A method, device and computer storage medium for treating wastewater in a plant

Abstract

The application provides a workshop wastewater treatment method, a workshop wastewater treatment device and a computer storage medium. The workshop wastewater treatment method comprises the following steps: acquiring first water quality sensor data and equipment operation logs; inputting the first water quality sensor data and the equipment operation logs into a large language model decision engine to obtain a large language model decision instruction; and performing classified treatment on the workshop wastewater according to the large language model decision instruction. The large language model decision instruction is introduced into the above-mentioned workshop wastewater treatment method, and the self-determination capability of the large language model is utilized to dynamically treat the workshop wastewater on the basis of automatic treatment. Compared with the establishment of an expert experience or a preset decision mode, a large amount of waste of human resources can be reduced, and the loss caused by human error in decision-making can be effectively reduced.

Description

Technical Field

[0001] This application relates to the field of wastewater treatment technology, and in particular to a workshop wastewater treatment method, a workshop wastewater treatment device, and a computer storage medium. Background Technology

[0002] There are many types of wastewater generated during daily life and production. Different types of wastewater require different treatment methods to effectively manage them and reduce the environmental pollution caused by direct discharge. Workshop cleaning wastewater includes emulsions, degreasing wastewater, degreasing cleaning wastewater, and cutting fluid cleaning wastewater, among which the concentration of organic matter is relatively high. Direct discharge of workshop cleaning wastewater would cause serious environmental pollution.

[0003] However, when faced with complex workshop wastewater, the current approach relies solely on operators to monitor the wastewater in real time and make corresponding manual decisions, resulting in a waste of human resources and poor timeliness and accuracy of decision-making. Summary of the Invention

[0004] To address the aforementioned technical problems, this application proposes a workshop wastewater treatment method, a workshop wastewater treatment device, and a computer storage medium.

[0005] To address the aforementioned technical problems, this application proposes a method for treating workshop wastewater, the method comprising:

[0006] Acquire data from the first water quality sensor, as well as equipment operation logs;

[0007] The first water quality sensor data and the device operation log are input into the large language model decision engine to obtain the large language model decision instructions;

[0008] The workshop wastewater is classified and treated according to the decision instructions of the large language model.

[0009] The large language model decision instructions include workshop wastewater diversion instructions;

[0010] The step of inputting the first water quality sensor data and the device operation log into the large language model decision engine to obtain large language model decision instructions includes:

[0011] Extract the traffic diversion instruction prompts from the prompt message text library;

[0012] The diversion instruction prompt information, the first water quality sensor data and the device operation log are input into the big language model decision engine to obtain the big language model decision instruction;

[0013] The process of classifying and treating workshop wastewater according to the decision instructions of the large language model includes:

[0014] The diversion valve of the workshop wastewater collection system is activated according to the decision instruction of the large language model to guide the workshop wastewater to the corresponding type of wastewater collection tank;

[0015] The wastewater collection tank includes: an acidic wastewater collection tank, an alkaline wastewater collection tank, an oily wastewater collection tank, and / or a combined wastewater collection tank.

[0016] The large language model decision instructions include workshop wastewater treatment instructions;

[0017] The step of inputting the first water quality sensor data and the device operation log into the large language model decision engine to obtain large language model decision instructions includes:

[0018] Extract processing instruction prompts from the prompt message text library;

[0019] The processing instruction prompt information, the first water quality sensor data and the device operation log are input into the big language model decision engine to obtain the big language model decision instruction;

[0020] The process of classifying and treating workshop wastewater according to the decision instructions of the large language model includes:

[0021] The diversion valve of the workshop wastewater treatment system is activated according to the decision instruction of the large language model to guide the workshop wastewater to the corresponding type of wastewater treatment tank;

[0022] The wastewater treatment tank includes: a wastewater equalization tank, a pH adjustment tank, an oxidation tank, and / or a comprehensive sludge tank.

[0023] Prior to classifying and treating the workshop wastewater according to the decision instructions of the large language model, the workshop wastewater treatment method further includes:

[0024] Obtain the data credibility, model prediction confidence, and engineering feasibility of the output of the large language model decision engine;

[0025] The confidence level of the data, the confidence level of the model prediction, and the feasibility of the project are weighted and fused to obtain the comprehensive confidence level of the decision instruction of the large language model.

[0026] An execution strategy for the decision instructions of the large language model is generated based on the comprehensive confidence score.

[0027] The workshop wastewater treatment method further includes, after classifying and treating the workshop wastewater according to the decision instructions of the large language model:

[0028] Acquire data from the second water quality sensor;

[0029] By comparing the data from the first water quality sensor and the data from the second water quality sensor, the execution result of the decision instruction of the large language model is obtained;

[0030] Adjust one or more of the overall confidence levels based on the execution result of the instruction;

[0031] The confidence model in the large language model decision engine is updated based on the adjusted confidence level.

[0032] The workshop wastewater treatment method further includes, after generating the execution strategy for the decision instructions of the large language model based on the comprehensive confidence level:

[0033] The decision instructions for the large language model whose overall confidence level is lower than the preset confidence threshold are output to the annotation platform;

[0034] Obtain the manual review information returned by the annotation platform;

[0035] The decision instructions of the large language model are annotated based on the manual review information to update the training set of the large language model decision engine.

[0036] The workshop wastewater treatment method further includes:

[0037] Based on the decision instructions from the large language model, human-machine collaborative instructions are generated;

[0038] Based on the human-machine collaborative instructions, the instruction basis and cost estimate of the large language model decision instructions are output in a natural language interpretation manner.

[0039] The acquisition of the first water quality sensor data includes:

[0040] Data on the workshop wastewater is collected using physicochemical property sensors, pollutant-specific sensors, and / or physical state sensors to obtain data from the first water quality sensor.

[0041] The physicochemical property sensors include: pH sensor, ORP sensor, conductivity sensor, turbidity sensor, and / or dissolved oxygen sensor; the pollutant-specific sensors include: ultraviolet fluorescence oil analyzer, ion-selective electrode, heavy metal sensor, and / or COD online analyzer; the physical state sensors include: electromagnetic flow meter, ultrasonic level gauge, and / or pressure transmitter.

[0042] To address the aforementioned technical problems, this application also proposes a workshop wastewater treatment device, which includes a memory and a processor coupled to the memory; wherein the memory is used to store program data, and the processor is used to execute the program data to implement the workshop wastewater treatment method described above.

[0043] To address the aforementioned technical problems, this application also proposes a computer storage medium for storing program data, which, when executed by a computer, is used to implement the aforementioned workshop wastewater treatment method.

[0044] Compared with existing technologies, the beneficial effects of this application are as follows: the workshop wastewater treatment device acquires data from a first water quality sensor and equipment operation logs; the first water quality sensor data and the equipment operation logs are input into a large language model decision engine to obtain large language model decision instructions; and the workshop wastewater is classified and treated according to the large language model decision instructions. By introducing large language model decision instructions into the above-mentioned workshop wastewater treatment method, the autonomous decision-making capability of the large language model is used to dynamically treat workshop wastewater on the basis of automated processing. Compared with the formulation of expert experience or preset decision-making models, this can reduce a large amount of human resource waste and effectively reduce losses caused by human error in decision-making. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] in:

[0047] Figure 1 This is a schematic flowchart of the first embodiment of the workshop wastewater treatment method provided in this application;

[0048] Figure 2 This is a schematic diagram of the architecture for multimodal data collaborative decision-making provided in this application;

[0049] Figure 3 This is a schematic flowchart of the second embodiment of the workshop wastewater treatment method provided in this application;

[0050] Figure 4 This is a schematic flowchart of the third embodiment of the workshop wastewater treatment method provided in this application;

[0051] Figure 5 This is a schematic diagram of the online feedback learning architecture of the large language model decision engine provided in this application;

[0052] Figure 6 This is a schematic flowchart of the fourth embodiment of the workshop wastewater treatment method provided in this application;

[0053] Figure 7 This is a schematic diagram of the training set update architecture of the large language model decision engine provided in this application;

[0054] Figure 8 This is a schematic diagram of the structure of an embodiment of the workshop wastewater treatment device provided in this application;

[0055] Figure 9 This is a schematic diagram of the structure of an embodiment of the computer storage medium provided in this application. Detailed Implementation

[0056] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0057] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0058] Please continue reading for details. Figure 1 and Figure 2 , Figure 1 This is a schematic flowchart of the first embodiment of the workshop wastewater treatment method provided in this application. Figure 2 This is a schematic diagram of the architecture for multimodal data collaborative decision-making provided in this application.

[0059] The workshop wastewater treatment method of this application is applied to a workshop wastewater treatment device. This device can be a server, a terminal device, or a system consisting of both. Accordingly, all components of the workshop wastewater treatment device, such as units, subunits, modules, and submodules, can be entirely housed in the server, entirely in the terminal device, or separately in both the server and the terminal device.

[0060] Furthermore, the aforementioned server can be either hardware or software. When the server is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When the server is software, it can be implemented as multiple software programs or software modules, such as software or software modules used to provide distributed server functionality, or as a single software program or software module; no specific limitations are made here.

[0061] like Figure 1 As shown, the specific steps are as follows:

[0062] Step S11: Obtain the data from the first water quality sensor and the equipment operation log.

[0063] In the embodiments of this application, such as Figure 2 As shown, the workshop wastewater treatment device needs to acquire water quality sensor data, equipment operation logs, and chemical substance maps of the workshop wastewater in advance. This multimodal data is then input into the large language model decision engine to provide the raw data for decision-making.

[0064] The water quality sensor data of the aforementioned workshop wastewater is mainly used to obtain water quality characteristics of the workshop wastewater in different dimensions, and different water quality sensors extract different water quality characteristics. Therefore, this application uses a large number and variety of water quality sensors to collaboratively extract water quality sensor data of workshop wastewater.

[0065] Specifically, the water quality sensors used in this application include any one or more of the following: physicochemical property sensors, pollutant-specific sensors, and physical state sensors. These types of water quality sensors are further subdivided according to the water composition.

[0066] For example, physicochemical property sensors include: pH sensors, ORP sensors, conductivity sensors, turbidity sensors, dissolved oxygen sensors, etc. Pollutant-specific sensors include: ultraviolet fluorescence oil separators, ion-selective electrodes, heavy metal sensors, online COD analyzers, etc. Physical state sensors include: electromagnetic flowmeters, ultrasonic level gauges, pressure transmitters, etc.

[0067] In workshop wastewater treatment systems, equipment operation logs are crucial data sources for recording equipment status, operational behavior, and system events, used for fault diagnosis, performance optimization, and large language model training. Specifically, the equipment operation logs for the aforementioned workshop wastewater treatment system include the following: basic equipment operating data, process control data, abnormal event records, and maintenance operation records. Furthermore, equipment operation logs may also include: equipment health indicators, energy efficiency assessment data, etc.

[0068] Step S12: Input the first water quality sensor data and equipment operation log into the big language model decision engine to obtain the big language model decision instructions.

[0069] In this embodiment of the application, the workshop wastewater treatment device simultaneously inputs the multimodal data from step S11. Figure 2 The large language model decision engine is used to obtain the large language model decision instructions output by the large language model decision engine.

[0070] The large language model decision engine can adopt mature large language models on the market, such as GPT and LLaMA. The method adopted in this application is to use relevant data on wastewater treatment to train and optimize the large language models on the market, so as to obtain a large language model decision engine that is close to the needs of the wastewater treatment technology field.

[0071] Specifically, this application can embed structured data such as the Integrated Wastewater Discharge Standard (GB 8978) and EPA regulations into the model, and inject 200,000 sets of wastewater component-treatment process correspondences to train and optimize the large language model decision engine.

[0072] Furthermore, the types of instructions output by the large language model decision engine include, but are not limited to, workshop wastewater diversion instructions and workshop wastewater treatment instructions. The specific type of instruction output by the large language model decision engine mainly depends on the prompt information input to the engine. Therefore, the workshop wastewater treatment device needs to periodically extract relevant prompt information from the prompt information text library based on the instructions from the staff or preset instructions, such as the diversion instruction prompt information corresponding to the workshop wastewater diversion instruction, and the treatment instruction prompt information corresponding to the workshop wastewater treatment instruction.

[0073] The instructions generated by the large language model in this application are multi-scale, adaptive, and interpretable, and their possible solutions include, but are not limited to:

[0074] Process level: Dynamically combine the optimal path from 20+ processing technologies.

[0075] Parameter level: Real-time fine-tuning of drug dosage (accuracy ±2%).

[0076] System level: Coordinate production scheduling and load handling (such as staggered operation).

[0077] In actual deployment, key instructions need to be verified through digital twin simulation, and manual review nodes need to be set up (such as mandatory confirmation of heavy metal handling instructions). In the future, federated learning can be combined to achieve cross-plant knowledge sharing and continuously optimize instruction generation capabilities.

[0078] It should be noted that the training and optimization process of the large model decision engine provided in this application is as follows:

[0079] During the data preparation phase, staff collected structured data in the field of wastewater treatment and 200,000 sets of wastewater component-treatment process correspondences in advance. The data was then cleaned, labeled, and used to construct training and validation sets.

[0080] During the model fine-tuning stage, the workshop wastewater treatment device adopts transfer learning technology, performs domain adaptation based on a pre-trained large model (such as GPT-4 or LLaMA), and sets a training objective: to minimize the prediction error of wastewater classification and treatment instructions, and adopts cross-entropy loss as the loss function.

[0081] During the model optimization phase, the workshop wastewater treatment unit adjusted the model parameters using the gradient descent algorithm, with the learning rate set to 1e-5 and the batch size to 32. Domain knowledge was also injected: regulatory provisions were transformed into prompt templates (such as "According to GB 8978, wastewater with excessive pH should be directed to ____ pool").

[0082] During the model validation and iteration phase, the workshop wastewater treatment system uses a digital twin system to simulate the execution effect of instructions, which is then fed back to the model optimization confidence weights (such as the adjustment formula for data confidence DC: DC_new=0.9*DC_old+0.1*actual effect score).

[0083] Step S13: Classify and treat workshop wastewater according to the decision instructions of the large language model.

[0084] In this embodiment, the workshop wastewater treatment device automatically performs classification and treatment of workshop wastewater according to the decision instructions of the large language model, thereby achieving automated management.

[0085] Specifically, the workshop wastewater treatment device can control the diversion valve according to the workshop wastewater diversion command, thereby guiding the workshop wastewater to the corresponding type of wastewater collection tank. Among them, the wastewater collection tanks involved in this application include, but are not limited to: acidic wastewater collection tanks, alkaline wastewater collection tanks, oily wastewater collection tanks, and comprehensive wastewater collection tanks.

[0086] In addition, the workshop wastewater treatment device can also guide workshop wastewater to corresponding types of wastewater treatment tanks according to workshop wastewater treatment instructions. Among them, the wastewater treatment tanks involved in this application include, but are not limited to: water equalization tanks, pH adjustment tanks, oxidation tanks, and integrated sludge tanks.

[0087] like Figure 2 As shown, the workshop wastewater treatment device can also control the dosing system to treat workshop wastewater according to the workshop wastewater treatment command. The type and amount of chemicals to be added in the dosing system can be indicated by the workshop wastewater treatment command.

[0088] In this application, the workshop wastewater treatment device acquires data from a first water quality sensor and equipment operation logs; it inputs the first water quality sensor data and the equipment operation logs into a large language model decision engine to obtain large language model decision instructions; and it classifies and treats the workshop wastewater according to the large language model decision instructions. By introducing large language model decision instructions into the above-mentioned workshop wastewater treatment method, the autonomous decision-making capability of the large language model is used to dynamically treat workshop wastewater on the basis of automated processing. Compared with expert experience or the formulation of preset decision-making models, this can reduce a large amount of waste of human resources and effectively reduce losses caused by human error in decision-making.

[0089] The automated processing logic of the workshop wastewater treatment method provided in this application can effectively shorten the process commissioning cycle from 2 weeks to 8 hours; handle process optimization problems under multiple constraints; and quickly adapt to new pollutant regulations (such as the new PFAS regulations).

[0090] Furthermore, in a large language model-driven wastewater treatment decision-making system, assigning confidence assessments to output commands is a crucial step in ensuring decision reliability. Please continue reading. Figure 3 , Figure 3 This is a schematic flowchart of the second embodiment of the workshop wastewater treatment method provided in this application.

[0091] like Figure 3 As shown, the specific steps are as follows:

[0092] Step S21: Obtain the data credibility, model prediction confidence, and engineering feasibility of the output of the large language model decision engine.

[0093] In the embodiments of this application, in addition to outputting large language model decision instructions, the large language model decision engine can also output, but not limited to, data confidence (DC), model prediction confidence (MC), and engineering feasibility (EF) through the confidence model.

[0094] Data confidence is used to assess the accuracy of sensor-collected data. For example, when a pH sensor is uncalibrated (DC < 0.5), the confidence of related neutralization commands is reduced. Model prediction confidence can be achieved through the Softmax probability output of a classification task (e.g., the probability of each category when BERT predicts wastewater type) or through the interval width predicted by a regression task (e.g., quantile regression in XGBoost). Evaluation factors for engineering feasibility include, but are not limited to, equipment availability, execution timeliness, and parameter safety margins.

[0095] Step S22: Weight and fuse the data confidence, model prediction confidence, and engineering feasibility to obtain the comprehensive confidence of the large language model decision instructions.

[0096] In this embodiment, the workshop wastewater treatment device performs a comprehensive confidence calculation on the confidence level output from step S21 above, and the weighted comprehensive formula is as follows:

[0097] Overall Confidence=0.4*DC+0.3*MC+0.3*EF

[0098] Step S23: Generate an execution strategy for the decision instructions of the large language model based on the comprehensive confidence score.

[0099] In this embodiment of the application, the workshop wastewater treatment device determines the execution depth of the large language model decision instruction based on the overall confidence level. For example, for large language model decision instructions with high confidence (≥0.8), the instruction is executed automatically; for large language model decision instructions with medium confidence (0.5-0.8), operator confirmation is required; and for large language model decision instructions with low confidence (<0.5), a re-evaluation is triggered.

[0100] Based on the aforementioned comprehensive confidence level and its implementation strategy, this application also provides a dynamic confidence level adjustment mechanism to optimize the training of the decision engine for large language models. Please refer to [link / reference] for details. Figure 4 and Figure 5 , Figure 4 This is a schematic flowchart of the third embodiment of the workshop wastewater treatment method provided in this application. Figure 5 This is a schematic diagram of the architecture of the online feedback learning of the large language model decision engine provided in this application.

[0101] like Figure 4 As shown, the specific steps are as follows:

[0102] Step S31: Obtain data from the second water quality sensor.

[0103] In this embodiment of the application, after executing the large language model decision instruction, the workshop wastewater treatment device uses the same water quality sensor as when collecting the first water quality sensor data to collect the second water quality sensor data of the treated workshop wastewater.

[0104] Step S32: By comparing the data from the first water quality sensor and the data from the second water quality sensor, obtain the instruction execution result of the large language model decision instruction.

[0105] In this embodiment, the workshop wastewater treatment device evaluates the instruction execution results of the large language model decision-making engine by analyzing changes in water quality sensor data at different times. The quality of the instruction execution results can be used to train and optimize the large language model decision engine.

[0106] Step S33: Adjust one or more confidence levels in the overall confidence level based on the instruction execution result.

[0107] In this embodiment of the application, when there are areas that need optimization in the instruction execution results, the workshop wastewater treatment device can guide the large language model decision engine to train and optimize through annotation, or it can update and optimize the confidence model of the large language model decision engine by adjusting the corresponding confidence in the comprehensive confidence.

[0108] like Figure 5 As shown, when the instruction execution result shows that the COD removal rate is lower than expected, the workshop wastewater treatment device can reduce the MC of similar instructions, thereby updating the confidence model; when the instruction execution result shows that the equipment alarms, the workshop wastewater treatment device can reduce the EF of similar instructions, thereby updating the confidence model.

[0109] Step S34: Update the confidence model in the large language model decision engine based on the adjusted confidence.

[0110] This application, through a multi-dimensional quantitative evaluation and dynamically adjusted confidence system, enables large language model instructions to be: more transparent, clearly displaying the decision-making basis and potential risks; more secure, automatically avoiding low-feasibility operations; and more intelligent, continuously learning and optimizing from execution feedback.

[0111] Please continue reading. Figure 6 and Figure 7 , Figure 6 This is a schematic flowchart of the fourth embodiment of the workshop wastewater treatment method provided in this application. Figure 7 This is a schematic diagram of the training set update architecture of the large language model decision engine provided in this application.

[0112] like Figure 6 As shown, the specific steps are as follows:

[0113] Step S41: Output the decision instructions of the large language model with a comprehensive confidence level lower than the preset confidence threshold to the annotation platform.

[0114] In this embodiment, the workshop wastewater treatment device directly puts the high-confidence content of the large language model pre-annotation results of the original sensor data into the instruction library; for the low-confidence content, subsequent manual review and manual annotation are required, mainly for updating the training set.

[0115] Step S42: Obtain the manual review information returned by the annotation platform.

[0116] Step S43: Annotate the decision instructions of the large language model based on the manual review information to update the training set of the large language model decision engine.

[0117] Furthermore, the workshop wastewater treatment device can also generate corresponding human-machine collaborative instructions based on the decision-making instructions of the large language model. These instructions are used to output the basis for the decision-making instructions of the large language model and cost estimates through natural language interpretation. For example, the following content can be output:

[0118] [Instruction Basis]

[0119] 1. Reasons for choosing Fenton oxidation instead of ozone:

[0120] The current wastewater contains phenol (which has strong absorption at 254 nm ultraviolet light).

[0121] The k of ·OH for p-phenol is 1.2 × 10⁻⁶. 10 M -1 s -1 (Much higher than ozone).

[0122] 2. Cost estimation:

[0123] H2O2: ¥2.3 / kg.

[0124] The total treatment cost is 38% lower than that of the ozone solution.

[0125] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0126] To achieve the above-mentioned workshop wastewater treatment method, this application also proposes a workshop wastewater treatment device, please refer to the details below. Figure 8 , Figure 8 This is a schematic diagram of an embodiment of the workshop wastewater treatment device provided in this application.

[0127] The workshop wastewater treatment device 500 of this embodiment includes a processor 51, a memory 52, an input / output device 53, and a bus 54.

[0128] The processor 51, memory 52, and input / output device 53 are connected to the bus 54. The memory 52 stores program data, and the processor 51 is used to execute the program data to implement the workshop wastewater treatment method described in the above embodiments.

[0129] In this embodiment, processor 51 can also be referred to as a CPU (Central Processing Unit). Processor 51 may be an integrated circuit chip with signal processing capabilities. Processor 51 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor, or processor 51 can be any conventional processor.

[0130] This application also provides a computer storage medium; please refer to the following: Figure 9 , Figure 9 This is a schematic diagram of a computer storage medium according to an embodiment of the present application. The computer storage medium 600 stores a computer program 61, which, when executed by a processor, is used to implement the workshop wastewater treatment method of the above embodiment.

[0131] When the embodiments of this application are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, 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 network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. 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.

[0132] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for treating workshop wastewater, characterized in that, The workshop wastewater treatment method includes: acquiring data from a first water quality sensor and equipment operation logs; The first water quality sensor data and the device operation log are input into the large language model decision engine to obtain the large language model decision instructions; the input of the first water quality sensor data and the device operation log into the large language model decision engine to obtain the large language model decision instructions includes: extracting diversion instruction prompt information from the prompt information text library; The diversion instruction prompt information, the first water quality sensor data and the device operation log are input into the big language model decision engine to obtain the big language model decision instruction; The workshop wastewater is classified and treated according to the decision instructions of the large language model; the classification and treatment of workshop wastewater according to the decision instructions of the large language model includes: activating the diversion valve of the workshop wastewater collection system according to the decision instructions of the large language model to guide the workshop wastewater to the corresponding type of wastewater collection tank; The wastewater collection tank includes: an acidic wastewater collection tank, an alkaline wastewater collection tank, an oily wastewater collection tank, and / or a combined wastewater collection tank. Before classifying and treating the workshop wastewater according to the decision instructions of the large language model, the workshop wastewater treatment method further includes: obtaining the data credibility, model prediction confidence, and engineering feasibility output by the large language model decision engine; The confidence level of the data, the confidence level of the model prediction, and the feasibility of the project are weighted and fused to obtain the comprehensive confidence level of the large language model decision instructions; the large language model decision instructions include workshop wastewater diversion instructions; An execution strategy for the decision instructions of the large language model is generated based on the comprehensive confidence score; After classifying and treating the workshop wastewater according to the decision instructions of the large language model, the workshop wastewater treatment method further includes: acquiring data from a second water quality sensor. By comparing the data from the first water quality sensor and the data from the second water quality sensor, the execution result of the decision instruction of the large language model is obtained; Adjust one or more of the overall confidence levels based on the execution result of the instruction; The confidence model in the large language model decision engine is updated based on the adjusted confidence level.

2. The workshop wastewater treatment method according to claim 1, characterized in that, The large language model decision instructions include workshop wastewater treatment instructions; The step of inputting the first water quality sensor data and the device operation log into the big language model decision engine to obtain big language model decision instructions includes: extracting processing instruction prompt information from the prompt information text library; The processing instruction prompt information, the first water quality sensor data and the device operation log are input into the big language model decision engine to obtain the big language model decision instruction; The step of classifying and treating workshop wastewater according to the decision instructions of the large language model includes: activating the diversion valve of the workshop wastewater treatment system according to the decision instructions of the large language model to guide the workshop wastewater to the corresponding type of wastewater treatment tank; The wastewater treatment tank includes: a wastewater equalization tank, a pH adjustment tank, an oxidation tank, and / or a comprehensive sludge tank.

3. The workshop wastewater treatment method according to claim 1, characterized in that, After generating the execution strategy for the decision instructions of the large language model based on the comprehensive confidence score, the workshop wastewater treatment method further includes: outputting the decision instructions of the large language model with a comprehensive confidence score lower than a preset confidence threshold to the annotation platform; Obtain the manual review information returned by the annotation platform; The decision instructions of the large language model are annotated based on the manual review information to update the training set of the large language model decision engine.

4. The workshop wastewater treatment method according to claim 1, characterized in that, The workshop wastewater treatment method further includes: Based on the decision instructions from the large language model, human-machine collaborative instructions are generated; Based on the human-machine collaborative instructions, the instruction basis and cost estimate of the large language model decision instructions are output in a natural language interpretation manner.

5. The workshop wastewater treatment method according to claim 1, characterized in that, The acquisition of the first water quality sensor data includes: collecting data on the workshop wastewater through a physicochemical property sensor, a pollutant-specific sensor, and / or a physical state sensor to acquire the first water quality sensor data; The physicochemical property sensors include: a pH sensor, an ORP sensor, a conductivity sensor, a turbidity sensor, and / or a dissolved oxygen sensor. The specific sensors for pollutants include: an ultraviolet fluorescence oil analyzer, an ion-selective electrode, a heavy metal sensor, and / or an online COD analyzer; The physical state sensors include: electromagnetic flowmeters, ultrasonic level gauges, and / or pressure transmitters.

6. A workshop wastewater treatment device, characterized in that, The workshop wastewater treatment device includes a memory and a processor coupled to the memory; The memory is used to store program data, and the processor is used to execute the program data to implement the workshop wastewater treatment method as described in any one of claims 1 to 5.

7. A computer storage medium, characterized in that, The computer storage medium is used to store program data, which, when executed by the computer, is used to implement the workshop wastewater treatment method as described in any one of claims 1 to 5.

Images