Multiple networked waste management system energy management optimization technique

Abstract

In waste management systems, energy management refers to dynamically adjusting the system's energy demand at different times, reducing energy consumption during peak electricity price periods and increasing energy consumption during off-peak periods to optimize energy costs in industrial waste management systems. Modern industrial systems typically consist of complex multi-layer industrial task networks. Traditional demand response techniques that only consider a single layer of the network and a single system are ineffective in reducing energy costs in multi-layer network environments, and may even lead to negative effects such as delays in overall system task completion times. This paper proposes a multi-networked waste management energy management optimization technique. This method divides multi-layer industrial network demand response energy optimization into inter-layer demand response and intra-layer demand response stages, which can improve the overall task execution efficiency in multi-networked waste management systems without affecting daily task completion requirements.

Description

Technical Field

[0001] This invention relates to a demand response energy management optimization method for multi-networked waste management systems, specifically an energy management optimization method for task scheduling and power load transfer under multi-network conditions. This method can adapt well to increasingly complex multi-network industrial scenarios, adaptively adjust the task execution of each network layer according to real-time electricity prices, and effectively reduce the electricity cost of waste management industrial systems without affecting the overall task objectives. Background Technology

[0002] In recent years, rapid global economic development and rising living standards have led to a surge in energy demand worldwide, inevitably posing significant challenges to the power sector. Demand response refers to the ability of intelligent user-side systems to flexibly adjust energy consumption based on electricity prices or other grid signals, thereby significantly reducing energy costs. Compared to residential and commercial electricity consumption, industrial plants have much greater electricity demands. Implementing demand response in industrial plants can provide flexible load solutions, improve power system stability, increase the development and utilization of renewable energy, and reduce energy costs.

[0003] In the industrial sector, industrial task networks dynamically adjust task execution based on demand response, thereby optimizing industrial electricity costs. Industrial scenarios are typically networked, with different network layers operating within different industrial systems. They are responsible for executing different types of industrial tasks. Previous research on energy management and demand response in industrial task networks usually only considered energy management and demand response within the current network layer, achieving optimization at a single network layer. However, with the continuous development of industrial modernization, current industrial tasks are becoming increasingly complex and multi-stage. The connections between different layers of industrial networks and different industrial systems are becoming increasingly close, meaning that the completion of an industrial task often requires the coordinated execution of multiple industrial production networks. For example… Figure 1 As shown, the execution of an industrial task requires the joint execution of three different industrial network layers. The failure of any one of the network layers will result in the final task not being completed.

[0004] While continuing to use traditional single-layer network energy management demand response techniques can achieve cost optimization for that single layer, it may also cause related network layer tasks to stagnate or become blocked. This means that the achieved local optimum may not be the global optimum, and could even be far from it. Waste disposal systems are typical examples of multi-networked industrial systems, such as... Figure 2As shown, in the waste treatment system of a certain waste treatment park, three layers of industrial networks are responsible for sludge treatment, waste incineration, and wastewater treatment, respectively. When the wastewater treatment system of network layer 3 reduces the wastewater treatment rate when the electricity price is high according to the demand response method, the wastewater generated by the waste incineration system of network layer 2 will be forced to accumulate because it cannot be treated in time. Once the accumulation reaches the upper limit, the waste incineration system of network layer 2 will be unable to perform the waste incineration task, which will further lead to the inability to incinerate the dried sludge generated by the sludge treatment system of network layer 1 in time, thus causing the paralysis of the entire system.

[0005] Therefore, it is necessary to propose a demand response energy management optimization method that is suitable for multi-networked waste treatment systems. This method needs to consider the task coupling relationship between multiple industrial networks, model and analyze the correlation between each layer of the network, and perform demand response from the perspective of the overall system, so as to achieve global optimization rather than optimization of a single layer of network. This reduces the system's electricity cost and improves the overall economic efficiency of the waste treatment system without affecting the overall task execution. Summary of the Invention

[0006] Technical Problem: The purpose of this invention is to propose a demand response energy management method for a multi-networked waste management system to solve the problems mentioned in the background. Specifically, the main part of this method is executed cyclically in two phases, once per hour. The first phase is the inter-network layer demand response method, which calculates the constraint impact of task resource coupling relationships between network layers, taking the overall system as the implementation object. The second phase is the intra-network layer demand response method, which adjusts the power demand based on the constraints obtained in the first phase and the specific task execution status within the network layer. According to this method, the recommended power consumption calculated in the second phase must satisfy the inter-layer constraints of the first phase. Based on this, the intra-layer component operation adjustment is achieved under the condition of satisfying the inter-layer constraints, that is, achieving global optimization rather than local optimization. This method can effectively reduce the overall system's power cost and solve the problem that the demand response under the previous single network could not consider the impact of multiple network coupling relationships on tasks.

[0007] Technical Solution: The power system adjusts electricity prices in real time based on the current power load, causing variations in the electricity costs for industrial components performing tasks at different times. Adopting a demand-response energy management scheme in industrial systems—reducing the execution rate of industrial tasks when electricity prices are high and increasing it when electricity prices are low—can significantly optimize electricity costs. In a waste management system with a multi-network context, different network layers handle different types of waste management tasks. Therefore, each network layer needs not only to rationally schedule component execution based on electricity prices at different times but also to schedule its own tasks based on the execution status of tasks in related network layers, maximizing system benefits while ensuring timely task completion. The main technical solutions of this demand-response scheduling strategy are as follows:

[0008] (1) Initialization. Using a calendar day as the scheduling target, it is stipulated that electricity price changes only occur at hourly boundaries, meaning the waste treatment system receives electricity price data once per hour. This price represents the actual electricity price for the current hour t. During the initialization phase, it is necessary to determine the minimum task completion target for the day (i.e., determine the minimum electricity demand for the day). Electricity price for the first hour .

[0009] (2) Vertical Inter-Layer Demand Response Phase. In this system, the network layer has resource dependencies on the upper-layer network. Specifically, this means that tasks executed between different network layers in the system have priority relationships. Upper-layer networks are required. The processed waste is used as the raw waste input for further waste processing, and the processed waste output is used as the lower-level network. The raw garbage input for the garbage disposal task. This stage computes the garbage at each layer of the network. For upper-layer networks This determines the resource dependence, thereby providing constraint boundaries for horizontal layer demand response calculations.

[0010] (3) Demand response stage within the horizontal layer. By using the electricity price of the past t-1 hours and the electricity price at the current time t, the electricity price range for the next 24-t hours is predicted. Based on this, the demand response model with uncertain electricity prices is transformed into a linear programming model based on a robust optimization model. Combining the current hourly electricity price and the load scheduling boundary constraints obtained in stage (2), the network at each layer is calculated. Optimal electricity demand for the current hour t And the electricity demand for the next few hours.

[0011] (4) Intra-layer component execution scheduling phase. Based on the recommended load power for the current time period obtained from the demand response phase above, the execution of schedulable components within the layer is adjusted so that the power consumption of each network layer reaches the optimal value calculated at the moment.

[0012] The above stages are executed once per hour to ensure timely updates of current electricity prices and future price ranges, thereby optimizing electricity costs based on real-time electricity prices. Beneficial effects

[0013] (1) Reduce energy costs in industrial plants. Through demand response, plants can adjust their electricity consumption strategies according to electricity prices, thereby avoiding excessive electricity consumption during peak periods. Without affecting task completion time, the load of components can be reasonably scheduled for different time periods, using more electricity when electricity prices are low and reducing electricity consumption when electricity prices are high, effectively reducing the plant's energy costs.

[0014] (2) Reduce the total time for the system to complete tasks. By responding to demands under multiple network relationships and taking the impact of the execution of related layer tasks on the execution of current network layer tasks as an important constraint, the execution of tasks can be reasonably scheduled and arranged, which can avoid task accumulation and congestion, thereby reducing the total time for the system to complete tasks at the overall level.

[0015] (3) Improve the stability and security of industrial power systems. In industrial systems, electricity consumption has obvious peak and trough periods. Frequent power fluctuations pose a huge challenge to the stability of the system, accelerate the aging of transmission lines and power loss. By adopting demand response methods, peak shaving and valley filling can be achieved, reducing the gap between peak and valley electricity demand, thereby improving the stability of industrial power systems. Attached Figure Description

[0016] Figure 1 This is an example diagram of a multi-industry network;

[0017] Figure 2 This is a schematic diagram of task execution in a multi-networked waste management system;

[0018] Figure 3 It refers to the relationship between the time subscripts t and h in the specific method;

[0019] Figure 4 This is a schematic diagram of the main principle of the method of the present invention. Detailed Implementation

[0020] We represent a multi-layered networked waste management system S containing N network layers as follows: Different network layers are responsible for performing different garbage disposal tasks, and adjacent network layers have task-resource coupling relationships, i.e., network layers Requires network layer The processed waste is used for task execution, and the processed waste is also used as the next layer of the network. The input, the last layer of the network The output is the clean resources obtained after all waste treatment processes are completed. Using a calendar day as the scheduling target, the completion status of waste treatment tasks is represented by electricity consumption. The minimum daily industrial task completion target is expressed as the minimum daily electricity consumption. The ultimate goal is to meet the daily task completion target. Minimize daily electricity costs The regulations stipulate that electricity price changes only occur at hourly boundaries, meaning S receives the electricity price once per hour. This price represents the electricity price for the next hour. The specific implementation steps are as follows:

[0021] (a) Initial Phase. This phase calculates the minimum task completion target for the day (i.e., determines the minimum electricity demand for the day). Electricity price for the first hour (Provided by the power company), maximum resource capacity of each network layer (Determined based on the capacity of each layer of the network in the actual waste treatment system).

[0022] (b) Vertical Inter-Layer Demand Response Phase. This phase calculates the resource coupling dependency constraints between layers and links the network layers... The task execution rate in time period t is equal to the garbage generation rate after processing by the current network layer. To indicate, network layer The rate of consumption of the original waste to be processed in time period t is expressed as: The unprocessed garbage in this layer is the upper-layer network. Processed waste; network layer The current amount of garbage processed in time period t is represented as follows: Network layer The electricity demand consumption in time period t is expressed as: Due to limitations of equipment, the network layer The maximum waste processing rate is expressed as Then the inter-layer network requirement response must meet the following:

[0023]

[0024] Among them, constraint (1) represents the network layer The initial garbage consumption rate at time t must satisfy the upper-layer network. Capacity limitations and non-negativity ensure the upper-layer network The amount of waste processed is never greater than that of the upper-layer network. Maximum storage limit; constraint (2) represents the network layer The initial garbage consumption rate at time t must satisfy the upper-layer network. The post-processing garbage retention limit, that is, to ensure the current network layer There is always raw waste available for processing; constraints (3) and (4) represent the network... At time t, the garbage processing rate must meet the current network requirements. The capacity limit and maximum processing rate limit are non-negative, ensuring that the current network layer... The amount of garbage processed is no higher than the maximum storage capacity, and the processing rate is no higher than the maximum processing rate; constraint (5) indicates the current network The functional relationship between power consumption and initial waste consumption and waste disposal can be used to analyze the network layer. The current task execution is converted into power consumption.

[0025] (c) Demand Response Phase within Horizontal Layers. This phase is responsible for calculating the optimal power demand for each network layer in the current time period. The objective function is: Must meet: Equation (6) indicates that the system optimization objective is to minimize the overall system's electricity cost for the current time period and the subsequent 24-t hours. This represents the actual electricity price at time t. This represents the uncertain electricity price at time t+h, and its value can be... any number; Indicates the network layer at time t Since the electricity price and electricity consumption at the first t-1 time point are already fixed, no additional constraints are needed for the electricity consumed; Equation (7) indicates that the daily electricity consumption must not be lower than the minimum electricity demand. Figure 3 This indicates the relationship between the time subscripts t and h.

[0026] Due to the electricity price after the current moment It is unknown and will fluctuate to some extent, so it is necessary to predict the future electricity price. Here, based on the electricity price of the past t-1 hour and the current hourly electricity price, the ARIMA model is used to predict the confidence interval of the electricity price for the next 24-t hours. The specific electricity value is limited to the electricity price range at a certain confidence level.

[0027] After obtaining the range of future electricity prices, a robust optimization model is used to solve for the objective function. The robust optimization model transforms the optimization objective of equation (6) into: Must meet: Equation (8) is the objective function derived from the duality of the robust optimization model and the linear equivalent transformation, where and It is the dual variable of the initial objective function (6), used to consider the influence of the price boundary; It is an auxiliary variable used to obtain linear expressions; This is a parameter used to control the robustness of the objective function; its value is a real number in the range [0, 2⁴ - t]. When the price deviation in the objective function is ignored, In this case, all price deviations in the objective function are taken into account, resulting in the most conservative but also the most costly solution. In practical applications, this can be flexibly adjusted based on the results. The value of .

[0028] After solving the above optimization model, the recommended power consumption for each layer of the network in the current hour is obtained. The system also updated the power demand for subsequent t+h hours. The next step is the in-layer component scheduling phase, with the following specific plan: ① First, the in-layer components are sensed and divided into schedulable and non-schedulable components to obtain the number of schedulable components in the network layer. Here are the definitions of schedulable and non-schedulable components: Schedulable components refer to devices in the current network layer whose usage can be adjusted. Their usage time or frequency can be changed to adjust power consumption. For example, the blower control component in the waste incineration layer is responsible for periodically blowing oxygen into the incinerator for waste combustion. The operation of this component can be controlled and it can operate at different times. Non-schedulable components refer to components in the current network layer whose operation cannot be adjusted, or whose operation would incur significant costs if adjusted. For example, the incinerator heating control component in the sludge treatment layer requires reheating the incinerator upon restarting after shutdown, resulting in very high costs. The system has multiple schedulable components. By enabling or disabling the use of schedulable components, the power consumption of network layer devices can be adjusted. ② Determine the relationship between the recommended power consumption and the current power demand. If the recommended power consumption is lower than the current power demand, shut down the use of schedulable components in sequence until the power consumption of the network layer reaches the recommended value. If the recommended power consumption is higher than the current power demand, start the schedulable components in sequence to increase the power demand of the network layer until the power consumption of the network layer reaches the recommended value.

[0029] The above stages run in a loop, executing once every hour at the hourly change boundary, until t=24. The specific execution flow is as follows: Figure 4 As shown.

[0030] To better understand the technical solution of the present invention, specific embodiments are described below:

[0031] like Figure 2 As shown, in a certain waste treatment system, the system is divided into three network layers according to the waste treatment tasks: sludge treatment layer, waste incineration layer, and wastewater treatment layer. There are task associations between different network layers.

[0032] First, during the initialization phase, the minimum waste disposal workload for the day is calculated. By converting the waste disposal workload into power consumption using historical data, the minimum power consumption target for the day is obtained. Initialize the power demand for the first hour to Then receive the electricity price for the first hour of the day. .

[0033] Secondly, in the demand response phase, based on the aforementioned vertical and horizontal demand response constraint models, the future hourly electricity price range is predicted, and the recommended electricity consumption for each network layer in the current time period is calculated. And the power demand for subsequent time periods, adjusting the operation of schedulable components based on recommended power consumption and the current power demand. Here, it is assumed that the calculated recommended power consumption for the sludge treatment layer in the first hour is... The amount of electricity required in the first hour is lower than the required amount in the first hour. Therefore, it is necessary to reduce the required power consumption. By sensing the sludge treatment layer components, the status of the schedulable components in this layer is as follows: the dewatering system can schedule sludge dewatering.

[0034] If the number of dryers is 'a' and the number of dryers that can be scheduled in the drying system is 'b', then the aforementioned scheduled components will be shut down sequentially, and the power demand for this layer during this time period will be adjusted accordingly. Other network layers are handled in the same way.

[0035] Receive the electricity price for the second hour at the start of the second hour. The above calculations are repeated to obtain recommended power consumption values, update power demand for subsequent time periods, and then adjust the operation of scheduleable components. This process is repeated until the end of the day. By using demand response to adjust the operation of scheduleable components from peak electricity price periods to off-peak periods, the overall electricity cost of the industrial system is reduced.

Claims

1. A method for optimizing energy management in a multi-networked waste treatment system, characterized in that, This method considers multiple industrial network task coupling relationships, including single-layer network execution and multiple network task flow execution. When adjusting the execution status of components in the demand response system, it also considers the coupling constraints of task execution within the same layer and task execution in adjacent layers. By predicting future electricity price ranges, and based on the current electricity price and the predicted future electricity price, the optimal recommended electricity volume for the current stage is calculated. The use of dispatchable components within the adjustment layer is then adjusted based on the recommended electricity volume, shifting the use of dispatchable components from periods with high electricity prices to periods with low electricity prices. An industrial system S containing N network layers is represented as follows: , Indicates the first Each network layer, Different network layers are responsible for performing different industrial tasks, and adjacent network layers have task resource coupling relationships. Consume network layer The generated materials are used for task execution, and the generated materials will also serve as the next layer of the network. The input, the last layer of the network The output is the final product; The scheduling objective is to adjust the execution status of component tasks over a natural day, optimize power costs while meeting a certain task volume, and represent the completion status of industrial tasks as power consumption. The minimum daily industrial task completion target is represented by the minimum daily electricity consumption. The ultimate goal is to meet the daily task completion target. Minimize daily electricity costs ,in This represents the electricity consumption in hour t. This represents the electricity price in hour t, where price changes only occur at hourly boundaries, meaning the industrial system receives the electricity price once per hour. This price represents the electricity price for the next hour; This method is mainly divided into an inter-layer demand response stage and an intra-layer demand response stage. The inter-layer demand response stage is used to calculate the task resource coupling constraints between adjacent network layers. The intra-layer demand response stage calculates the recommended electricity consumption of the network layer by predicting future electricity price ranges. The calculation of intra-layer demand response is based on the coupling relationship calculated by the inter-layer demand response, that is, the intra-layer demand response must meet the task resource coupling constraints determined by the inter-layer demand response stage. Task resource coupling constraints are represented as the dependency relationship between raw materials and products, specifically including: network layer The product generation rate in time period t is Network layer The rate of raw material consumption in time period t is expressed as: The raw material is the upper-layer network. The generated product; network layer The remaining product in time period t is represented as Network layer The electricity demand consumption in time period t is expressed as: Due to limitations of equipment, the network layer Having the maximum production rate, denoted as Network layer The raw material consumption rate at time t satisfies the upper-layer network. The capacity limitation and non-negativity make the upper-layer network The number of products is no greater than that of the upper-layer network. Maximum storage limit of materials; network layer The raw material consumption rate at time t satisfies the upper-layer network. The limited remaining quantity of products makes the current network layer Raw materials are used; network At time t, the product generation rate satisfies the current network speed. The capacity limit and maximum production rate limit are non-negative, making the current network layer... The quantity of products generated does not exceed the maximum storage capacity, and the production rate does not exceed the maximum production rate; current network The power consumption is functionally related to the raw material consumption and product generation, and the task execution is converted into power consumption accordingly. Inter-layer resource coupling dependency constraints are represented as raw material and product dependencies satisfying the following constraints: ; ; ; ; The objective of intra-layer demand response is to minimize electricity costs. A robust optimization model is constructed to handle variable real-time electricity prices. Through the duality property and linear equivalence transformation of the robust optimization model, the optimization objective is transformed into: minimizing... ;in This represents the actual electricity price at time t. This represents the uncertain electricity price at time t+h. and It is the dual variable of the target of the demand response within the layer, used to consider the impact of the price boundary; It is an auxiliary variable used to obtain linear expressions; It is a parameter used to control the robustness of the objective function; By using the ARIMA model to predict future electricity price ranges, and then using a robust optimization model to calculate the objective function, we obtain the recommended electricity consumption for the current time period and the projected electricity demand for subsequent time periods. In this process, the predicted electricity price range is used to calculate the following constraints in the robust optimization model: ; in This represents the upper boundary of the predicted electricity price range at time t+h. This indicates the lower boundary of the predicted electricity price range at time t+h; By using the constraints of the robust optimization model and the predicted electricity price range, a mixed-integer linear programming solver is employed to determine the recommended electricity consumption for the current time period. The recommended power consumption is fed back to each layer of the network. The network layer dynamically adjusts the operation of the schedulable components within the layer based on the recommended power consumption. The recommended power consumption is recalculated at the next moment, and the power demand is updated until the end of a cycle. The network layer receives the recommended power consumption. Next, the relationship between the recommended power consumption and the current power demand is determined. If the recommended power consumption is lower than the current power demand, the demand for schedulable components is reduced until the power consumption of the network layer is not higher than the recommended value. If the recommended power consumption is higher than the current power demand, the demand for schedulable components is increased until the power consumption of the network layer is not lower than the recommended value.

Images