Plateau sewer-free public toilet self-sustaining operation and maintenance method and system based on intelligent optimization

Abstract

The invention discloses a plateau sewer-free public toilet self-sustaining operation and maintenance method and system based on intelligent optimization. Relates to the technical field of plateau public toilet operation and maintenance. The method is improved on the basis of an existing passive ecological treatment technology for a scene of a plateau environment, and based on algorithms of a real-time sensing network, a long and short-term memory network and a deep Q network, influences of unsteady-state characteristics such as large plateau day and night temperature difference and wind and light energy fluctuation are considered in the operation and maintenance process. The ecological health degree, the energy sustainability and the processing efficiency in operation are comprehensively considered in the decision making process, manual intervention and energy waste are reduced, and the long-term unattended operation self-sustaining operation and maintenance method suitable for the plateau extreme environment is provided.

Description

Technical Field

[0001] This application relates to the field of operation and maintenance technology for public toilets in high-altitude areas, and in particular to a self-sustaining operation and maintenance method and system for public toilets without sewers in high-altitude areas based on intelligent optimization. Background Technology

[0002] The high altitude, low oxygen levels, strong ultraviolet radiation, and extreme temperature differences in plateau regions pose severe challenges to the sustainable operation and maintenance of sanitation facilities. Public toilets without sewers serve as a core solution, achieving on-site treatment of wastewater through ecological recycling, eliminating reliance on municipal sewer networks, and constructing a micro-ecosystem where microbial communities and treatment units work in synergy, aiming to maintain stable operation with minimal intervention. However, the diurnal temperature range often exceeding 30°C and the abrupt changes in wind speed and sunlight in plateau regions make the system prone to instability, and insufficient self-sufficiency has become a technical bottleneck.

[0003] Existing public toilets mostly employ passive technologies, relying on fixed parameter controls such as temperature threshold-based heating activation and deactivation, and fixed-cycle ventilation and agitation. While these solutions are designed based on steady-state assumptions and initially reduced waste collection and energy consumption, their shortcomings become apparent in high-altitude areas and under high-standard requirements: fixed strategies cannot adapt to environmental uncertainties; on cloudy days, high-power equipment easily depletes its energy storage, leading to sterilization failure; delayed ventilation for energy saving, however, results in excessive humidity inside the chamber, inhibiting bacterial activity and reducing degradation efficiency, creating a vicious cycle. More importantly, existing systems lack a self-sustaining capacity quantification mechanism, making it difficult to balance the dynamic interplay between ecological health, energy sustainability, and processing efficiency, and lacking a unified evaluation function to support global decision-making.

[0004] Existing technologies lack static control logic and evaluation, making it impossible to dynamically weigh multi-dimensional indicators or provide forward-looking control capabilities. Therefore, there is an urgent need for a new operation and maintenance paradigm that integrates real-time perception, dynamic modeling, and intelligent decision-making to overcome existing limitations and achieve long-term unattended self-sustaining operation of public toilets in extreme high-altitude environments. Summary of the Invention

[0005] In order to at least overcome the above-mentioned shortcomings in the prior art, the purpose of this application is to provide a self-sustaining operation and maintenance method and system for public toilets without sewers in high-altitude areas based on intelligent optimization to solve the above problems.

[0006] Firstly, this application provides a self-sustaining operation and maintenance method for high-altitude, sewer-free public toilets based on intelligent optimization, including: To obtain environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers in plateau areas, and to obtain ecological health index, energy sustainability index, and treatment efficiency index; Based on the aforementioned ecological health index, energy sustainability index, and treatment efficiency index, a self-sufficiency index is obtained using an evaluation function. Based on the self-sufficiency index and the corresponding environmental and equipment status at that time, a health assessment model is obtained using a long short-term memory network to obtain a predicted self-sufficiency index and a predicted environmental and equipment status. Based on the predicted self-sufficiency index and the predicted environment and equipment status, operation and maintenance decisions are obtained; Based on the aforementioned operation and maintenance decision, and the difference between the real-time environment and equipment status and the set threshold, the equipment operating parameters are obtained; In response to the completion of equipment operation and maintenance decisions, update the self-sufficiency index.

[0007] In one possible implementation, the acquisition of environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers in plateau areas, and the obtaining of ecological health index, energy sustainability index, and treatment efficiency index, includes: Sensors that collect data on environmental monitoring, energy storage equipment data, and the environmental and equipment status of sewage treatment equipment will be configured as a sensor network through the Internet of Things; By setting a sampling period through the sensor network, data is collected periodically to obtain the first dataset; Based on the first dataset, the cabin temperature, relative humidity, redox potential, fecal pH, measured anaerobic bacteria activity, and measured aerobic bacteria activity are obtained. The ecological health index is then calculated. Based on the first dataset, the total renewable energy generation and total energy consumption of equipment within a set time period are obtained, the real-time collected state of charge of energy storage batteries is obtained, and an energy sustainability index is obtained through calculation. Based on the first dataset, the real-time ammonia concentration and hydrogen sulfide concentration are obtained, and the processing efficiency index is calculated.

[0008] In one possible implementation, the evaluation function is configured to: obtain the value of the current evaluation function by performing a weighted summation based on the current ecological health index, energy sustainability index, and treatment efficiency index, according to adjustable weight coefficients; The process of obtaining the self-sufficiency index based on the ecological health index, energy sustainability index, and treatment efficiency index, using an evaluation function, includes: The value of the current evaluation function is used as the self-sustainability index for this sampling period; The current state is obtained based on the self-sufficiency index corresponding to the previous sampling period; Based on the current state, and using a preset adjustable weight coefficient selection strategy, the specific value of the adjustable weight coefficient for this period is obtained.

[0009] In one possible implementation, the step of obtaining a health assessment model based on a long short-term memory network, according to the self-sufficiency index and the corresponding environmental and equipment status at a given time, to obtain a predicted self-sufficiency index and a predicted environmental and equipment status, includes: Based on the self-sufficiency index, the environmental and equipment status of the sewage treatment equipment at the corresponding time is obtained to obtain the second dataset; Based on the environmental monitoring, energy storage equipment data, and environmental and equipment status of sewage treatment equipment in historical plateau public toilets without sewers, the historical self-sufficiency index and the corresponding environmental and equipment status of sewage treatment equipment at the time are obtained as historical datasets. Based on the historical dataset, obtain the training dataset and the validation dataset; The initial health evaluation model is input based on the training dataset, and the health evaluation model after iterative training is obtained by using the mean squared error loss function. The health evaluation model is input into the validation dataset after iterative training. The difference between the predicted value and the actual value is calculated. The parameters of the health evaluation model are adjusted to obtain the health evaluation model. The second dataset is input into the health assessment model, which outputs a predicted self-sufficiency index and a predicted environmental and equipment status.

[0010] In one possible implementation, the health assessment model is configured to be constructed using a long short-term memory network, including an input layer, a hidden layer, and an output layer. The step of inputting the second dataset into the health assessment model and outputting a predicted self-sufficiency index and a predicted environmental and equipment status includes: The input layer receives the second dataset; Hidden layers include long short-term memory network hidden layers with a set number of layers; The output layer includes two parallel branches to obtain the predicted self-sufficiency index and the predicted environment and equipment status; Branch 1 is the branch for predicting self-sufficiency index, which is configured to output the predicted ecological health, predicted energy sustainability, and predicted processing efficiency for a set number of future time steps. Based on the predicted ecological health, predicted energy sustainability, and predicted treatment efficiency, the predicted self-sufficiency index for each time step is calculated. Branch 2 is a branch for predicting the environmental and equipment status, and is configured to output the predicted temperature, predicted relative humidity, predicted energy storage battery state of charge, predicted solar irradiance, and predicted wind speed for a set number of time steps in the future. Based on the predicted temperature, relative humidity, state of charge of the energy storage battery, solar irradiance, and wind speed, the predicted environment and equipment status at each time step are stitched together.

[0011] In one possible implementation, obtaining the operation and maintenance decision based on the predicted self-sufficiency index and the predicted environment and equipment status includes: Based on the predicted self-sufficiency index and the predicted environment and equipment status, an intelligent agent is obtained using an algorithm based on a deep Q-network. Based on the agent, a trained and fine-tuned agent is obtained using a preset state space, action space, and reward function. Based on the trained and fine-tuned agent, input real-time samples and output the action with the highest action value as the optimal action. Based on the optimal action, an operation and maintenance decision is obtained.

[0012] In one possible implementation, the agent is configured with a main network and a target network, and is also configured with a buffer of a set capacity to store experience samples. The step of obtaining a trained and fine-tuned agent based on the agent, using a preset state space, action space, and reward function, includes: Based on the empirical samples in the buffer, a predetermined number of empirical samples are extracted with replacement; based on the main network, the action value of the current state is obtained; based on the target network, the maximum action value of the next state is obtained; the number of samples is less than the set capacity of the buffer. The maximum action value of the next state is obtained based on the target network, and the target action value is obtained based on the Bellman equation. Based on the main network, the action value of the current state and the target action value are obtained, and the loss function is obtained based on the mean square error. The main network is updated by performing training iterations based on the loss function. The main network, after being trained for a set number of iterations, is synchronized to the target network to update the target network; When the main network and the target network reach the preset hyperparameters, a trained agent is obtained; Deploy the trained agents to public toilets without sewers on the plateau; The intelligent agent obtains real-time samples based on real-time collected data; Input real-time samples and output the action corresponding to the action with the highest action value; After the action corresponding to the action with the highest output value is completed, data is collected in real time; The real-time reward is obtained based on the reward function. For every set number of real-time samples obtained, the main network and the target network are updated. Once all the real-time rewards are non-negative, the trained and fine-tuned agent is obtained.

[0013] In one possible implementation, the empirical sample includes: The state space is obtained based on the current state and the next state, the real-time battery data, and the temperature difference between the outdoor temperature and the cabin temperature. The action space is obtained based on the discrete action combinations; After the action is executed, data is collected in real time to obtain changes in the self-sustainability index, changes in battery state of charge, and total energy consumption of the device that performed the action. Reward range constraints are set to obtain the reward function. The current state, action, reward, and next state are used as experience samples.

[0014] Secondly, this application provides a self-sustaining operation and maintenance system for a high-altitude public toilet without sewers based on intelligent optimization, including a data acquisition unit, a self-sustaining capacity index calculation unit, a health evaluation and prediction unit, an operation and maintenance decision-making unit, and an equipment operation unit connected in sequence. The data acquisition unit is configured to: acquire environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers on plateaus, and obtain ecological health index, energy sustainability index and treatment efficiency index; The self-sufficiency index calculation unit is configured to: obtain the self-sufficiency index based on the ecological health index, energy sustainability index, and treatment efficiency index, using an evaluation function; The health assessment and prediction unit is configured to: obtain a health assessment model based on a long short-term memory network according to the self-sufficiency index and its corresponding historical environment and equipment status, so as to obtain a predicted self-sufficiency index and a predicted environment and equipment status. The operation and maintenance decision unit is configured to: obtain operation and maintenance decisions based on the predicted self-sufficiency index and the predicted environment and equipment status; The equipment operation unit is configured to: obtain equipment operation parameters based on the operation and maintenance decision and the difference between the real-time environment and equipment status and a set threshold.

[0015] In summary, the beneficial effects that this application can achieve are: This application proposes a multi-dimensional index quantification method that integrates ecological, energy, and processing efficiency parameters to construct an evaluation function, solving the problem of the lack of a unified measure for self-sufficiency and enabling precise assessment of the operational status of public toilets. The proposed health evaluation model, trained with time-series data to capture environmental dependencies, addresses the inability to proactively predict risks and achieves accurate prediction of future states. The proposed intelligent agent training and fine-tuning method, through dynamic weighting and real-time deviation adjustment, solves the problem of poor adaptability of static control and achieves optimal output of maintenance actions. The methods in this application represent significant technological advancements and beneficial effects in improving the self-sufficiency and maintenance capabilities of public toilets in high-altitude areas. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the method steps in an embodiment of this application; Figure 2 This is a schematic diagram of the system structure according to an embodiment of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0018] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] Example 1 Please refer to the following: Figure 1The above is a schematic diagram of the steps of the self-sustaining operation and maintenance method for public toilets without sewers in high-altitude areas based on intelligent optimization provided in the embodiments of the present invention. Further, the self-sustaining operation and maintenance method for public toilets without sewers in high-altitude areas based on intelligent optimization may specifically include the contents described in steps S1-S6.

[0020] Step S1: Obtain environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers in plateau areas, and obtain the ecological health index, energy sustainability index, and treatment efficiency index.

[0021] Step S2: Based on the ecological health index, energy sustainability index, and treatment efficiency index, obtain the self-sufficiency index using the evaluation function.

[0022] Step S3: Based on the self-sufficiency index and its corresponding environmental and equipment status at a given time, obtain a health assessment model using a long short-term memory network to obtain a predicted self-sufficiency index and a predicted environmental and equipment status; input a real-time environmental and equipment status to obtain a predicted environmental and equipment status.

[0023] Step S4: Obtain operation and maintenance decisions based on the predicted self-sufficiency index and the predicted environmental and equipment status.

[0024] Step S5: Based on the operation and maintenance decision, obtain the equipment operating parameters based on the difference between the real-time environment and equipment status and the set threshold.

[0025] In this embodiment, step S6 is also included: updating the self-sufficiency index in response to the completion of the operation and maintenance decision of the equipment.

[0026] In the implementation of this application embodiment, a self-sustaining operation and maintenance method is designed for public toilets without sewers in high-altitude areas. Through the above steps, a comprehensive perception of the public toilet's operating status is achieved, and assessment and trend prediction are carried out to update and optimize decisions, thereby ensuring the long-term stable self-sustaining operation of the public toilet in the complex environment of high-altitude areas.

[0027] Public toilets without sewer systems in high-altitude areas refer to public health facilities built in high-altitude regions that do not rely on traditional municipal sewer networks and achieve on-site resource utilization and harmless treatment of fecal waste through internal treatment systems. These toilets typically integrate fecal waste treatment units, energy supply units, and environmental control units, designed to adapt to the unique environment of high-altitude areas and operate independently.

[0028] Self-sustaining operation and maintenance refers to the ability of public toilets in high-altitude areas without sewers to maintain stable operation and functionality over the long term through internal material circulation, energy flow, and intelligent control mechanisms, with little or no human intervention. This operation and maintenance model is characterized by greater autonomy, adaptability, and resilience.

[0029] Specifically, it can include the following: In this embodiment, a public toilet without sewers in a high-altitude area is deployed in a high-altitude region above 4,000 meters, specifically taking a county-level administrative region in a high-altitude area as an example.

[0030] In step S1, environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers in plateau areas are obtained, and ecological health index, energy sustainability index, and treatment efficiency index are obtained.

[0031] Sensors that collect data on environmental monitoring, energy storage equipment data, and the environmental and equipment status of sewage treatment equipment will be configured as a sensor network through the Internet of Things; By using a sensor network and setting a sampling period, data is collected periodically to obtain the first dataset; Based on the first dataset, the cabin temperature, relative humidity, redox potential, fecal pH, measured anaerobic bacteria activity, and measured aerobic bacteria activity were obtained. The ecological health index was then calculated. Based on the first dataset, the total renewable energy generation and total energy consumption of equipment within a set time period are obtained, the state of charge of energy storage batteries is collected in real time, and an energy sustainability index is obtained through calculation. Based on the first dataset, the real-time ammonia and hydrogen sulfide concentrations are obtained, and the processing efficiency index is calculated.

[0032] In the implementation of this application embodiment, sensors are deployed inside and around the public toilet. These sensors include an environmental monitoring sensor group, a sewage treatment equipment status sensor group, and an energy storage and energy harvesting equipment status sensor group.

[0033] Then, different sensors are connected to form a sensor network through the Internet of Things. This sensor network is then used to collect the first dataset in real time. The first dataset includes, but is not limited to: environmental monitoring of public toilets without sewers in the plateau, data from energy storage devices, and environmental and equipment status of sewage treatment equipment. Environmental monitoring of public toilets without sewers in high-altitude areas includes: chamber temperature, relative humidity, fecal pH, oxidation-reduction potential, oxygen concentration, methane concentration, aerobic bacteria activity index (represented by the total number of colonies per gram of sample, expressed in CFU / g, where CFU is colony-forming unit), and anaerobic bacteria activity index (unit: CFU / g). Energy storage device data includes: solar photovoltaic panel output power, wind turbine output power, and energy storage battery state of charge, denoted as SOC; The environmental and equipment status of sewage treatment equipment includes: ammonia concentration, hydrogen sulfide concentration, stirring motor current, ventilation fan speed, ultraviolet germicidal lamp working power, and heater power.

[0034] The first dataset was collected according to the same sampling period. In the early stage of collection, when it was used to train the subsequent health evaluation model and operation and maintenance decision, the sampling period could be set to once every 10 minutes. However, after it was put into actual use, it was uniformly set to 30 minutes, which is more energy-efficient and also takes into account the user experience.

[0035] Based on the first dataset, the specific index values ​​of the ecological health index, energy sustainability index, and treatment efficiency index were calculated respectively.

[0036] Ecological Health Index The ecological health index primarily depends on the functional state of the microbial community within the reaction chamber. Better community function, such as higher activity, leads to better fecal decomposition and less impact on the surrounding ecosystem. The better, the better, the ecological health index , can be represented as: ; Where T is the chamber temperature, RH is the relative humidity, ORP is the oxidation-reduction potential, pH is the acidity or alkalinity of the fecal waste, and Na is the measured anaerobic bacteria activity (CFU / g). The reference anaerobic bacterial activity threshold is the average value of the high-altitude adaptive bacterial community under optimal working conditions, determined through laboratory testing and subsequent calibration. To measure the aerobic bacteria activity (CFU / g). The aerobic bacteria activity threshold, which is used as a reference, is also calibrated in the laboratory. function to These are all functions with values ​​ranging from [0,1]. The closer they are to 1, the better the state. They can be represented as: ; Among them, let The optimal degradation temperature for microorganisms was determined, and then the temperature fluctuation tolerance was set to... For example, at T=35℃, At T=43℃, .

[0037] , where settings ; In practice, anaerobic digestion is often used to treat feces, converting organic matter into methane and carbon dioxide. The ORP (Operating Resource Ratio) value is directly related to the activity of microorganisms in decomposing organic matter. For example, a reducing environment of -200mV is conducive to the work of anaerobic bacteria, therefore, setting... ; , where settings Then set the bacterial community activity threshold. By continuously culturing bacteria at a simulated altitude of 4000 meters in the laboratory for 100 days under the same diurnal temperature range as the most common temperature in the deployment site (20°C), the activity threshold of the bacterial community can be obtained.

[0038] Weighting coefficient to satisfy Furthermore, based on the characteristics of plateau microbial communities, and to ensure that the functional state of the microbial community occupies the most important position in the ecological health index, after extensive experiments, the weights are set as follows: , , , , , When calculating the ecological health index, multidimensional ecological parameters are mapped to the [0,1] interval to achieve a comprehensive quantification of the functional status of the microbial community and avoid misjudgment by a single indicator.

[0039] Energy Sustainability Index The calculation relies on energy balance data over a certain period of time, such as 24 hours, and can be expressed as: ; in, This represents the total renewable energy generation over the past 24 hours. Because it's designed for high-altitude areas, it benefits from abundant solar and wind power resources. Through built-in energy metering sensors, it records the instantaneous power of solar and wind power every 30 minutes and calculates the total generation through numerical integration. ; It is the total energy consumption of the public toilet equipment within the same time period. Taking this embodiment as an example, it includes the sum of the energy consumption of the stirring motor, ventilation fan, heater, and ultraviolet germicidal lamp. SOC is the current state of charge of the energy storage battery, which is obtained directly from the battery itself. It is the contribution weight of renewable energy.

[0040] The energy sustainability index reflects the degree of dependence on local energy and the energy storage buffer capacity. The value range is [0,1], and the closer it is to 1, the more sustainable the energy supply is.

[0041] Processing efficiency index This is directly related to public health and safety, and its calculation is based on real-time gas concentration, which can be expressed as: ; in, This is the real-time ammonia concentration. The limit value was obtained according to the national standard GB / T17217-2021 "Hygiene Standard for Public Toilets" and thus serves as the limit value. This is the real-time hydrogen sulfide concentration, obtained according to national standards. As a limit; Then put Set as This is the ammonia weight, because ammonia has a lower olfactory threshold, and people are more sensitive to it.

[0042] In actual use, because ammonia and hydrogen sulfide concentrations are also related to public health and safety, to ensure that the self-managed operation and maintenance of public toilets meets the standards, if the concentration of either gas exceeds the standard, the corresponding value will be negative, resulting in... It immediately enters a high-priority alarm state, including: stopping people from using it, waiting for the reaction chamber to continue reacting, and the equipment operating at the highest level. After a set time, staff will carry out maintenance.

[0043] In step S2, the evaluation function is configured as follows: based on the current ecological health index, energy sustainability index, and treatment efficiency index, a weighted sum is performed according to adjustable weight coefficients to obtain the value of the current evaluation function. Based on the ecological health index, energy sustainability index, and treatment efficiency index, and using the evaluation function, a self-sufficiency index is obtained.

[0044] The value of the current evaluation function is used as the self-sustainability index for this sampling period; The current state is obtained based on the self-sufficiency index corresponding to the previous sampling period; Based on the current state, and using a preset selection strategy for adjustable weight coefficients, the specific value of the adjustable weight coefficient for this period is obtained.

[0045] In the implementation of this application embodiment, the existing IoT sensor data has been quantified into an ecological health index, an energy sustainability index, and a processing efficiency index. Then, an evaluation function is constructed to obtain a self-sufficiency index, which serves as a unified measure of the resilience of the overall high-altitude sewer-free public toilets in self-sufficiency operation and maintenance. The evaluation function can be expressed as: ; Among them, the evaluation function It is an adjustable weighting coefficient that satisfies And adjustments are made according to the operational phase.

[0046] Initially, the design prioritized the ecological health index, with the energy sustainability index being slightly higher than the treatment efficiency index. The weighting coefficients were directly set to... Based on this idea, a set of adjustable weight coefficient selection strategies is also set up: When calculated in real time If the current state is considered fragile, then prioritizing ecological health and amplifying the importance of the ecological health index is crucial. Therefore, the next cycle will be set as follows: Strengthen ecological restoration; when At this point, it is considered to be in a stable state, with the three factors in equilibrium. It remains unchanged; when At this point, it is considered to be in a robust state, and processing efficiency can be appropriately increased to release redundant capacity. .

[0047] This evaluation function transforms the multi-objective optimization problem into a single-objective maximization problem, giving subsequent decisions a clear optimization direction, and Adjustable weighting coefficients are implemented using a lookup table method, eliminating the need for complex logical judgments. Because the weights of each index are set to sum to one, the final calculated self-sustaining capacity index... Then set It is considered to have basic self-sufficiency. It is considered a high-level self-sufficiency.

[0048] In step S3, based on the self-sufficiency index and its corresponding historical environment and equipment status, a health assessment model is obtained using a long short-term memory network to obtain a predicted self-sufficiency index and a predicted environment and equipment status.

[0049] Based on the self-sufficiency index, the environmental and equipment status of the sewage treatment equipment at the corresponding time is obtained to obtain the second dataset; Based on the environmental monitoring, energy storage equipment data, and environmental and equipment status of sewage treatment equipment in historical plateau public toilets without sewers, the historical self-sufficiency index and the corresponding environmental and equipment status of sewage treatment equipment at the time are obtained as historical datasets. Based on historical datasets, obtain training and validation datasets; The initial health assessment model is input based on the training dataset, and the health assessment model after iterative training is obtained by using the mean squared error loss function. The health assessment model is iteratively trained based on the validation dataset. The difference between the predicted value and the actual value is calculated. The parameters of the health assessment model are then adjusted to obtain the health assessment model. Input the second dataset into the health assessment model, and output the predicted self-sufficiency index and the predicted environmental and equipment status.

[0050] The health assessment model is configured to be constructed using a long short-term memory network, including an input layer, hidden layers, and an output layer. The second dataset is input into the health assessment model, which outputs a predicted self-sufficiency index and predicted environmental and equipment conditions, including: The input layer receives the second dataset; The hidden layer includes a long short-term memory network hidden layer with a set number of layers to capture the long and short-term dependencies of time-series data; for example, the dependency between day and night temperature changes and cabin temperature changes, and the dependency between the time-varying patterns of photovoltaic and wind power and the state of charge of energy storage batteries, etc.

[0051] The output layer includes two parallel branches to obtain the predicted self-sufficiency index and the predicted environment and equipment status; Branch 1 is the branch for predicting self-sufficiency index, which is configured to output the predicted ecological health, predicted energy sustainability, and predicted processing efficiency for a set number of future time steps. Based on the predicted ecological health, predicted energy sustainability, and predicted treatment efficiency, the predicted self-sufficiency index for each time step is calculated. Branch 2 is a branch for predicting the environmental and equipment status, and is configured to output the predicted temperature, predicted relative humidity, predicted energy storage battery state of charge, predicted solar irradiance, and predicted wind speed for a set number of time steps in the future. Based on the predicted temperature, relative humidity, state of charge of the energy storage battery, solar irradiance, and wind speed, the predicted environment and equipment status at each time step are stitched together.

[0052] In the implementation of this application embodiment, a health assessment model is constructed using a long short-term memory network model, mainly for the purpose of predicting future health and equipment status.

[0053] Based on the self-sufficiency index, the environmental and equipment status of the sewage treatment equipment at the corresponding time is obtained. Then, outliers are removed and missing values ​​are filled from the collected data. All data is then normalized and mapped to the [0,1] interval as the data basis for the second dataset. The second dataset consists of a sequence of N past time steps, for example, N=48, corresponding to 24 hours, with a sampling interval of 30 minutes. This forms a sequence, with each time step including: Then the health assessment model takes the second dataset as input; in The solar irradiance at time t is obtained by inverse calculation from the open-circuit voltage of the photovoltaic panel. The wind speed at time t is calculated from the wind turbine's rotational speed.

[0054] Based on the historical data obtained from environmental monitoring, energy storage equipment data, and environmental and equipment status of sewage treatment equipment in public toilets without sewers on plateaus, the historical self-sufficiency index and the corresponding environmental and equipment status of sewage treatment equipment at that time are obtained as historical datasets.

[0055] The health assessment model is constructed using a long short-term memory network, designed here to include two hidden layers, each with 128 memory units, and the output layer has two branches: One branch utilizes historical environmental monitoring and energy storage equipment data to generate a historical self-sufficiency index, predicts the self-sufficiency index, and outputs the predicted ecological health level for the k-th step in the future. Energy sustainability Processing efficiency ,For example This refers to the forecast for the next 3 hours.

[0056] Then there's branch two, which uses historical environment and equipment states at corresponding times to predict the environment and equipment states, outputting a predicted environment and equipment state vector for the k-th future step. .

[0057] Based on the historical dataset, obtain the training dataset and the validation dataset; divide the historical dataset into the training dataset and the validation dataset in a 7:3 ratio. The training dataset is used for model parameter learning, and the validation dataset is used to monitor the training effect and avoid overfitting.

[0058] Model training uses the mean squared error to calculate the loss function, which can be expressed as: ; in It is a true historical self-sufficiency index obtained from real ecological health, energy sustainability, and treatment efficiency. To ensure the authentic historical environment and equipment condition, then This refers to the environmental and equipment condition prediction weights. The environmental and equipment condition prediction weights are directly set to... .

[0059] Set the number of training iterations to 200 rounds and enable the early stopping strategy, which means stopping training when the validation set loss does not decrease for 10 consecutive rounds to ensure that the model converges and does not overfit. Finally, save the health evaluation model after training.

[0060] Finally, the reserved validation set data is input into the trained health assessment model to calculate the difference between the predicted value and the true value, and to verify the model's prediction accuracy. If the accuracy is not up to standard, such as the difference between the predicted value and the true value being greater than the preset threshold, the model parameters are adjusted and the model is retrained until the accuracy requirement is met, which is then used as the health assessment model after iterative training.

[0061] When real-time environmental monitoring data, energy storage device data, and the environmental and equipment status of sewage treatment equipment are used as current real-time data and input into the health evaluation model after iterative training, the real-time environmental and equipment status vector at the current time t is then used in the inference phase. After concatenating with historical vectors, the data is input into a trained health assessment model, which outputs a predicted sequence for a specified future time period.

[0062] In order to make predictions by rolling over time, the current real-time data is concatenated with the historical data of the adjacent previous time steps. For example, within 3 hours, the latest state is concatenated to the end of the previous historical data every 30 minutes. Then, the model is input from the most recent 48 steps to obtain the rolling prediction for the next 3 hours. That is, the real-time data of the current time t is concatenated with the historical data of the past 47 time steps to form a 48×8 input sequence.

[0063] The concatenated input is fed into the trained health assessment model. After capturing temporal dependencies through two hidden layers, the results are output from two output branches: Based on the output of branch 1, six time steps within the next three hours. , , The predicted self-sustainability index for the corresponding time step is calculated using a dynamic weighting formula. .

[0064] Then directly output the 6 time steps within the next 3 hours of branch two. It includes predicted values ​​for temperature, humidity, SOC, solar irradiance, and wind speed.

[0065] Then, the prediction results are used to determine whether to enter the predicted self-sufficiency index. The risk range can be identified, allowing for proactive regulation. For example, taking 15:00 as a typical winter daytime, real-time data collected at this time yields... The external temperature is -20℃, the internal temperature is 20℃, the solar irradiance is 300W / m², and the wind speed is 4m / s. The health assessment model directly predicts that in the next 3 hours, after sunset, the energy storage battery's state of charge will drop to 10% due to the approach of night, and the cabin temperature will drop to 10℃. The predicted self-sufficiency index... If the price drops directly below the 0.6 threshold, the prediction result will be used to trigger a preventative heating strategy to avoid entering a vulnerable state.

[0066] In step S4, an operation and maintenance decision is obtained based on the predicted self-sufficiency index and the predicted environmental and equipment status.

[0067] Based on the predicted self-sufficiency index and the predicted environmental and equipment status, an intelligent agent is obtained using an algorithm based on a deep Q-network. Based on the agent, a trained and fine-tuned agent is obtained using a preset state space, action space, and reward function. Based on the trained and fine-tuned agent, input real-time samples and output the action with the highest action value as the optimal action; Based on the optimal action, obtain operation and maintenance decisions.

[0068] Based on the agent, and using a preset state space, action space, and reward function, a trained and fine-tuned agent is obtained, including: The agent is configured with a main network and a target network, and is also configured with a buffer of a set size to store experience samples; Furthermore, hyperparameters including discount factor and exploration rate are preset; Based on the empirical samples in the buffer, extract a specified number of empirical samples with replacement; based on the main network, obtain the action value of the current state; based on the target network, obtain the maximum action value of the next state; the number of samples is less than the set capacity of the buffer. Based on the target network, obtain the maximum action value of the next state, and then obtain the target action value based on the Bellman equation. Based on the action value and target action value obtained from the main network, and the loss function is obtained based on the mean squared error; Based on the loss function, training iterations are performed to update the main network; The main network, after a set number of training iterations, is synchronized to the target network to update the target network; When the main network and the target network reach the preset hyperparameters, the trained agent is obtained.

[0069] This embodiment can also involve deploying the trained agent to a public toilet without sewers on a plateau, and then making further fine-tuning.

[0070] The agent is configured with a main network and a target network, and is also configured with a buffer of a set size to store experience samples; Furthermore, hyperparameters including discount factor and exploration rate are preset; Based on the empirical samples in the buffer, extract a specified number of empirical samples with replacement; based on the main network, obtain the action value of the current state; based on the target network, obtain the maximum action value of the next state.

[0071] In this embodiment, in order to ensure the randomness and independence of the samples and reduce the interference of data correlation on training, the number of samples is not only less than the set capacity of the buffer, but also much less than the set capacity of the buffer. Here, "much less than" means that the number of samples is less than one-thousandth of the set capacity of the buffer.

[0072] Based on the target network, obtain the maximum action value of the next state, and then obtain the target action value based on the Bellman equation. Based on the action value and target action value obtained from the main network, and the loss function is obtained based on the mean squared error; Based on the loss function, training iterations are performed to update the main network; The main network, after a set number of training iterations, is synchronized to the target network to update the target network; When the main network and the target network reach the preset hyperparameters, the trained agent is obtained; Then, the trained agent obtains real-time samples based on the real-time collected data; Input real-time samples and output the action corresponding to the action with the highest action value; After the action with the highest output value is completed, data is collected in real time. Receive real-time rewards based on the reward function; Once a set number of real-time samples are obtained, the main network and the target network are updated. Once all real-time rewards are non-negative, the trained and fine-tuned agent is obtained.

[0073] Empirical samples, including: The state space is obtained based on the current state and the next state, the real-time battery data, and the temperature difference between the outdoor temperature and the cabin temperature. The action space is obtained based on the discrete action combinations; After the action is executed, data is collected in real time to obtain changes in the self-sustainability index, changes in battery state of charge, and total energy consumption of the device that performed the action. Reward range constraints are set to obtain the reward function. The current self-sustaining capacity index, actions, rewards, and predicted self-sustaining capacity index are used as empirical samples.

[0074] In this embodiment, step S4, which involves obtaining an operation and maintenance decision based on the predicted self-sufficiency index and the predicted environmental and equipment status, may further include: Based on the predicted self-sufficiency index and the predicted environmental and equipment status, an intelligent agent is obtained using an algorithm based on a deep Q-network. Based on the agent, a trained agent is obtained using a preset state space, action space, and reward function; Based on the trained agent, input real-time samples and output the action with the highest action value as the optimal action; Based on the optimal action, obtain operation and maintenance decisions.

[0075] In the implementation of this application embodiment, based on the prediction results, in order to convert them into actions to be executed, a learning agent is designed using the algorithm of a deep Q-network to generate operation and maintenance decisions.

[0076] State space of the agent Defined to include the current Predictions for the next 1-3 steps ,current Temperature difference between inside and outside This can be represented as: ; Then among them, It is the current self-sufficiency index. It is a predictive self-sufficiency index. ; This is the temperature difference between the outside and inside of the cabin; it reflects the heat load. .

[0077] The equipment included in a public toilet is the minimum required to meet the basic requirements. Additional equipment can be added based on the actual budget of a specific project. The motion spaces of all this equipment are discretized into equipment operation combinations, and then the motion space is denoted as... ,include: Ventilation fan: Off, Low speed (set to 500 rpm), Medium speed (set to 1000 rpm), High speed (set to 1500 rpm); Mixing motor: Off, low frequency (set to mix once every 30 minutes for 2 minutes each time), high frequency (set to mix once every 15 minutes for 3 minutes each time). Heater: Off, Low Power (set to 50W heating), Medium Power (set to 100W heating), High Power (set to 200W heating); Ultraviolet germicidal lamp: Off, On (to allow the ultraviolet lamp to work continuously).

[0078] In this embodiment, the action space The total number of actions is 4×3×4×2=96 combinations, which is 96 discrete actions.

[0079] reward function Designed as follows: ; in This is to increase self-sufficiency. The change is used for energy storage, and then the positive value is defined as charging. The instantaneous power consumption corresponding to the selected action; Then set the reward weight to The reward weighting is set to encourage improvements in the self-sufficiency index and the state of charge of energy storage batteries, while penalizing situations with high energy consumption.

[0080] Then, using predicted data, historical monitoring data, and device power consumption parameters, a simulated training environment can be constructed to train the agent. This generates empirical samples of the current self-sustainability index, actions, rewards, and the next self-sustainability index. Using a deep Q-network algorithm, the parameters of the main network and target network are first initialized. Then, a batch of samples is randomly sampled from the empirical samples, and the main network is used to calculate the current self-sustainability index. of Value, calculate the next self-sustainability index using the target network. The largest Value; calculate the target based on the Bellman equation. Value, represented as , It's a discount factor, and then it's based on the output of the main network. and The mean squared error is calculated as the loss function, and the main network parameters are updated through the Adam optimizer; After setting the number of iteration rounds, the latest parameters of the main network are synchronized to the target network to avoid excessive fluctuations in the target value; To ensure the accuracy and stability of the agent, a discount factor γ is set to 0.95, and the exploration rate is... The exploration rate decreases linearly from 1.0 to 0.1, and finally, when the exploration rate decreases from the initial full exploration, it becomes... =1.0 linearly decays to 0.1, that is, after stabilization, explore 10%, and iterate training until the loss function converges. For example, if the loss value of a continuously set number of rounds is lower than the preset threshold, offline training can be stopped.

[0081] After offline training, the target network, i.e., the agent, is deployed to an actual public toilet on the plateau for fine-tuning to ensure adaptation to real-world environmental fluctuations, based on the current self-sufficiency index. Select Action Early stage Higher level, exploring actions that are effective after actual execution in real-world environments; later stage Stable, primarily for utilization, with actual data collected after actions are executed. Calculate real-time rewards ,Will , , , It is stored as a new sample in the experience sample.

[0082] Every 100 real-time experience samples collected triggers a fine-tuning of the main network parameters, synchronously updating the target network as an intelligent agent to adapt to dynamic scenarios such as diurnal temperature differences, irradiance fluctuations, and changes in manure load in the plateau environment, thereby improving the robustness of decision-making.

[0083] After completing the fine-tuning above, when the loss function is stable and the reward value remains positive, the agent is considered to have completed training. The agent's optimal action can then be directly expressed as... The action with the highest Q value is directly selected, which is the operation and maintenance decision, including the specific operating settings of the ventilation fan, stirring motor, heater, and ultraviolet germicidal lamp.

[0084] Take a specific operation and maintenance decision as an example: If the forecast shows that S will drop from 0.80 to 0.55 in the next 3 hours, and SOC is sufficient (defined as above 50% for sufficient energy), then the intelligent agent may choose to use a heater with medium power (100W) + a ventilation fan with medium speed (1000rpm) + a stirring motor with high frequency to enhance microbial activity and inhibit ammonia accumulation. If the SOC is low, defined as below 30% indicating insufficient energy, but the S is still above 0.7, then only low-speed ventilation may be activated to prioritize power supply.

[0085] In step S5, based on the operation and maintenance decision, the equipment operating parameters are obtained based on the difference between the real-time environment and equipment status and the set threshold.

[0086] In the implementation of this application embodiment, the discrete actions of the output operation and maintenance decision are mapped to specific equipment operating parameters, and a proportional-integral-derivative fine-tuning mechanism is introduced to deal with prediction deviations.

[0087] Then, taking a heater as an example, if the operation and maintenance decision is "medium power", then the basic set power is... However, in actual operation, if the real-time temperature... With the preset target temperature There is a deviation, the deviation is The final output power is: ; In this embodiment, the equipment is designed to prevent steady-state deviation in the integral term and suppress overshoot in the derivative term. However, for the safety of equipment operation, limitations are imposed, and the amplitude is limited to ensure... Based on numerous experiments, specific values ​​that can be obtained directly are derived, for example... .

[0088] Similarly, taking a ventilation fan as an example, the ventilation fan speed... The setting reference is the specified gear of the action, and then based on the ammonia concentration deviation. ,in Perform PID control.

[0089] Then, what's a bit hard to think of is controlling the relative humidity inside the chamber. This is done by using a stirring motor to control the stirring frequency, adjusting it based on the deviation between the moisture content of the sewage and the target value. If the moisture content is too high, for example, greater than 60%, the stirring frequency is increased to promote water evaporation and increase the relative humidity inside the chamber; if the moisture content is too low, for example, less than 40%, the stirring is reduced to preserve humidity and lower the relative humidity inside the chamber.

[0090] In step S6, in response to the completion of the operation and maintenance decision of the equipment, the self-sufficiency index is updated.

[0091] In the implementation of this application embodiment, after the equipment makes an operation and maintenance decision, after one sampling period (30 minutes), the environmental and equipment status is re-collected, and steps S1 to S2 are repeated to calculate a new self-sufficiency index. This index serves as state feedback and is used for: Update the input sequence of the health assessment model, and then rolling predictions can be achieved; The immediate reward of the agent's reward function in a deep Q-network directly drives policy optimization; In this embodiment, a trigger-based tiered alarm is also added: if three consecutive cycles... Then, a remote maintenance request is sent via the satellite communication module.

[0092] Based on algorithms using real-time sensor networks, long short-term memory networks, and deep Q-networks, and considering the impact of non-steady-state characteristics such as large diurnal temperature variations and fluctuations in wind and solar energy during operation and maintenance, this invention comprehensively incorporates ecological health, energy sustainability, and processing efficiency into the decision-making process. Through the closed-loop iteration of the above six steps, this invention realizes a complete self-sustaining operation and maintenance cycle from data collection, evaluation, prediction, strategy acquisition, strategy execution, and then to maintenance, reducing human intervention and energy waste, and providing a long-term unattended self-sustaining operation and maintenance method for public toilets adapted to the extreme environment of high altitudes.

[0093] Example 2 This is the second embodiment of the present invention. Based on Embodiment 1, please refer to the following: Figure 2This is a structural diagram of a self-sustaining operation and maintenance system for a public toilet without sewers in a high-altitude area based on intelligent optimization, provided in an embodiment of the present invention. This embodiment provides a self-sustaining operation and maintenance system for a public toilet without sewers in a high-altitude area based on intelligent optimization, including a data acquisition unit, a self-sustaining capacity index calculation unit, a prediction unit, an operation and maintenance decision unit, and an equipment operation unit connected in sequence by electrical connection. The data acquisition unit is configured to: acquire environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers on plateaus, and obtain ecological health index, energy sustainability index and treatment efficiency index; The self-sufficiency index calculation unit is configured to obtain the self-sufficiency index based on the ecological health index, energy sustainability index, and treatment efficiency index, using an evaluation function. The prediction unit is configured to: obtain a health assessment model based on a long short-term memory network, according to the self-sufficiency index and its corresponding historical environment and equipment status, so as to obtain the predicted self-sufficiency index and the predicted environment and equipment status. The operation and maintenance decision unit is configured to: make operation and maintenance decisions based on the predicted self-sufficiency index and the predicted environmental and equipment status; The equipment operation unit is configured to obtain equipment operation parameters based on the difference between the real-time environment and equipment status and the set threshold, according to operation and maintenance decisions.

[0094] Example 3 The third embodiment of the present invention differs from the previous embodiments in that: Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0095] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0096] The units described as separate components may or may not be physically separate. As will be apparent to those skilled in the art, the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0097] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0098] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or grid device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0099] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A self-sustaining operation and maintenance method for public toilets without sewers in high-altitude areas based on intelligent optimization, characterized in that: include: To obtain environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers in plateau areas, and to obtain ecological health index, energy sustainability index, and treatment efficiency index; Based on the aforementioned ecological health index, energy sustainability index, and treatment efficiency index, a self-sufficiency index is obtained using an evaluation function. Based on the self-sufficiency index and the corresponding environmental and equipment status at that time, a health assessment model is obtained using a long short-term memory network to obtain a predicted self-sufficiency index and a predicted environmental and equipment status. Based on the predicted self-sufficiency index and the predicted environment and equipment status, operation and maintenance decisions are obtained; Based on the aforementioned operation and maintenance decision, and the difference between the real-time environment and equipment status and the set threshold, the equipment operating parameters are obtained; In response to the completion of equipment operation and maintenance decisions, update the self-sufficiency index.

2. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 1, characterized in that, The acquisition of environmental monitoring and energy storage equipment data for public toilets without sewers in plateau areas, as well as the environmental and equipment status of sewage treatment equipment, yields ecological health index, energy sustainability index, and treatment efficiency index, including: Sensors that collect data on environmental monitoring, energy storage equipment data, and the environmental and equipment status of sewage treatment equipment will be configured as a sensor network through the Internet of Things; By setting a sampling period through the sensor network, data is collected periodically to obtain the first dataset; Based on the first dataset, the cabin temperature, relative humidity, redox potential, fecal pH, measured anaerobic bacteria activity, and measured aerobic bacteria activity are obtained. The ecological health index is then calculated. Based on the first dataset, the total renewable energy generation and total energy consumption of equipment within a set time period are obtained, the real-time collected state of charge of energy storage batteries is obtained, and an energy sustainability index is obtained through calculation. Based on the first dataset, the real-time ammonia concentration and hydrogen sulfide concentration are obtained, and the processing efficiency index is calculated.

3. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 1, characterized in that, The evaluation function is configured to: obtain the value of the current evaluation function by performing a weighted summation based on the current ecological health index, energy sustainability index, and treatment efficiency index, according to adjustable weight coefficients; The process of obtaining the self-sufficiency index based on the ecological health index, energy sustainability index, and treatment efficiency index, using an evaluation function, includes: The value of the current evaluation function is used as the self-sustainability index for this sampling period; The current state is obtained based on the self-sufficiency index corresponding to the previous sampling period; Based on the current state, and using a preset adjustable weight coefficient selection strategy, the specific value of the adjustable weight coefficient for this period is obtained.

4. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 1, characterized in that, The step of obtaining a health assessment model based on the self-sufficiency index and the corresponding environmental and equipment status at a given time, using a long short-term memory network, to obtain a predicted self-sufficiency index and a predicted environmental and equipment status, includes: Based on the self-sufficiency index, the environmental and equipment status of the sewage treatment equipment at the corresponding time is obtained to obtain the second dataset; Based on the environmental monitoring, energy storage equipment data, and environmental and equipment status of sewage treatment equipment in historical plateau public toilets without sewers, the historical self-sufficiency index and the corresponding environmental and equipment status of sewage treatment equipment at the time are obtained as historical datasets. Based on the historical dataset, obtain the training dataset and the validation dataset; The initial health evaluation model is input based on the training dataset, and the health evaluation model after iterative training is obtained by using the mean squared error loss function. The health evaluation model is input into the validation dataset after iterative training. The difference between the predicted value and the actual value is calculated. The parameters of the health evaluation model are adjusted to obtain the health evaluation model. The second dataset is input into the health assessment model, which outputs a predicted self-sufficiency index and a predicted environmental and equipment status.

5. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 4, characterized in that, The health assessment model is configured to be constructed using a long short-term memory network, including an input layer, a hidden layer, and an output layer. The step of inputting the second dataset into the health assessment model and outputting a predicted self-sufficiency index and a predicted environmental and equipment status includes: The input layer receives the second dataset; Hidden layers include long short-term memory network hidden layers with a set number of layers; The output layer includes two parallel branches to obtain the predicted self-sufficiency index and the predicted environment and equipment status; Branch 1 is the branch for predicting self-sufficiency index, which is configured to output the predicted ecological health, predicted energy sustainability, and predicted processing efficiency for a set number of future time steps. Based on the predicted ecological health, predicted energy sustainability, and predicted treatment efficiency, the predicted self-sufficiency index for each time step is calculated. Branch 2 is a branch for predicting the environmental and equipment status, and is configured to output the predicted temperature, predicted relative humidity, predicted energy storage battery state of charge, predicted solar irradiance, and predicted wind speed for a set number of time steps in the future. Based on the predicted temperature, relative humidity, state of charge of the energy storage battery, solar irradiance, and wind speed, the predicted environment and equipment status at each time step are stitched together.

6. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 1, characterized in that, The step of obtaining operation and maintenance decisions based on the predicted self-sufficiency index and the predicted environmental and equipment status includes: Based on the predicted self-sufficiency index and the predicted environment and equipment status, an intelligent agent is obtained using an algorithm based on a deep Q-network. Based on the agent, a trained agent is obtained using a preset state space, action space, and reward function; Based on the trained agent, input real-time samples and output the action corresponding to the action with the highest action value as the optimal action; Based on the optimal action, an operation and maintenance decision is obtained.

7. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 6, characterized in that, The agent is configured with a main network and a target network, and is also configured with a buffer of a set capacity to store experience samples. The step of obtaining a trained agent based on the agent, using a preset state space, action space, and reward function, includes: Based on the empirical samples in the buffer, a predetermined number of empirical samples are extracted with replacement; based on the main network, the action value of the current state is obtained; based on the target network, the maximum action value of the next state is obtained; the number of samples is less than the set capacity of the buffer. The maximum action value of the next state is obtained based on the target network, and the target action value is obtained based on the Bellman equation. Based on the main network, the action value of the current state and the target action value are obtained, and the loss function is obtained based on the mean square error. The main network is updated by performing training iterations based on the loss function. The main network, after being trained for a set number of iterations, is synchronized to the target network to update the target network; When the main network and the target network reach the preset hyperparameters, a trained agent is obtained.

8. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 7, characterized in that, Also includes: Deploy the trained agents to public toilets without sewers on the plateau; The intelligent agent obtains real-time samples based on real-time collected data; Input real-time samples and output the action corresponding to the action with the highest action value; After the action corresponding to the action with the highest output value is completed, data is collected in real time; The real-time reward is obtained based on the reward function. For every set number of real-time samples obtained, the main network and the target network are updated. Once all the real-time rewards are non-negative, the trained and fine-tuned agent is obtained.

9. The self-sustaining operation and maintenance method for high-altitude sewer-free public toilets based on intelligent optimization according to claim 7, characterized in that, The empirical sample includes: The state space is obtained based on the current state and the next state, the real-time battery data, and the temperature difference between the outdoor temperature and the cabin temperature. The action space is obtained based on the discrete action combinations; After the action is executed, data is collected in real time to obtain changes in the self-sustainability index, changes in battery state of charge, and total energy consumption of the device that performed the action. Reward range constraints are set to obtain the reward function. The current state, action, reward, and next state are used as experience samples.

10. A self-sustaining operation and maintenance system for high-altitude public toilets without sewers, based on intelligent optimization, is characterized in that: It includes a data acquisition unit with sequential electrical connection, a self-sufficiency index calculation unit, a health assessment and prediction unit, an operation and maintenance decision-making unit, and an equipment operation unit; The data acquisition unit is configured to: acquire environmental monitoring data, energy storage equipment data, and environmental and equipment status data of public toilets without sewers on plateaus, and obtain ecological health index, energy sustainability index and treatment efficiency index; The self-sufficiency index calculation unit is configured to: obtain the self-sufficiency index based on the ecological health index, energy sustainability index, and treatment efficiency index, using an evaluation function; The health assessment and prediction unit is configured to: obtain a health assessment model based on a long short-term memory network according to the self-sufficiency index and its corresponding historical environment and equipment status, so as to obtain a predicted self-sufficiency index and a predicted environment and equipment status. The operation and maintenance decision unit is configured to: obtain operation and maintenance decisions based on the predicted self-sufficiency index and the predicted environment and equipment status; The equipment operation unit is configured to: obtain equipment operation parameters based on the operation and maintenance decision and the difference between the real-time environment and equipment status and a set threshold.

Images