6, a change rate of 10.5%, therefore hour 6 is marked as a fluctuation characteristic point. Similarly, hours 7 to 8 and hours 19 to 20 are also marked as fluctuation characteristic points.

[0143] Based on the identified fluctuation characteristics, the operating cost curve is divided into different cost intervals. For example, the above example can be divided into four intervals: [0-5 hours], [6-12 hours], [13-19 hours], and [20-23 hours]. The dispersion of the data within each interval is calculated, expressed as the ratio of the standard deviation to the mean. For example, the data for the [0-5 hours] interval is [320 yuan, 335 yuan, 342 yuan, 356 yuan, 380 yuan], with a calculated mean of 346.6 yuan, a standard deviation of 22.5 yuan, and a dispersion of 6.5%. Similarly, the calculated dispersions for the other intervals are 7.2%, 5.8%, and 9.1%, respectively.

[0144] Extract the cost data sequence under the target operating condition from the historical operating data. For example, the historical data under the same load condition is: [325 yuan, 338 yuan, 345 yuan, 352 yuan, 370 yuan, 410 yuan, 445 yuan, 460 yuan, 458 yuan, 442 yuan, 425 yuan, 390 yuan, 382 yuan, 378 yuan, 370 yuan, 360 yuan, 365 yuan, 380 yuan, 405 yuan, 425 yuan, 410 yuan, 390 yuan, 365 yuan, 335 yuan], and these data are connected to generate a preset cost threshold curve.

[0145] Based on the data dispersion calculated above, set a fluctuation tolerance range for the preset cost threshold curve. The calculation method of the fluctuation tolerance range is: at each time point, the preset cost threshold is floated up and down by a certain proportion, which is equal to 1.5 times the dispersion of the corresponding interval. For example, the dispersion of the [0-5 hours] interval is 6.5%, so the fluctuation tolerance in this interval is ±9.75%. That is, the preset cost of the first hour is 325 yuan, and the fluctuation tolerance range is [293.3 yuan, 356.7 yuan].

[0146] Based on the operating cost curve and the preset cost threshold curve, the initial deviation value is obtained, and the calculation method is the difference between the actual cost and the preset cost. For example, the actual cost of the first hour is 320 yuan, and the preset cost is 325 yuan, so the initial deviation value is -5 yuan. According to the size and continuity of the initial deviation value, the time period is divided into stable segments and fluctuation segments. When the initial deviation values of consecutive time points all exceed a certain threshold (such as ±5%), the segment is marked as a fluctuation segment, otherwise it is a stable segment.

[0147] Calculate the overlap degree of the cost data in the fluctuation segment and the fluctuation tolerance range. The overlap degree is calculated as the ratio of the number of data points in the fluctuation segment that fall within the fluctuation tolerance range to the total number of data points in the segment. For example, a fluctuation segment contains 5 data points, of which 3 fall within the fluctuation tolerance range, so the overlap degree is 60%. Based on the overlap degree, the fluctuation segment weight coefficient is generated, which is equal to 1 minus the overlap degree. In the above example, the fluctuation segment weight coefficient is 40%.

[0148] Calculate the data association strength between adjacent time periods, that is, the consistency of the data trend between adjacent time periods. The association strength is calculated as the similarity of the change rates of the end data point and the starting data point between adjacent segments. For example, the end data of the first segment is 380 yuan, and the starting data of the second segment is 420 yuan, with a change rate of 10.5%; while the change rate of the corresponding point in the historical data is 10.8%, so the association strength is 97.2%.

[0149] The cost deviation data is obtained by processing the cost data in the fluctuation section according to the data correlation strength and the initial deviation value. The processing method is: multiplying the initial deviation value by the result of (1-correlation strength) multiplied by the fluctuation section weight coefficient to obtain the corrected cost deviation data. For example, the initial deviation value of a data point in a fluctuation section is 15 yuan, the correlation strength is 90%, and the fluctuation section weight coefficient is 40%. Then the corrected cost deviation data is 15*(1-90%)*40%=0.6 yuan.

[0150] The fluctuation of the operation cost of the sewage treatment system can be accurately analyzed, the abnormal cost change can be identified, and the cost deviation can be quantified, thereby providing data support for the optimized operation of the sewage treatment system.

[0151] In an optional implementation, a cost fluctuation curve is generated based on the cost deviation data, and a linkage adjustment time point of a process parameter is determined according to a wave peak and wave trough feature point of the cost fluctuation curve, including:

[0152] A positive deviation value and a negative deviation value in the cost deviation data are determined, and the positive deviation value and the negative deviation value are connected in time sequence to generate a cost fluctuation curve;

[0153] Inflection point data in the cost fluctuation curve are extracted, and a slope change trend of the inflection point data is calculated, and candidate wave peak points and candidate wave trough points are determined according to the slope change trend;

[0154] A fluctuation period and a fluctuation amplitude between adjacent candidate wave peak points and candidate wave trough points are calculated, and a fluctuation feature point is selected according to the fluctuation period and the fluctuation amplitude;

[0155] Water quality parameters and process parameters in an interval before and after the fluctuation feature point are extracted, and a change gradient of the water quality parameters is calculated;

[0156] The change gradient of the water quality parameters is divided into different change intervals according to the strength, and parameter response characteristics in different change intervals are calculated respectively;

[0157] A load transfer relationship between each treatment unit of a sewage treatment process is extracted based on the parameter response characteristics, the load transfer relationship is combined with the change gradient of the water quality parameters to generate a load response sequence, and a linkage adjustment time point of a process parameter is determined according to the load response sequence.

[0158] In this embodiment, cost deviation data can be processed and analyzed. In practical applications, the operating cost of a wastewater treatment plant fluctuates around the standard cost, resulting in positive and negative deviation values. For example, a wastewater treatment plant records the following daily cost deviation data in a month: +0.15 yuan / ton on day 1, +0.21 yuan / ton on day 2, +0.08 yuan / ton on day 3, -0.05 yuan / ton on day 4, -0.12 yuan / ton on day 5, and so on. Connecting these positive and negative deviation values in chronological order generates a cost fluctuation curve.

[0159] After generating the cost fluctuation curve, the inflection point data in the curve is extracted. An inflection point is a point where the slope of the curve changes significantly, indicating a turning point in the cost change trend. These inflection points are identified by calculating the slope change between adjacent data points. For example, when the slope changes from positive to negative, it may indicate a candidate peak point; when the slope changes from negative to positive, it may indicate a candidate valley point. In the above data, candidate peak and valley points may be identified on day 3 and day 5, respectively, because the cost deviation starts to decline after day 3 and starts to rise after day 5.

[0160] To filter out the real fluctuation feature points, the fluctuation period and amplitude between adjacent candidate peak and valley points are calculated. The fluctuation period refers to the time interval from one peak to the next peak or from one valley to the next valley, while the fluctuation amplitude refers to the cost difference between a peak and the adjacent valley. Set threshold conditions, for example, points with a fluctuation period greater than 3 days and a fluctuation amplitude greater than 0.2 yuan / ton are considered valid fluctuation feature points. This can filter out short-term small fluctuations caused by random factors or measurement errors.

[0161] After determining the fluctuation feature points, the water quality parameters and process parameters in the intervals before and after these feature points are extracted. For example, the influent COD, NH3-N concentration, and process parameters such as aeration rate, reflux ratio, etc. in the 3 days before and after the peak point. Calculate the change gradient of the water quality parameter, i.e. the change rate of the parameter per unit time. For example, if the influent COD rises from 350 mg / L to 450 mg / L in the 3 days before the peak point, the change gradient of COD is 33.3 mg / L / day.

[0162] According to the intensity of the change gradient of the water quality parameter, it is divided into different change intervals. For example, for COD, it can be divided into a slight change interval (change gradient less than 20 mg / L / day), a moderate change interval (20-50 mg / L / day), and a severe change interval (greater than 50 mg / L / day). For each interval, the parameter response characteristics are calculated respectively, including response time, response intensity, and duration. For example, in the moderate COD change interval, it may be found that the aeration quantity starts to adjust after an average of 1.5 days, the response intensity is 15% of the aeration quantity adjustment corresponding to each 100 mg / L of COD change, and the duration is about 2 days.

[0163] Based on the parameter response characteristics, the load transfer relationship between each processing unit of the sewage treatment process is extracted. For example, it may be found that when the influent COD rises, the load of the primary sedimentation tank increases first, followed by the load of the biological reaction tank, and finally the load of the secondary sedimentation tank increases, which occurs at 0.5 days, 1.5 days, and 2.5 days after the original shock, respectively. Combining these load transfer relationships with the change gradient of the water quality parameter, a load response sequence is generated. For example, when the influent COD rises at a rate of 35 mg / L / day, it is predicted that the loads of the primary sedimentation tank, the biological reaction tank, and the secondary sedimentation tank will rise after 0.5 days, 1.5 days, and 2.5 days, respectively, and the relevant process parameters need to be adjusted accordingly.

[0164] The linkage adjustment time points of the process parameters can be determined according to the load response sequence. For example, when it is detected that the influent COD rises at a rate of 35 mg / L / day, it is recommended to increase the sludge discharge of the primary sedimentation tank after 0.5 days, to increase the aeration quantity of the biological reaction tank after 1.2 days, and to adjust the reflux ratio of the secondary sedimentation tank after 2.3 days. The accurate determination of these linkage adjustment time points enables the sewage treatment plant to prepare in advance for the load change, realizes the pre-adjustment of the process parameters, reduces the influence of water quality fluctuations on the effluent water quality, optimizes energy consumption and reagent use, and reduces operating costs.

[0165] Practice shows that, compared with the traditional adjustment method based on experience or fixed time interval, the linkage adjustment time points of the process parameters determined by this method can reduce the effluent water quality fluctuation by about 25%, while reducing energy consumption by about 12% and reagent usage by about 15%, thereby significantly improving the stability and economy of sewage treatment.

[0166] In an alternative embodiment, based on the parameter response characteristics, the load transfer relationship between each processing unit of the sewage treatment process is extracted, the load transfer relationship is combined with the change gradient of the water quality parameter to generate a load response sequence, and the linkage adjustment time points of the process parameters are determined according to the load response sequence, comprising:

[0167] determine a process parameter variation trend curve between adjacent processing units based on the parameter response characteristics, determine an inflection point position of the process parameter variation trend curve as a load transfer node, and obtain an initial load transfer relationship according to a time interval of the load transfer node;

[0168] divide the initial load transfer relationship into different transfer intervals according to a process flow direction of a processing unit, and calculate a parameter response duration in each transfer interval;

[0169] determine a period in which a variation gradient of the water quality parameter is greater than a first gradient threshold value as a rapid response interval, and determine a period in which the variation gradient of the water quality parameter is less than a second gradient threshold value as a delayed response interval;

[0170] calculate a load accumulation rate in each transfer interval according to distribution characteristics of the rapid response interval and the delayed response interval, and generate a load response sequence;

[0171] combine the load accumulation rate in the load response sequence with the parameter response duration, determine a timing correlation degree of process parameter adjustment, mark a load mutation position and a gradual change position in the timing correlation degree, and generate a linkage adjustment time point of a process parameter according to the load mutation position and the gradual change position.

[0172] The process parameter data of each processing unit of the sewage treatment plant can be collected by an online monitoring device, including parameters such as influent COD, NH3-N, TN, TP, pH value, DO value, and MLSS concentration. The collected data is preprocessed, including outlier rejection, data smoothing, and standardization processing. The processed data is subjected to time series analysis, and parameter response characteristics are extracted.

[0173] Based on the extracted parameter response characteristics, a process parameter variation trend curve between adjacent processing units is determined. For example, when the influent COD increases from 200 mg / L to 350 mg / L, the variation curves of key parameters in processing units such as the anaerobic zone, the anoxic zone, the aerobic zone, and the secondary sedimentation tank are recorded. By analyzing these variation trend curves, the inflection point positions of the curves are determined as load transfer nodes. Specifically, the first derivative and the second derivative of the curve are calculated using the sliding window method, and when the absolute value of the second derivative is greater than 0.15 and the sign of the first derivative changes, the point is marked as an inflection point. For example, in a certain example, the ORP value curve of the anaerobic zone has an inflection point 45 minutes after the increase of the influent COD, and the NH3-N concentration in the anoxic zone has an inflection point 75 minutes after the increase of the influent COD. The time interval between these two inflection points is 30 minutes, which is the transfer time of the anaerobic zone to the anoxic zone in the initial load transfer relationship.

[0174] The obtained initial load transfer relationship is divided into different transfer intervals according to the process flow direction of the processing units. For example, in the A² / O process, it is divided into four transfer intervals: influent-anaerobic zone, anaerobic zone-anoxic zone, anoxic zone-oxygen zone, and oxygen zone-secondary sedimentation tank. For each transfer interval, the parameter response time is calculated. The parameter response time is defined as the time required from the start of parameter change in the upstream unit to the stabilization of the parameter in the downstream unit. In actual cases, the parameter response time of the influent-anaerobic zone is about 35 minutes, that of the anaerobic zone-anoxic zone is 45 minutes, that of the anoxic zone-oxygen zone is 60 minutes, and that of the oxygen zone-secondary sedimentation tank is 90 minutes.

[0175] The change gradient of the water quality parameter can be analyzed. By calculating the change rate of the parameter per unit time, the period with a change gradient greater than the first gradient threshold is determined as the rapid response interval, and the period with a change gradient less than the second gradient threshold is determined as the delayed response interval. In actual applications, for the NH3-N parameter, when the change gradient is greater than 0.1 mg / (L•min), it is determined as the rapid response interval; when the change gradient is less than 0.02 mg / (L•min), it is determined as the delayed response interval. For the DO parameter, when the change gradient is greater than 0.05 mg / (L•min), it is the rapid response interval, and when it is less than 0.01 mg / (L•min), it is the delayed response interval.

[0176] According to the distribution characteristics of the rapid response interval and the delayed response interval, the load accumulation rate in each transfer interval is calculated to generate a load response sequence. The load accumulation rate represents the accumulation degree of the load in the processing unit per unit time. In the anoxic zone-oxygen zone transfer interval, the time proportions of the rapid response interval and the delayed response interval are counted, and the parameter change amplitude is combined to calculate the load accumulation rate. For example, when the rapid response interval accounts for 40% and the delayed response interval accounts for 25%, and the average parameter change amplitude is 0.8 mg / L, the load accumulation rate of this transfer interval is 0.32 mg / (L•min).

[0177] The load accumulation rate in the load response sequence can be combined with the parameter response time to determine the timing correlation degree of the process parameter adjustment. The timing correlation degree reflects the time-dependent relationship of parameter adjustment between different processing units. On the load accumulation rate change curve, the load mutation position and the gradual change position are marked by threshold judgment. When the load accumulation rate changes by more than 50% within a short period of time (such as 15 minutes), it is marked as a mutation position; when the load accumulation rate changes slowly and the change rate is not more than 20% within a longer period of time (such as more than 60 minutes), it is marked as a gradual change position.

[0178] According to the marked load mutation position and the gradual change position, a linkage adjustment time point of the process parameters is generated. At the mutation position, it is suggested to adjust the parameters in advance before the load is transferred to the downstream unit, and the adjustment time point is the mutation position time minus 30% of the parameter response duration. At the gradual change position, it is suggested to adjust the parameters synchronously during the load transfer, and the adjustment time point is the gradual change position time plus 10% of the parameter response duration. For example, when the water inflow COD mutation is detected, it is suggested to adjust the anaerobic zone stirring intensity after 25 minutes, to adjust the anoxic zone reflux ratio after 62 minutes, and to adjust the aerobic zone aeration amount after 108 minutes.

[0179] Through the above, the cooperative operation between the treatment units of the sewage treatment process can be realized, the response capability of the system to the load impact can be improved, the energy consumption can be reduced, and the treatment efficiency can be improved.

[0180] The sewage treatment operation cost optimization system based on big data provided in the embodiment comprises:

[0181] A first unit is configured to collect real-time operation data of a sewage treatment plant, calculate and process the treatment cost according to the real-time operation data in segments according to the sewage treatment load, and obtain load cost relationship data;

[0182] A second unit is configured to map the sewage treatment load in the time dimension and the water quality change dimension based on the load cost relationship data, and establish a corresponding relationship between the sewage treatment load and the comprehensive resource consumption;

[0183] A third unit is configured to generate an operation cost curve of the sewage treatment system according to the corresponding relationship, compare the operation cost curve with a preset cost threshold curve, and obtain cost deviation data;

[0184] A fourth unit is configured to generate a cost fluctuation curve based on the cost deviation data, and determine a linkage adjustment time point of the process parameters according to the peak and valley feature points of the cost fluctuation curve;

[0185] A fifth unit is configured to generate a process parameter adjustment instruction according to the linkage adjustment time point, and send the process parameter adjustment instruction to an execution end of a sewage treatment device;

[0186] A sixth unit is configured to receive effluent water quality data returned by the execution end of the sewage treatment device, and record the process parameters corresponding to the effluent water quality data meeting the discharge standard to a sewage treatment operation database.

[0187] In a third aspect, the embodiment provides an electronic device, which comprises:

[0188] a processor;

[0189] a memory for storing processor-executable instructions;

[0190] The processor is configured to invoke instructions stored in the memory to perform the method described above.

[0191] In a fourth aspect, the present application provides a computer readable storage medium having stored thereon computer program instructions, which when executed by a processor implement the method described above.

[0192] The present application can be a method, apparatus, system, and / or computer program product. The computer program product can include a computer readable storage medium having computer readable program instructions stored therein, which when executed by a processor, perform various aspects of the present application.

[0193] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for optimizing wastewater treatment operation costs based on big data, characterized in that, include: Collect real-time operating data of the wastewater treatment plant, and calculate the treatment cost in segments according to the wastewater treatment load to obtain load-cost relationship data; Based on the load-cost relationship data, the wastewater treatment load is cross-combined and mapped in terms of time and water quality change to establish a correspondence between wastewater treatment load and comprehensive resource consumption. The operating cost curve of the wastewater treatment system is generated based on the correspondence, and the operating cost curve is compared with the preset cost threshold curve to obtain cost deviation data; A cost fluctuation curve is generated based on the cost deviation data, and the linkage adjustment time point of the process parameters is determined according to the peak and trough characteristic points of the cost fluctuation curve. Based on the aforementioned linkage adjustment time point, a process parameter adjustment instruction is generated, and the process parameter adjustment instruction is sent to the execution end of the wastewater treatment equipment. The system receives effluent water quality data returned by the execution terminal of the wastewater treatment equipment and records the process parameters corresponding to the effluent water quality data meeting the discharge standards into the wastewater treatment operation database.

2. The method according to claim 1, characterized in that, Real-time operating data of the wastewater treatment plant is collected, and the treatment cost is calculated in segments according to the wastewater treatment load to obtain load-cost relationship data, including: Collect real-time operating data of the wastewater treatment plant, extract wastewater treatment load characteristic parameters based on the real-time operating data, and construct a real-time curve of wastewater treatment load; The real-time curve of the wastewater treatment load is segmented in the time dimension to obtain the mapping relationship between time periods and load changes. Obtain the equipment operating parameters within the time period, calculate the energy consumption, reagent usage, and equipment operating time within the time period, and generate an operating cost calculation factor; The operating cost calculation factor is correlated with the time period to obtain initial cost distribution data; Based on the initial cost distribution data, the characteristic values ​​of the changes in wastewater treatment load and operating cost are extracted, and the initial cost distribution data is divided into multiple treatment intervals according to the characteristic values ​​of the changes. Calculate the average load value and average cost value within each processing interval, and then associate and pair the average load value and average cost value to obtain load-cost relationship data.

3. The method according to claim 1, characterized in that, Based on the aforementioned load-cost relationship data, the wastewater treatment load is cross-mapped across time and water quality change dimensions to establish a correspondence between wastewater treatment load and comprehensive resource consumption, including: Based on the load cost relationship data, time-scale sequence distribution data is generated, and the sequence distribution data is divided according to dynamic time windows to generate time slice data of load cost. Obtain the influent water quality monitoring data for the corresponding time period of the time slice data, and extract the water quality change parameters from the influent water quality monitoring data to obtain the water quality fluctuation characteristic curve; The water quality fluctuation characteristic curve is divided into multiple fluctuation intervals. The coordinated change characteristic value of water quality parameters in each interval is calculated. The water quality characteristics at the interval boundary are compensated based on the difference in the coordinated change characteristic values ​​of adjacent intervals to generate the water quality fluctuation characteristic curve. The time slice data and the water quality fluctuation characteristic curve are reconstructed to generate a two-dimensional feature matrix. A two-layer recursive mapping structure is determined based on the two-dimensional feature matrix. A fluctuation compensation threshold is determined in the two-layer recursive mapping structure. The two-dimensional feature matrix is ​​processed according to the fluctuation compensation threshold to generate multi-dimensional mapping data. Based on the multi-dimensional mapping data, a correspondence between wastewater treatment load and comprehensive resource consumption is established.

4. The method according to claim 3, characterized in that, A two-layer recursive mapping structure is determined based on the two-dimensional feature matrix. A fluctuation compensation threshold is determined within the two-layer recursive mapping structure. The two-dimensional feature matrix is ​​then processed according to the fluctuation compensation threshold to generate multi-dimensional mapping data, including: Based on the dual-dimensional feature matrix, the temporal fluctuation characteristics of load data and cost data are extracted, and the gradient of change of load data and cost data in different time periods is calculated. The change gradients of the different time periods are divided into different fluctuation intervals according to the fluctuation intensity, and the correlation coefficients of load data and cost data in each fluctuation interval are calculated. The correlation coefficient is used to determine the data correlation degree of each fluctuation interval, and the data correlation degree is used as a weighting coefficient to determine the two-level recursive mapping structure of the load data and the cost data. In the two-layer recursive mapping structure, a fluctuation compensation threshold is determined based on the correlation coefficient and the weight coefficient. The fluctuation compensation threshold is used to characterize the data conversion intensity between different fluctuation intervals. Based on the fluctuation compensation threshold, the load data and cost data in the dual-dimensional feature matrix are recursively mapped to different regions. At the boundary of adjacent fluctuation intervals, data smoothing transition is performed based on the weight coefficient. Based on the weight coefficient and the fluctuation compensation threshold, a mapping correction coefficient is obtained. Based on the mapping correction coefficient, the load and cost data are adjusted to different regions to generate multi-dimensional mapping data.

5. The method according to claim 1, characterized in that, Based on the aforementioned correspondence, an operating cost curve for the wastewater treatment system is generated, and this operating cost curve is compared with a preset cost threshold curve to obtain cost deviation data, including: Based on the correspondence, the operating cost data of the sewage treatment system at different time points are connected to form an operating cost curve, and the operating cost fluctuation characteristic points are determined by calculating the operating cost curve. Based on the operating cost fluctuation characteristic points, the operating cost curve is divided into different cost intervals, and the data dispersion within each cost interval is calculated. Extract cost data sequences under target operating conditions from historical operating data, connect the cost data sequences to generate a preset cost threshold curve, and set a fluctuation tolerance range for the preset cost threshold curve based on the data dispersion. An initial deviation value is obtained based on the operating cost curve and the preset cost threshold curve. The time period is divided into a stable segment and a fluctuating segment according to the initial deviation value. The overlap between the cost data in the fluctuating segment and the fluctuation tolerance range is calculated, and a fluctuating segment weight coefficient is generated. The data correlation strength between adjacent time periods is calculated based on the weighting coefficient of the fluctuation segment. The cost data within the fluctuation segment is then processed according to the data correlation strength and the initial deviation value to obtain cost deviation data.

6. The method according to claim 1, characterized in that, A cost fluctuation curve is generated based on the cost deviation data, and the timing points for the linkage adjustment of process parameters are determined according to the peak and trough characteristic points of the cost fluctuation curve, including: Determine the positive and negative deviation values ​​in the cost deviation data, and connect the positive and negative deviation values ​​in chronological order to generate a cost fluctuation curve; Extract the inflection point data from the cost fluctuation curve and calculate the slope change trend of the inflection point data. Based on the slope change trend, determine the candidate peak points and candidate trough points. Calculate the fluctuation period and fluctuation amplitude between adjacent candidate peak points and candidate trough points, and select fluctuation feature points based on the fluctuation period and fluctuation amplitude; Extract the water quality parameters and process parameters before and after the fluctuation feature point, and calculate the change gradient of the water quality parameters; The gradient of the water quality parameter change is divided into different change intervals according to the intensity, and the parameter response characteristics in different change intervals are calculated respectively. Based on the parameter response characteristics, the load transfer relationship between each treatment unit of the wastewater treatment process is extracted. The load transfer relationship is combined with the change gradient of the water quality parameters to generate a load response sequence. The linkage adjustment time point of the process parameters is determined according to the load response sequence.

7. The method according to claim 6, characterized in that, Based on the parameter response characteristics, the load transfer relationship between each treatment unit of the wastewater treatment process is extracted. This load transfer relationship is then combined with the change gradient of the water quality parameters to generate a load response sequence. Based on the load response sequence, the linkage adjustment time points of the process parameters are determined, including: Based on the parameter response characteristics, the process parameter change trend curve between adjacent processing units is determined, and the inflection point of the process parameter change trend curve is determined as the load transfer node. The initial load transfer relationship is obtained according to the time interval of the load transfer node. The initial load transfer relationship is divided into different transfer intervals according to the process flow direction of the processing unit, and the parameter response time in each transfer interval is calculated. The time period in which the gradient of the water quality parameter change is greater than the first gradient threshold is defined as the fast response interval, and the time period in which the gradient of the water quality parameter change is less than the second gradient threshold is defined as the delayed response interval. Based on the distribution characteristics of the fast response interval and the delayed response interval, the load accumulation rate within each of the transmission intervals is calculated to generate a load response sequence; By combining the load accumulation rate in the load response sequence with the parameter response duration, the temporal correlation of process parameter adjustment is determined. The load abrupt change position and the gradual change position are marked in the temporal correlation. The linkage adjustment time point of process parameters is generated based on the load abrupt change position and the gradual change position.

8. A wastewater treatment operation cost optimization system based on big data, used to implement the method as described in any one of claims 1-7, characterized in that, include: The first unit is used to collect real-time operating data of the sewage treatment plant, and to calculate the treatment cost in segments according to the sewage treatment load to obtain load-cost relationship data. The second unit is used to cross-combine and map the wastewater treatment load in the time dimension and water quality change dimension based on the load cost relationship data, and establish the correspondence between wastewater treatment load and comprehensive resource consumption. The third unit is used to generate the operating cost curve of the sewage treatment system according to the correspondence, and compare the operating cost curve with the preset cost threshold curve to obtain cost deviation data. The fourth unit is used to generate a cost fluctuation curve based on the cost deviation data, and to determine the linkage adjustment time point of the process parameters according to the peak and trough characteristic points of the cost fluctuation curve. The fifth unit is used to generate process parameter adjustment instructions based on the linkage adjustment time point, and send the process parameter adjustment instructions to the execution end of the sewage treatment equipment. The sixth unit is used to receive the effluent water quality data returned by the execution end of the wastewater treatment equipment, and record the process parameters corresponding to the effluent water quality data meeting the discharge standards into the wastewater treatment operation database.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images