A highway service area sewage treatment system based on the Internet of Things

Abstract

The invention discloses a highway service area sewage treatment system based on the Internet of Things, and relates to the technical field of sewage treatment. The invention deploys a multimodal sensor and an edge computing unit to realize real-time monitoring and data processing of sewage flow and water quality parameters, quickly identify water quality change trends and parameter change rates, and significantly improve the system's perception of load changes. The system makes instant adjustments according to water quality changes to avoid fluctuations in treatment effects and waste of energy consumption. An LSTM model is used to dynamically predict water quality parameters, and in combination with real-time monitoring data, an aeration strategy is optimized through intelligent adaptive adjustment and closed-loop refined control. The LSTM model can predict future sewage loads and changes in ammonia nitrogen concentrations, adjust aeration intensity and time in advance, and quickly respond to fluctuations in ammonia nitrogen concentrations through PID control and closed-loop control of a feedback mechanism, thereby improving ammonia nitrogen removal efficiency and aeration accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and in particular to a highway service area sewage treatment system based on the Internet of Things. Background Art

[0002] In the sewage treatment of highway service areas, the sewage sources are complex and the load changes irregularly, and there will be significant fluctuations in water volume and pollutant concentrations between peak and trough periods.

[0003] The current sewage treatment system regularly monitors water quality parameters to adjust the process, and adjusts aeration, dosing, stirring and other treatment processes according to the set thresholds. Such adjustments mainly rely on manual settings or fixed control logic, lacking the ability to respond to water quality changes in real time. The water quality monitoring frequency is low, and the adjustment response is delayed. The system cannot respond to the rapidly changing sewage load in real time, resulting in unstable treatment effects. Failure to adjust in time during peak periods will lead to overload operation, affecting the quality of effluent; energy will be wasted during low-load periods.

[0004] In addition, in order to control the ammonia nitrogen concentration in sewage, the traditional method is to adjust the activity of aerobic microorganisms by setting the dissolved oxygen concentration in the aeration tank, thereby removing ammonia nitrogen. Most of this type of control is based on preset aeration time and fixed aeration intensity. A small number of systems will use PID controllers to make adjustments based on real-time dissolved oxygen data, but there is still a problem of slow response of aeration control to changes in ammonia nitrogen concentration. When the ammonia nitrogen concentration rises or falls sharply, the adjustment of aeration intensity often lags behind, and PID control performs poorly when dealing with nonlinear and complex fluctuating ammonia nitrogen loads, and is prone to overshoot or undercontrol, resulting in low ammonia nitrogen removal efficiency. Therefore, there is an urgent need for a highway service area sewage treatment system based on the Internet of Things to solve such problems. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a highway service area sewage treatment system based on the Internet of Things to solve the problem of relying on real-time monitoring data, lacking the ability to predict future load changes, and being unable to make adjustments in advance. It can only passively respond to fluctuations in sewage load. When the load is high or low, there is no corresponding adjustment method, which easily causes energy waste.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] On the one hand, the present invention provides a highway service area sewage treatment system based on the Internet of Things, which includes:

[0009] The monitoring and data acquisition module deploys multimodal sensors in the sewage treatment system to monitor the flow and water quality parameters of sewage in real time, and outputs water quality change trends, change rates of water quality parameters, early warning signals, and characterized data sets;

[0010] The prediction and analysis module uses the long short-term memory network LSTM model to analyze the water quality data processed by the monitoring module and predict the changing trend of sewage load and ammonia nitrogen concentration in the short term in the future;

[0011] Sewage treatment module, including sedimentation tank, anaerobic tank, aerobic tank and dosing system treatment unit;

[0012] Intelligent adaptive load regulation system, which dynamically adjusts the operating parameters of each unit in the sewage treatment module, including the dissolved oxygen level in the aeration tank, the residence time in the sedimentation tank, and the dosage of the chemical according to the results provided by the prediction and analysis module;

[0013] The biological treatment and aeration control module adjusts the aeration intensity and time in real time based on the prediction results of the prediction and analysis module;

[0014] Edge computing unit, which is used for preliminary processing of data collected by sensors and closed-loop control of the aeration process to dynamically adjust the dissolved oxygen level;

[0015] Energy consumption monitoring module, which combines dynamic load forecasting and intelligent adaptive adjustment results to monitor and schedule the energy consumption of each processing unit in real time.

[0016] On the other hand, the present invention provides a highway service area sewage treatment method based on the Internet of Things, comprising:

[0017] Step S1, real-time data collection, deploying multimodal sensors at the water inlet, aeration tank and sedimentation tank to monitor sewage flow and water quality parameters in real time;

[0018] The sensor data is initially processed in the edge computing unit and water quality data is output in real time;

[0019] Step S2, sewage load and water quality parameter prediction, based on the feature data generated in step S1, the long short-term memory network LSTM model is used for analysis to predict the change trend of sewage flow and water quality parameters in the short term in the future, generate short-term prediction data of load change and water quality parameters, and form a prediction trend chart and data sequence;

[0020] The forecast results are used to identify peak loads or sudden load changes in advance;

[0021] Step S3, intelligent adaptive process adjustment, combining the real-time data of step S1 and the predicted data of step S2, the intelligent adaptive adjustment module dynamically adjusts the treatment process according to the current and predicted sewage load;

[0022] Step S4, closed-loop aeration control, uses edge computing units to combine real-time and predicted water quality data to implement closed-loop control of the aeration process.

[0023] Dynamically adjust aeration time and intensity, and adjust aeration strategies according to the changing trends of dissolved oxygen levels and ammonia nitrogen concentrations to improve removal efficiency and avoid control lag;

[0024] Step S5, energy consumption optimization and intelligent scheduling, combines the load prediction of step S2 and the adjustment result of step S3, monitors the energy consumption of each processing unit in real time, identifies high and low load operating states, and performs intelligent scheduling.

[0025] Further, in step S1, the water quality parameters include pH, COD, ammonia nitrogen concentration and dissolved oxygen;

[0026] The real-time output water quality data includes characteristic data sets of water quality change trends, parameter change rates, and early warning signals.

[0027] Furthermore, in step S1, the edge computing unit performs preliminary processing on the data collected by the sensor:

[0028] De-noise and smooth the raw data collected by the sensor to remove abnormal points.

[0029] in, represents the smoothed data at time t, x i (t) represents the original data value at time t, which is composed of the water quality parameters collected by the i-th sensor, and α represents the smoothing coefficient, ranging from 0<α<1;

[0030] Calculate the changing trend of water quality parameters and use the weighted moving average method to extract the changing direction of the data:

[0031] Among them, T i (t) represents the water quality change trend at time t, N represents the window length, w k Represents the weight of the kth time step, set as a decreasing series;

[0032] Calculate the rate of change of water quality parameters, characterize the speed of change of water quality parameters over time, and use it to monitor mutations in real time. Among them, R i (t) represents the parameter change rate at time t, Δt represents the time interval, that is, the time difference between adjacent sampling points;

[0033] Generates an early warning signal. When the parameter change rate exceeds the set threshold, the early warning is triggered.

[0034] If |Ri (t)|>θ i , then E i (t)=1, otherwise E i (t) = 0, where E i (t) represents the warning signal at time t, 1 means triggering the warning, 0 means normal, θ i Indicates the warning threshold.

[0035] Furthermore, in step S2, the long short-term memory network LSTM model is used to analyze the characteristic data generated in step S1 to predict the change trend of sewage flow and water quality parameters in the short term in the future. Specifically:

[0036] The input data of the LSTM model is the feature data set processed in step S1, including the water quality change trend T i (t), parameter change rate R i (t) and warning signal E i (t), normalize the input data,

[0037] Among them, X i (t) represents the normalized data of the i-th parameter at time t, represents the smoothed data after processing in step S1, Indicates the minimum and maximum values ​​of the data;

[0038] The calculation process of LSTM unit:

[0039] Forget gate: f t =σ(W f ·[h t-1 ,X i (t)]+b f ), where f t represents the activation value of the forget gate, indicating whether to retain the memory of the previous moment, σ represents the Sigmoid function, mapping the output to the [0,1] interval, W f represents the weight matrix of the forget gate, h t-1 represents the hidden state at the previous moment, b f Represents the bias term of the forget gate;

[0040] Input gate: i t =σ(W i ·[h t-1 ,X i (t)]+b i ), where i t represents the activation value of the input gate, W i represents the weight matrix of the input gate, b i represents the bias term of the input gate;

[0041] Candidate memory: in, represents the candidate memory state, W C represents the weight matrix of candidate memory, b C The bias term representing the candidate memory;

[0042] Update memory status: Among them, C t Indicates the memory state at the current moment, C t-1 Indicates the memory state at the previous moment;

[0043] Output gate: o t =σ(W o ·[h t-1 ,X i (t)]+b o ), where o t represents the activation value of the output gate, W o represents the weight matrix of the output gate, b o Represents the bias term of the output gate;

[0044] Hide status update:h t =o t tanh(C t ), where h t Represents the hidden state at the current moment, that is, the model output;

[0045] Based on the hidden state h calculated by LSTM t , input to the fully connected layer, and output the prediction results of sewage flow and water quality parameters in the future time period,

[0046] in, Represents the predicted value at the future time t+Δt, including sewage flow and water quality parameters, W y represents the weight matrix of the fully connected layer, b y Represents the bias term of the fully connected layer; the LSTM model processes real-time water quality data, captures the changing rules of time series, generates short-term forecast data of sewage flow and water quality parameters, and forms a forecast trend chart and data sequence.

[0047] Furthermore, the adjustment measures in step S3 include:

[0048] Adjust the dissolved oxygen level in the aeration tank, increase or decrease the aeration intensity,

[0049] Dynamically adjust the residence time of the sedimentation tank,

[0050] Control the dosage according to water quality requirements.

[0051] Furthermore, in step S3, the intelligent adaptive process adjustment method is:

[0052] According to the real-time and predicted water quality data, the dissolved oxygen level and aeration intensity in the aeration tank are dynamically adjusted. The dissolved oxygen level adjustment formula is: Where DO(t) represents the dissolved oxygen concentration at time t, DO set Indicates the set reference dissolved oxygen concentration, k1 indicates the dissolved oxygen adjustment coefficient, which is set according to the system response speed. represents the predicted ammonia nitrogen concentration at time t, Indicates the ammonia nitrogen concentration at time t of real-time monitoring;

[0053] The aeration intensity adjustment formula is: A(t) = A base +k2·(DO set -DO(t)), where A(t) represents the aeration intensity at time t, A base represents the basic aeration intensity, i.e. the minimum aeration requirement of the system, and k2 represents the aeration intensity adjustment coefficient;

[0054] The retention time of the sedimentation tank is adjusted according to the real-time sewage flow and predicted load changes to optimize the solid-liquid separation effect. The retention time adjustment formula is: Among them, T s (t) represents the residence time of the sedimentation tank at time t, V s represents the effective volume of the sedimentation tank, Q(t) represents the real-time sewage flow at time t, represents the predicted sewage flow at time t, k3 represents the residence time adjustment coefficient, reflecting the impact of flow fluctuation on residence time;

[0055] According to the real-time status and prediction results of water quality, the dosage of chemical agents is dynamically adjusted to ensure that the effluent meets the standard and avoid waste of agents. The dosage control formula is: D(t) = D set +k4·(C target -C(t)), where D(t) represents the dosage at time t, D set Indicates the set basic dosage, k4 indicates the dosage adjustment coefficient, which is set according to the sensitivity of water quality fluctuation, C target represents the target water quality parameter concentration, and C(t) represents the real-time water quality parameter concentration at time t.

[0056] Furthermore, in step S4, the closed-loop aeration control method is:

[0057] A real-time feedback control mechanism is used to make instant adjustments based on the current dissolved oxygen level and ammonia nitrogen concentration. The dissolved oxygen control formula is:

[0058]

[0059] Among them, u DO (t) represents the dissolved oxygen regulation signal at time t, K p Represents the proportional gain, which controls the direct impact of dissolved oxygen error on the regulation signal, K i Indicates the integral gain, K d represents the differential gain, which is the speed of response to the change of dissolved oxygen. dτ represents the small increment of the time variable τ. set represents the target dissolved oxygen level, which is set according to the optimal microbial activity, and DO(t) represents the actual dissolved oxygen concentration at time t, which is obtained from real-time monitoring;

[0060] Combined with the predicted trend of ammonia nitrogen concentration, the aeration time and intensity are dynamically adjusted. The aeration intensity adjustment formula is:

[0061] Among them, A(t) represents the aeration intensity at time t, which affects the oxygen supply, A base It represents the basic aeration intensity, i.e. the system's default minimum oxygen demand. K1 represents the ammonia nitrogen adjustment coefficient, which adjusts the effect of ammonia nitrogen concentration deviation on aeration intensity. represents the predicted ammonia nitrogen concentration at time t, from LSTM prediction, It represents the actual ammonia nitrogen concentration at time t, the real-time monitoring value, and k2 represents the ammonia nitrogen change rate coefficient, reflecting the dynamic effect of the ammonia nitrogen change rate on the aeration intensity;

[0062] Dynamically optimize aeration time, adjust aeration cycle according to current and predicted water quality trends, avoid over-aeration or under-aeration, and improve system efficiency. Aeration time adjustment formula:

[0063] Among them, T air (t) represents the aeration time length at time t, T base It represents the basic aeration time, i.e. the minimum aeration time period set by the system. K3 represents the cumulative ammonia nitrogen error influence coefficient, which adjusts the influence of the total error on the aeration time. K4 represents the dissolved oxygen adjustment signal influence coefficient, which reflects the influence of aeration intensity on the time length. Indicates the cumulative deviation of ammonia nitrogen concentration;

[0064] The edge computing unit performs closed-loop optimization between real-time feedback control and predictive adjustment. The dissolved oxygen regulation control uses the proportional integral differential PID algorithm to achieve rapid response to real-time oxygen demand, enabling the system to operate stably under different load conditions and avoiding waste of resources and excessive pollution.

[0065] Furthermore, the intelligent scheduling method in step S5 is:

[0066] When the load is low, reduce the equipment's operating power or shut down some equipment; when the load is high, adjust the working mode of each processing unit to balance operating efficiency and energy consumption.

[0067] Furthermore, in step S5, the energy consumption of each processing unit is monitored in real time in combination with the load forecast of step S2 and the adjustment result of step S3:

[0068] To monitor energy consumption in real time, the edge computing unit collects the voltage and current of each processing unit in real time and calculates the current energy consumption level. The energy consumption calculation formula is:

[0069] P i (t) = V i (t)×I i (t)×η i (t), where P i (t) represents the energy consumption of the i-th processing unit at time t, V i (t) represents the real-time voltage of the i-th processing unit at time t, I i (t) represents the real-time current of the i-th processing unit at time t, η i (t) represents the efficiency factor of the i-th processing unit at time t, reflecting the energy efficiency status of the equipment;

[0070] According to the load forecast in step S2 and the current real-time flow data, the current load state factor of the system is calculated to determine whether it is currently in a high load or low load state. The load state factor is:

[0071] Among them, L(t) represents the load state factor at time t, reflecting the system load status. represents the predicted sewage flow at time t, provided by step S2, Q max represents the maximum processing flow of the system design, α represents the weight factor, balancing the impact of flow and energy consumption in the load state, P total (t) represents the total energy consumption at time t, the sum of the energy consumption of all processing units, P max Indicates the maximum energy consumption level of the system design;

[0072] According to the load state factor L(t), an intelligent dispatching strategy is implemented. When the load is low, the equipment operating power is adjusted or some equipment is shut down; when the load is high, the operating mode of each unit is adjusted to optimize energy consumption and efficiency. The dispatching adjustment method is:

[0073] Low load state L(t)<θ low :

[0074] Among them, P i adj(t) represents the adjusted energy consumption of the i-th processing unit at time t, k1 represents the adjustment coefficient, which determines the extent of energy consumption reduction, θ low Indicates the low load threshold, the minimum load benchmark for system operation;

[0075] High load state L(t)>θ high :

[0076] Among them, k2 represents the high load adjustment coefficient, which determines the increase in energy consumption, θ high Indicates the high load threshold, the benchmark point for efficient operation of the system, L max It represents the maximum load state factor of the system, indicating full load operation;

[0077] Under high load, each processing unit is optimized and deployed, and the optimization formula is:

[0078] Among them, E eff (t) represents the system energy efficiency at time t, N represents the number of processing units, Q i (t) represents the treatment flow of the i-th treatment unit at time t, ∈ represents a small constant. By adjusting the equipment operation status and optimizing the scheduling mode, intelligent energy consumption management of the sewage treatment system can be achieved.

[0079] The beneficial effects of the present invention are:

[0080] The present invention deploys multimodal sensors and edge computing units to realize real-time monitoring and data processing of sewage flow and water quality parameters, quickly identify water quality change trends and parameter change rates, and significantly improve the system's ability to perceive load changes. The system makes instant adjustments based on water quality changes to avoid fluctuations in treatment effects and waste of energy.

[0081] The present invention adopts the LSTM model to dynamically predict water quality parameters, combines real-time monitoring data, and optimizes the aeration strategy through intelligent adaptive adjustment and closed-loop refined control. The LSTM model can predict future sewage load and ammonia nitrogen concentration changes, adjust the aeration intensity and time in advance, and quickly respond to fluctuations in ammonia nitrogen concentration through PID control and closed-loop control of the feedback mechanism, thereby improving the ammonia nitrogen removal efficiency and the accuracy of aeration.

[0082] The present invention uses an intelligent adaptive regulation module to dynamically adjust the treatment process according to real-time and predicted water quality data, including the control of the dissolved oxygen level in the aeration tank, the residence time in the sedimentation tank and the amount of added chemicals. When the load is high, the operating efficiency of the treatment unit is improved, and when the load is low, the equipment power is reduced or part of the equipment is shut down, thereby significantly reducing energy consumption.

[0083] The present invention adopts aeration and dosing control strategies to enable the treatment process to flexibly adapt to changes in sewage load and ensure the stability of effluent water quality. Through real-time feedback and predictive adjustment, the dosage of chemical agents is controlled according to water quality requirements, thus avoiding agent waste, improving the sewage treatment effect and significantly reducing the cost of using chemicals. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0085] Figure 1 This is a schematic diagram of the structure of the highway service area sewage treatment system based on the Internet of Things of the present invention;

[0086] Figure 2 The figure is a schematic flow chart of the highway service area sewage treatment method based on the Internet of Things of the present invention. DETAILED DESCRIPTION

[0087] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0088] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0089] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0090] Example 1, reference Figure 1 , this embodiment provides a highway service area sewage treatment system based on the Internet of Things, including:

[0091] The monitoring and data acquisition module deploys multimodal sensors in the sewage treatment system to monitor the flow and water quality parameters of sewage in real time, and outputs water quality change trends, change rates of water quality parameters, early warning signals, and characterized data sets;

[0092] The prediction and analysis module uses the long short-term memory network LSTM model to analyze the water quality data processed by the monitoring module and predict the changing trend of sewage load and ammonia nitrogen concentration in the short term in the future;

[0093] Sewage treatment module, including sedimentation tank, anaerobic tank, aerobic tank and dosing system treatment unit;

[0094] Intelligent adaptive load regulation system, which dynamically adjusts the operating parameters of each unit in the sewage treatment module, including the dissolved oxygen level in the aeration tank, the residence time in the sedimentation tank, and the dosage of the chemical according to the results provided by the prediction and analysis module;

[0095] The biological treatment and aeration control module adjusts the aeration intensity and time in real time based on the prediction results of the prediction and analysis module;

[0096] Edge computing unit, which is used for preliminary processing of data collected by sensors and closed-loop control of the aeration process to dynamically adjust the dissolved oxygen level;

[0097] Energy consumption monitoring module, which combines dynamic load forecasting and intelligent adaptive adjustment results to monitor and schedule the energy consumption of each processing unit in real time.

[0098] Example 2, reference Figure 2 , this embodiment provides a highway service area sewage treatment method based on the Internet of Things, comprising the following steps:

[0099] Step S1, real-time data collection,

[0100] Deploy multimodal sensors at water inlets, aeration tanks, and sedimentation tanks to monitor sewage flow and water quality parameters in real time;

[0101] The sensor data is initially processed in the edge computing unit and water quality data is output in real time;

[0102] Specifically, by deploying multimodal sensors at water inlets, aeration tanks and sedimentation tanks, sewage flow and key water quality parameters (such as pH, COD, ammonia nitrogen concentration and dissolved oxygen) are monitored in real time, and edge computing units are used for data processing. Edge computing significantly reduces the delay in data processing, can quickly denoise, smooth data, extract change trends and parameter change rates, and greatly improve the ability to perceive dynamic changes in water quality. It solves the problems of low frequency and delayed response of water quality monitoring in traditional methods, and ensures that the system can quickly respond to load changes and sudden changes in water quality.

[0103] Water quality parameters include pH, COD, ammonia nitrogen concentration, and dissolved oxygen;

[0104] The real-time output of water quality data includes water quality change trends, parameter change rates, and characteristic data sets of early warning signals;

[0105] In step S1, the edge computing unit performs preliminary processing on the data collected by the sensor:

[0106] De-noise and smooth the raw data collected by the sensor to remove abnormal points.

[0107] in, represents the smoothed data at time t, x i (t) represents the original data value at time t, which is composed of the water quality parameters collected by the i-th sensor, and α represents the smoothing coefficient, ranging from 0<α<1;

[0108] Calculate the changing trend of water quality parameters and use the weighted moving average method to extract the changing direction of the data:

[0109] Among them, T i (t) represents the water quality change trend at time t, N represents the window length, w k Represents the weight of the kth time step, set as a decreasing series;

[0110] Calculate the rate of change of water quality parameters, characterize the speed of change of water quality parameters over time, and use it to monitor mutations in real time. Among them, R i (t) represents the parameter change rate at time t, Δt represents the time interval, that is, the time difference between adjacent sampling points;

[0111] Generates an early warning signal. When the parameter change rate exceeds the set threshold, the early warning is triggered.

[0112] If |R i (t)|>θ i , then E i (t)=1, otherwise E i (t) = 0, where E i (t) represents the warning signal at time t, 1 means triggering the warning, 0 means normal, θ i Indicates the warning threshold;

[0113] Step S2, sewage load and water quality parameter prediction,

[0114] Based on the characteristic data generated in step S1, a long short-term memory network LSTM model is used for analysis to predict the change trend of sewage flow and water quality parameters in the short term in the future, generate short-term prediction data of load change and water quality parameters, and form a prediction trend chart and data sequence;

[0115] The forecast results are used to identify peak loads or sudden load changes in advance;

[0116] Specifically, the LSTM model is introduced to perform short-term predictions on sewage load and water quality parameters, and to identify peak loads and sudden load changes in advance. The LSTM model can accurately predict future trends in sewage flow and water quality parameters by virtue of its excellent ability to capture time series data, enabling the system to make process adjustments in advance when the load is about to change, thereby avoiding the problem of passive adjustment lag in traditional systems when the load changes, and significantly improving the stability of the system and the quality of the effluent.

[0117] In step S2, the long short-term memory network LSTM model is used to analyze the feature data generated in step S1 to predict the trend of sewage flow and water quality parameters in the short term in the future. Specifically:

[0118] The input data of the LSTM model is the feature data set processed in step S1, including the water quality change trend T i (t), parameter change rate R i (t) and warning signal E i (t), normalize the input data,

[0119] Among them, X i (t) represents the normalized data of the i-th parameter at time t, represents the smoothed data after processing in step S1, Indicates the minimum and maximum values ​​of the data;

[0120] The calculation process of LSTM unit:

[0121] Forget gate: f t =σ(W f ·[h t-1 ,X i (t)]+b f ), where f t represents the activation value of the forget gate, indicating whether to retain the memory of the previous moment, σ represents the Sigmoid function, mapping the output to the [0,1] interval, W f represents the weight matrix of the forget gate, h t-1 represents the hidden state at the previous moment, b f Represents the bias term of the forget gate;

[0122] Input gate: i t =σ(W i ·[h t-1 ,X i (t)]+b i ), where i t represents the activation value of the input gate, W i represents the weight matrix of the input gate, b i represents the bias term of the input gate;

[0123] Candidate memory: in, represents the candidate memory state, W C represents the weight matrix of candidate memory, b C The bias term representing the candidate memory;

[0124] Update memory status: Among them, C t Indicates the memory state at the current moment, C t-1 Indicates the memory state at the previous moment;

[0125] Output gate: o t =σ(W o ·[h t-1 ,X i (t)]+b o ), where o t represents the activation value of the output gate, W o represents the weight matrix of the output gate, b o Represents the bias term of the output gate;

[0126] Hide status update:h t =o t tanh(C t ), where h t Represents the hidden state at the current moment, that is, the model output;

[0127] Based on the hidden state h calculated by LSTM t , input to the fully connected layer, and output the prediction results of sewage flow and water quality parameters in the future time period,

[0128] in, Represents the predicted value at the future time t+Δt, including sewage flow and water quality parameters, W y represents the weight matrix of the fully connected layer, b y Represents the bias term of the fully connected layer; the LSTM model processes real-time water quality data, captures the time series change rules, generates short-term forecast data of sewage flow and water quality parameters, and forms a forecast trend chart and data sequence;

[0129] Step S3, intelligent adaptive process adjustment,

[0130] Combining the real-time data of step S1 and the predicted data of step S2, the intelligent adaptive adjustment module dynamically adjusts the treatment process according to the current and predicted sewage loads;

[0131] The adjustment measures in step S3 include:

[0132] Adjust the dissolved oxygen level in the aeration tank, increase or decrease the aeration intensity,

[0133] Dynamically adjust the residence time of the sedimentation tank,

[0134] Control the dosage according to water quality requirements;

[0135] Specifically, the intelligent adaptive regulation module dynamically adjusts the sewage treatment process according to real-time monitoring data and prediction results, including adjusting the dissolved oxygen level and aeration intensity of the aeration tank, optimizing the residence time of the sedimentation tank, and accurately controlling the dosage of chemical agents, so that the system can maintain the optimal treatment state under different load conditions, effectively improving the ammonia nitrogen removal efficiency and solid-liquid separation effect.

[0136] In step S3, the intelligent adaptive process adjustment method is:

[0137] According to the real-time and predicted water quality data, the dissolved oxygen level and aeration intensity in the aeration tank are dynamically adjusted. The dissolved oxygen level adjustment formula is: Where DO(t) represents the dissolved oxygen concentration at time t, DO set Indicates the set reference dissolved oxygen concentration, k1 indicates the dissolved oxygen adjustment coefficient, which is set according to the system response speed. represents the predicted ammonia nitrogen concentration at time t, Indicates the ammonia nitrogen concentration at time t of real-time monitoring;

[0138] The aeration intensity adjustment formula is: A(t) = A base +k2·(DO set -DO(t)), where A(t) represents the aeration intensity at time t, A base represents the basic aeration intensity, i.e. the minimum aeration requirement of the system, and k2 represents the aeration intensity adjustment coefficient;

[0139] The retention time of the sedimentation tank is adjusted according to the real-time sewage flow and predicted load changes to optimize the solid-liquid separation effect. The retention time adjustment formula is: Among them, T s (t) represents the residence time of the sedimentation tank at time t, V s represents the effective volume of the sedimentation tank, Q(t) represents the real-time sewage flow at time t, represents the predicted sewage flow at time t, k3 represents the residence time adjustment coefficient, reflecting the impact of flow fluctuation on residence time;

[0140] According to the real-time status and prediction results of water quality, the dosage of chemical agents is dynamically adjusted to ensure that the effluent meets the standard and avoid waste of agents. The dosage control formula is: D(t) = D set +k4·(C target-C(t)), where D(t) represents the dosage at time t, D set Indicates the set basic dosage, k4 indicates the dosage adjustment coefficient, which is set according to the sensitivity of water quality fluctuation, C target represents the target water quality parameter concentration, C(t) represents the real-time water quality parameter concentration at time t;

[0141] Step S4, closed loop aeration control,

[0142] Using edge computing units, combined with real-time and predicted water quality data, closed-loop control of the aeration process is implemented.

[0143] Dynamically adjust aeration time and intensity, and adjust aeration strategies according to the changing trends of dissolved oxygen levels and ammonia nitrogen concentrations to improve removal efficiency and avoid control lag;

[0144] In step S4, the closed-loop aeration control method is:

[0145] A real-time feedback control mechanism is used to make instant adjustments based on the current dissolved oxygen level and ammonia nitrogen concentration. The dissolved oxygen control formula is:

[0146]

[0147] Among them, u DO (t) represents the dissolved oxygen regulation signal at time t, K p Represents the proportional gain, which controls the direct impact of dissolved oxygen error on the regulation signal, K i Indicates the integral gain, K d represents the differential gain, which is the speed of response to the change of dissolved oxygen. dτ represents the small increment of the time variable τ. set represents the target dissolved oxygen level, which is set according to the optimal microbial activity, and DO(t) represents the actual dissolved oxygen concentration at time t, which is obtained from real-time monitoring;

[0148] Combined with the predicted trend of ammonia nitrogen concentration, the aeration time and intensity are dynamically adjusted. The aeration intensity adjustment formula is:

[0149] Among them, A(t) represents the aeration intensity at time t, which affects the oxygen supply, A base It represents the basic aeration intensity, i.e. the system's default minimum oxygen demand. K1 represents the ammonia nitrogen adjustment coefficient, which adjusts the effect of ammonia nitrogen concentration deviation on aeration intensity. represents the predicted ammonia nitrogen concentration at time t, from LSTM prediction, It represents the actual ammonia nitrogen concentration at time t, the real-time monitoring value, and k2 represents the ammonia nitrogen change rate coefficient, reflecting the dynamic effect of the ammonia nitrogen change rate on the aeration intensity;

[0150] Dynamically optimize aeration time, adjust aeration cycle according to current and predicted water quality trends, avoid over-aeration or under-aeration, and improve system efficiency. Aeration time adjustment formula:

[0151] Among them, T air (t) represents the aeration time length at time t, T base It represents the basic aeration time, i.e. the minimum aeration time period set by the system. K3 represents the cumulative ammonia nitrogen error influence coefficient, which adjusts the influence of the total error on the aeration time. K4 represents the dissolved oxygen adjustment signal influence coefficient, which reflects the influence of aeration intensity on the time length. Indicates the cumulative deviation of ammonia nitrogen concentration;

[0152] The edge computing unit performs closed-loop optimization between real-time feedback control and predictive adjustment. The dissolved oxygen regulation control uses the proportional integral differential PID algorithm to achieve rapid response to real-time oxygen demand, enabling the system to operate stably under different load conditions, avoiding resource waste and excessive pollution.

[0153] Specifically, the closed-loop refined aeration control comprehensively optimizes real-time feedback and predictive adjustments through the edge computing unit, combines the current dissolved oxygen level and ammonia nitrogen concentration, and adjusts the aeration strategy in real time through the proportional integral differential PID control algorithm to achieve a rapid response to dissolved oxygen demand, dynamically optimize aeration time and intensity, avoid over-aeration or under-aeration, improve ammonia nitrogen removal efficiency, and significantly reduce energy waste.

[0154] Step S5: energy consumption optimization and intelligent scheduling.

[0155] Combined with the load forecast of step S2 and the adjustment result of step S3, the energy consumption of each processing unit is monitored in real time, the high and low load operating states are identified, and intelligent scheduling is performed;

[0156] The intelligent scheduling method in step S5 is:

[0157] When the load is low, reduce the equipment operating power or shut down some equipment; when the load is high, adjust the working mode of each processing unit to balance the operating efficiency and energy consumption;

[0158] In step S5, the energy consumption of each processing unit is monitored in real time in combination with the load forecast of step S2 and the adjustment result of step S3:

[0159] To monitor energy consumption in real time, the edge computing unit collects the voltage and current of each processing unit in real time and calculates the current energy consumption level. The energy consumption calculation formula is:

[0160] P i (t) = V i (t)×I i (t)×ηi (t), where P i (t) represents the energy consumption of the i-th processing unit at time t, V i (t) represents the real-time voltage of the i-th processing unit at time t, I i (t) represents the real-time current of the i-th processing unit at time t, η i (t) represents the efficiency factor of the i-th processing unit at time t, reflecting the energy efficiency status of the equipment;

[0161] According to the load forecast in step S2 and the current real-time flow data, the current load state factor of the system is calculated to determine whether it is currently in a high load or low load state. The load state factor is:

[0162] Among them, L(t) represents the load state factor at time t, reflecting the system load status. represents the predicted sewage flow at time t, provided by step S2, Q max represents the maximum processing flow of the system design, α represents the weight factor, balancing the impact of flow and energy consumption in the load state, P total (t) represents the total energy consumption at time t, the sum of the energy consumption of all processing units, P max Indicates the maximum energy consumption level of the system design;

[0163] According to the load state factor L(t), an intelligent dispatching strategy is implemented. When the load is low, the equipment operating power is adjusted or some equipment is shut down; when the load is high, the operating mode of each unit is adjusted to optimize energy consumption and efficiency. The dispatching adjustment method is:

[0164] Low load state L(t)<θ low :

[0165] Among them, P i adj (t) represents the adjusted energy consumption of the i-th processing unit at time t, k1 represents the adjustment coefficient, which determines the extent of energy consumption reduction, θ low Indicates the low load threshold, the minimum load benchmark for system operation;

[0166] High load state L(t)>θ high :

[0167] Among them, k2 represents the high load adjustment coefficient, which determines the increase in energy consumption, θ high Indicates the high load threshold, the benchmark point for efficient operation of the system, L max It represents the maximum load state factor of the system, indicating full load operation;

[0168] Under high load, each processing unit is optimized and deployed, and the optimization formula is:

[0169] Among them, E eff (t) represents the system energy efficiency at time t, N represents the number of processing units, Q i (t) represents the treatment flow of the i-th treatment unit at time t, ∈ represents a small constant, and the intelligent energy consumption management of the sewage treatment system is realized by adjusting the equipment operation status and optimizing the scheduling mode;

[0170] Specifically, in terms of energy consumption optimization and intelligent scheduling, the energy consumption of each processing unit is monitored in real time, and the high and low load states are identified in combination with load forecast data to achieve fine energy consumption management. When the load is low, the operating power of the equipment can be automatically reduced or some equipment can be shut down to avoid ineffective energy consumption. When the load is high, the intelligent scheduling module reasonably allocates the working mode of each processing unit, reducing overall energy consumption and improving economy.

[0171] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A highway service area sewage treatment system based on the Internet of Things, characterized by: include, The monitoring and data acquisition module deploys multimodal sensors in the sewage treatment system to monitor the flow and water quality parameters of sewage in real time, and outputs water quality change trends, change rates of water quality parameters, early warning signals, and characterized data sets; The prediction and analysis module uses the long short-term memory network LSTM model to analyze the water quality data processed by the monitoring module and predict the changing trend of sewage load and ammonia nitrogen concentration in the short term in the future; Sewage treatment module, including sedimentation tank, anaerobic tank, aerobic tank and dosing system treatment unit; Intelligent adaptive load regulation system, which dynamically adjusts the operating parameters of each unit in the sewage treatment module, including the dissolved oxygen level in the aeration tank, the residence time in the sedimentation tank, and the dosage of the chemical according to the results provided by the prediction and analysis module; The operating parameters are adjusted as follows: According to the real-time and predicted water quality data, the dissolved oxygen level and aeration intensity in the aeration tank are dynamically adjusted. The dissolved oxygen level adjustment formula is: Where DO(t) represents the dissolved oxygen concentration at time t, DO set Indicates the set reference dissolved oxygen concentration, k1 indicates the dissolved oxygen adjustment coefficient, which is set according to the system response speed. represents the predicted ammonia nitrogen concentration at time t, Indicates the ammonia nitrogen concentration at time t of real-time monitoring; The aeration intensity adjustment formula is: A(t) = A base +k2·(DO set -DO(t)), where A(t) represents the aeration intensity at time t, A base represents the basic aeration intensity, i.e. the minimum aeration requirement of the system, and k2 represents the aeration intensity adjustment coefficient; The retention time of the sedimentation tank is adjusted according to the real-time sewage flow and the predicted load change. The retention time adjustment formula is: Among them, T s (t) represents the residence time of the sedimentation tank at time t, V s represents the effective volume of the sedimentation tank, Q(t) represents the real-time sewage flow at time t, represents the predicted sewage flow at time t, k3 represents the residence time adjustment coefficient, reflecting the impact of flow fluctuation on residence time; According to the real-time status and prediction results of water quality, the dosage of chemical agents is dynamically adjusted. The dosage control formula is: D(t) = D set +k4·(C target -C(t)), where D(t) represents the dosage at time t, D set Indicates the set basic dosage, k4 indicates the dosage adjustment coefficient, which is set according to the sensitivity of water quality fluctuation, C target represents the target water quality parameter concentration, C(t) represents the real-time water quality parameter concentration at time t; The biological treatment and aeration control module adjusts the aeration intensity and time in real time based on the prediction results of the prediction and analysis module; Edge computing unit, which is used for preliminary processing of data collected by sensors and closed-loop control of the aeration process to dynamically adjust the dissolved oxygen level; Energy consumption monitoring module, which combines dynamic load forecasting and intelligent adaptive adjustment results to monitor and schedule the energy consumption of each processing unit in real time.

2. A highway service area sewage treatment method based on the Internet of Things, characterized in that: Based on the highway service area sewage treatment system based on the Internet of Things as described in claim 1, The following steps are included: Step S1, real-time data collection, Deploy multimodal sensors at water inlets, aeration tanks, and sedimentation tanks to monitor sewage flow and water quality parameters in real time; The sensor data is initially processed in the edge computing unit and water quality data is output in real time; Step S2, sewage load and water quality parameter prediction, Based on the characteristic data generated in step S1, a long short-term memory network LSTM model is used for analysis to predict the change trend of sewage flow and water quality parameters in the short term in the future, generate short-term prediction data of load change and water quality parameters, and form a prediction trend chart and data sequence; Step S3, intelligent adaptive process adjustment, Combining the real-time data of step S1 and the predicted data of step S2, dynamically adjusting the treatment process according to the current and predicted sewage loads; Step S4, closed loop aeration control, Using edge computing units, combined with real-time and predicted water quality data, closed-loop control of the aeration process is implemented. Dynamically adjust aeration time and intensity, and adjust aeration strategy according to the changing trends of dissolved oxygen levels and ammonia nitrogen concentrations; Step S5: energy consumption optimization and intelligent scheduling. Combined with the load prediction of step S2 and the adjustment result of step S3, the energy consumption of each processing unit is monitored in real time, the high and low load operating states are identified, and intelligent scheduling is performed.

3. A highway service area sewage treatment method based on the Internet of Things according to claim 2, characterized in that: In step S1, water quality parameters include pH, COD, ammonia nitrogen concentration and dissolved oxygen; The real-time output water quality data includes characteristic data sets of water quality change trends, parameter change rates, and early warning signals.

4. A highway service area sewage treatment method based on the Internet of Things according to claim 3, characterized in that: In step S1, the edge computing unit performs preliminary processing on the data collected by the sensor: De-noise and smooth the raw data collected by the sensor to remove abnormal points. in, represents the smoothed data at time t, x i (t) represents the original data value at time t, which is composed of the water quality parameters collected by the i-th sensor, and α represents the smoothing coefficient, ranging from 0<α<1; Calculate the changing trend of water quality parameters and use the weighted moving average method to extract the changing direction of the data: Among them, T i (t) represents the water quality change trend at time t, N represents the window length, w k Represents the weight of the kth time step, set as a decreasing series; Calculate the rate of change of water quality parameters, characterize the speed of change of water quality parameters over time, and use it to monitor mutations in real time. Among them, R i (t) represents the parameter change rate at time t, Δt represents the time interval, that is, the time difference between adjacent sampling points; Generates an early warning signal. When the parameter change rate exceeds the set threshold, the early warning is triggered. If |R i (t)|>θ i , then E i (t)=1, otherwise E i (t) = 0, where E i (t) represents the warning signal at time t, 1 means triggering the warning, 0 means normal, θ i Indicates the warning threshold.

5. A highway service area sewage treatment method based on the Internet of Things according to claim 4, characterized in that: In step S2, the long short-term memory network LSTM model is used to analyze the feature data generated in step S1 to predict the trend of sewage flow and water quality parameters in the short term in the future. Specifically: The input data of the LSTM model is the feature data set processed in step S1, including the water quality change trend T i (t), parameter change rate R i (t) and warning signal E i (t), normalize the input data, Among them, X i (t) represents the normalized data of the i-th parameter at time t, represents the smoothed data after processing in step S1, Indicates the minimum and maximum values ​​of the data; The calculation process of LSTM unit: Forget gate: f t =σ(W f ·[h t-1 ,X i (t)]+b f ), where f t represents the activation value of the forget gate, indicating whether to retain the memory of the previous moment, σ represents the Sigmoid function, mapping the output to the [0,1] interval, W f represents the weight matrix of the forget gate, h t-1 represents the hidden state at the previous moment, b f Represents the bias term of the forget gate; Input gate: i t =σ(W i ·[h t-1 ,X i (t)]+b i ), where i t represents the activation value of the input gate, W i represents the weight matrix of the input gate, b i represents the bias term of the input gate; Candidate memory: in, represents the candidate memory state, W C represents the weight matrix of candidate memory, b C The bias term representing the candidate memory; Update memory status: Among them, C t Indicates the memory state at the current moment, C t-1 Indicates the memory state at the previous moment; Output gate: o t =σ(W o ·[h t-1 ,X i (t)]+b o ), where o t represents the activation value of the output gate, W o represents the weight matrix of the output gate, b o Represents the bias term of the output gate; Hide status update:h t =o t tanh(C t ), where h t Represents the hidden state at the current moment, that is, the model output; Based on the hidden state h calculated by LSTM t , input to the fully connected layer, and output the prediction results of sewage flow and water quality parameters in the future time period, in, Represents the predicted value at the future time t+Δt, including sewage flow and water quality parameters, W y represents the weight matrix of the fully connected layer, b y Represents the bias term of the fully connected layer.

6. A highway service area sewage treatment method based on the Internet of Things according to claim 5, characterized in that: The adjustment measures in step S3 include: Adjust the dissolved oxygen level in the aeration tank, increase or decrease the aeration intensity, Dynamically adjust the residence time of the sedimentation tank, Control the dosage according to water quality requirements.

7. A highway service area sewage treatment method based on the Internet of Things according to claim 6, characterized in that: In step S4, the closed-loop aeration control method is: A real-time feedback control mechanism is used to make instant adjustments based on the current dissolved oxygen level and ammonia nitrogen concentration. The dissolved oxygen control formula is: Among them, u DO (t) represents the dissolved oxygen regulation signal at time t, K p Represents the proportional gain, which controls the direct impact of dissolved oxygen error on the regulation signal, K i Indicates the integral gain, K d represents the differential gain, which is the speed of response to the change of dissolved oxygen. dτ represents the small increment of the time variable τ. set represents the target dissolved oxygen level, which is set according to the optimal microbial activity, and DO(t) represents the actual dissolved oxygen concentration at time t, which is obtained from real-time monitoring; Combined with the predicted trend of ammonia nitrogen concentration, the aeration time and intensity are dynamically adjusted. The aeration intensity adjustment formula is: Among them, A(t) represents the aeration intensity at time t, which affects the oxygen supply, A base It represents the basic aeration intensity, that is, the system's default minimum oxygen demand. K1 represents the ammonia nitrogen adjustment coefficient, which adjusts the effect of ammonia nitrogen concentration deviation on aeration intensity. represents the predicted ammonia nitrogen concentration at time t, from LSTM prediction, It represents the actual ammonia nitrogen concentration at time t, the real-time monitoring value, and k2 represents the ammonia nitrogen change rate coefficient, reflecting the dynamic effect of the ammonia nitrogen change rate on the aeration intensity; Dynamically optimize aeration time, adjust aeration cycle according to current and predicted water quality trends, aeration time adjustment formula: Among them, T air (t) represents the aeration time length at time t, T base It represents the basic aeration time, i.e. the minimum aeration time period set by the system. K3 represents the cumulative ammonia nitrogen error influence coefficient, which adjusts the influence of the total error on the aeration time. K4 represents the dissolved oxygen adjustment signal influence coefficient, which reflects the influence of aeration intensity on the time length. Indicates the cumulative deviation of ammonia nitrogen concentration.

8. A highway service area sewage treatment method based on the Internet of Things according to claim 7, characterized in that: The intelligent scheduling method in step S5 is: When the load is low, reduce the equipment's operating power or shut down some equipment; when the load is high, adjust the working mode of each processing unit to balance operating efficiency and energy consumption.

9. A highway service area sewage treatment method based on the Internet of Things according to claim 8, characterized in that: In step S5, the energy consumption of each processing unit is monitored in real time in combination with the load forecast of step S2 and the adjustment result of step S3: To monitor energy consumption in real time, the edge computing unit collects the voltage and current of each processing unit in real time and calculates the current energy consumption level. The energy consumption calculation formula is: P i (t) = V i (t)×I i (t)×η i (t), where P i (t) represents the energy consumption of the i-th processing unit at time t, V i (t) represents the real-time voltage of the i-th processing unit at time t, I i (t) represents the real-time current of the i-th processing unit at time t, η i (t) represents the efficiency factor of the i-th processing unit at time t, reflecting the energy efficiency status of the equipment; According to the load forecast in step S2 and the current real-time flow data, the current load state factor of the system is calculated to determine whether it is currently in a high load or low load state. The load state factor is: Among them, L(t) represents the load state factor at time t, reflecting the system load status. represents the predicted sewage flow at time t, provided by step S2, Q max represents the maximum processing flow of the system design, α represents the weight factor, balancing the impact of flow and energy consumption in the load state, P total (t) represents the total energy consumption at time t, the sum of the energy consumption of all processing units, P max Indicates the maximum energy consumption level of the system design; According to the load state factor L(t), an intelligent dispatching strategy is implemented. When the load is low, the equipment operating power is adjusted or some equipment is shut down; when the load is high, the operating mode of each unit is adjusted to optimize energy consumption and efficiency. The dispatching adjustment method is: Low load state L(t)<θ low : in, represents the adjusted energy consumption of the i-th processing unit at time t, k1 represents the adjustment coefficient, which determines the extent of energy consumption reduction, θ low Indicates the low load threshold, the minimum load benchmark for system operation; High load state L(t)>θ high : Among them, k2 represents the high load adjustment coefficient, which determines the increase in energy consumption, θ high Indicates the high load threshold, the benchmark point for efficient operation of the system, L max It represents the maximum load state factor of the system, indicating full load operation; Under high load, each processing unit is optimized and deployed, and the optimization formula is: Among them, E eff (t) represents the system energy efficiency at time t, N represents the number of processing units, Q i (t) represents the processing flow of the i-th processing unit at time t, and ∈ represents a small constant.

Images