JPH04322710A - Method for controlling filter bed by fuzzy interface - Google Patents

Abstract

PURPOSE:To optimize the macroscopic operatin of plural filter beds, to increase an use rate of clean water and at the same time to realize saving energy and long life when controlling a filter bed by a control device, such as a process controller. CONSTITUTION:3-stage fuzzy interfaces are used to judge each quantity of plural filter beds 5a to 5n. In the first fuzzy interface (1), a target figure of raw water flow is set by the integrated value of raw water flow, water head loss of each filter bed and the water level in a clean water reservoir. In the second fuzzy interface (2), washing of the filter bed is determined by the target figure of raw water flow, the number of the filter beds in operation and the clean water level. In the third fuzzy interface (3), the number of filter beds required in a filtration process is determined by the target figure of raw water flow and the clean water level. The result of these judgements makes a command to each filter bed.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a filter control system for controlling a filter using a control device such as a process controller, and more particularly to a filter control system that uses fuzzy inference to achieve optimal control.

0002

[Prior Art] Among the various processes in water treatment plants, there is a filtration process that removes small blocks and suspended matter that cannot be removed by the sedimentation process.By removing suspended matter, highly purified treated water can be obtained. can. A sand filtration method is a filtration process for removing suspended solids with a relatively low concentration, and a filtration pond is used. A filtration basin is a device that separates and retains suspended matter contained in water by passing water through a layer filled with inert filter media at an appropriate speed.If used for a long time, the filtration layer becomes dirty due to deposits, making it difficult to properly filter the filter. Since the speed cannot be maintained and the filtration ability decreases, cleaning of the filtration layer is required to avoid this. The conventional filter control method detects the instantaneous value of head loss,
The filtration ability is determined only based on this instantaneous value.

0003

[Problems to be Solved by the Invention] However, in the above-mentioned conventional filter control system, the cleaning timing is controlled only by the instantaneous value of head loss, which is proportional to the filtration speed and the amount of suspended matter suppressed. Since the capacity cannot be maximized and the capacity of the filtration basins has not been comprehensively determined, it is difficult to achieve optimal control when considering the macroscopic operation of multiple filtration basins.

[0004] The present invention was devised in view of these problems, and it is possible to make maximum use of the capacity of a filtration basin, optimize the macro operation of a plurality of filtration basins, and improve the utilization rate of purified water. At the same time, the objective is to provide a filter control method with a flexible algorithm that realizes power saving and life extension.

[0005]

[Means for Solving the Problems] The means for solving the above problems in the present invention is to remove suspended solids from raw water using a plurality of filtration ponds, and to wash the filtration ponds with purified water returned from the water purification ponds. In a filtration pond control method for a water purification plant system, a raw water flow rate target value is set using the first fuzzy inference based on the raw water integrated flow rate, head loss of each filtration basin, and water treatment pond water level, and this raw water flow rate target value and filtration during filtration are A second fuzzy inference is used to determine the cleaning of the filtration basins based on the number of ponds and the water level of the water purification pond, and a third fuzzy inference is used to determine the number of filtration basins for the required filtration process based on the raw water flow rate target value and the water level of the water purification pond. It will be based on the filter control method.

[0006]

[Operation] The present invention is a method that comprehensively determines the capacity of the filtration basin based on each quantity of the filtration process, such as head loss, integrated raw water flow rate, and water treatment tank water level, and determines the target raw water flow rate, determines the cleaning of the filtration basin, and so on. The number of filters is determined using fuzzy reasoning. In the present invention, three stages of fuzzy inference are provided. In the first fuzzy inference, a raw water flow rate target value is set based on the raw water integrated flow rate, the head loss of each filtration basin, and the water treatment pond water level, and in the second fuzzy inference, this Cleaning of the filtration basin is determined from the target raw water flow rate, the number of filtration basins undergoing filtration, and the water level of the water purification pond, and in the third fuzzy inference, the cleaning of the filtration basin for the required filtration process is determined from the target raw water flow rate and the water level of the water purification pond. Decide on the number of machines.

[0007]

Embodiments Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a configuration diagram of an embodiment of the present invention. In the figure, 1 is the first fuzzy inference (1), 2 is the second fuzzy inference (2), 3 is the third fuzzy inference (3),
4 is a settling basin pit, 5a to 5n are a plurality of filter basins, and 6 is a water purification basin. In the same figure, the solid line indicates the water system, and the water that is filtered from the settling tank pit 4 through the raw water valves 7a to 7n and the flow meters 8a to 8n and in each of the filtration tanks 5a to 5n is transferred to the water purification tank 6.
A part of it is returned to the filter basins 5a to 5n by the cleaning pump 9 and used for cleaning them. As a filter control method for such a system, in this embodiment, the raw water flow rate detected by the flowmeters 8a to 8n is integrated by the integrator 10, and the raw water integrated flow rate and the head loss of each filter basin 5a to 5n are calculated. , and the water level of the water purification pond 6 using the first fuzzy inference (1
) to set the target value of raw water flow rate, and
1, the raw water valves 7a to 7n are controlled. Based on the target value of the raw water flow rate, the number of filters undergoing filtration, and the water level of the water purification basin 6, the second fuzzy inference (2) determines cleaning of the filtration basin and issues a filtration basin cleaning command. Third fuzzy inference (
3) determines the number of filters for the required filtration process from the target value of the raw water flow rate and the water level of the water purification tank 6, and issues a filter filter command.

Next, the fuzzy inference of the embodiment will be explained. It is well known that fuzzy inference is established using membership functions and rule matrices, and these are as follows.

(1) In the first fuzzy inference, the membership function has three types as phenomenon items: loss head (LS), raw water integrated flow rate (FQ), and water purification tank water level (JL), and as cause items: There is a raw water flow rate target value (SQ), and three stages of L, M, and S are set for each, and the rule matrix for this is as shown in the table below.

0010

[Table 1]

(2) In the second fuzzy inference, the membership function is based on the target raw water flow rate (SQ) as a phenomenon item.
, Number of filtration ponds (NR) and water purification pond water level (JL) during filtration
The three types and the cause item are filtration basin cleaning directives (SE).
There are three stages of L, M, and S, respectively.
The rule matrix for this is shown in the table below.

0012

[Table 2]

(3) In the third fuzzy inference, the membership function is based on the target raw water flow rate (SQ) as a phenomenon item.
, There are two types of water purification tank water level (JL) and a cause item, filtration tank filtration command (RK), respectively L, M, and S.
Three stages are set, and the rule matrix for these stages is shown in the table below.

[0014]

[Table 3]

[0015] In this embodiment, each command to the filtration basin is determined by such three fuzzy inferences.

The following effects are evident in this embodiment.

(1) The capacity of the filtration pond can be utilized to the fullest.

(2) The macroscopic operation of multiple filters can be optimized.

(3) It is possible to reduce the number of times a filtration basin, which uses a large amount of purified water, is washed, thereby improving the utilization rate of purified water, shortening the operating time of washing pumps, etc., and saving power and extending the service life.

(4) Filtration basin cleaning commands and the like can be made using fuzzy reasoning to construct flexible algorithms, which can be easily changed and modified.

[0021]

[Effects of the Invention] As explained above, according to the present invention, it is possible to make maximum use of the capacity of the filtration basin, optimize the macroscopic operation of multiple filtration basins, and clean the filtration basins that use a large amount of purified water. A filtration control method based on fuzzy reasoning that reduces the number of times the cleaning pump is used, improves the utilization rate of purified water, shortens the operating time of the cleaning pump, saves power, and extends the life of the cleaning pump.It also has a flexible algorithm that can be easily changed and modified. can be provided.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram of an embodiment of the present invention.

[Explanation of symbols]

1, 2, 3...fuzzy reasoning, 4...sedimentation basin pit, 5...filtration basin, 6...purification pond, 7...raw water valve, 8...flow meter, 9...washing pump, 10...integrator, 11...PI controller .

Claims (1)

[Claims] Claim 1. A filtration pond control method for a water treatment plant system in which suspended solids are removed from raw water using a plurality of filtration ponds, and the filtration ponds are cleaned with purified water returned from the water treatment ponds,
A raw water flow rate target value is set using the first fuzzy inference based on the raw water integrated flow rate, the head loss of each filtration pond, and the water purification pond water level, and the second raw water flow rate target value is set from this raw water flow target value, the number of filter ponds under filtration, and the water purification pond water level. A filtration pond control method characterized in that cleaning of the filtration basin is determined by fuzzy inference, and the number of filtration units for a required filtration process is determined by third fuzzy inference from the raw water flow rate target value and the water purification pond water level.