TW202100472A - Formula decision making method for controlling water quality generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model - Google Patents

Formula decision making method for controlling water quality generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model Download PDF

Info

Publication number
TW202100472A
TW202100472A TW108121985A TW108121985A TW202100472A TW 202100472 A TW202100472 A TW 202100472A TW 108121985 A TW108121985 A TW 108121985A TW 108121985 A TW108121985 A TW 108121985A TW 202100472 A TW202100472 A TW 202100472A
Authority
TW
Taiwan
Prior art keywords
water quality
formula
predicted
data
quality parameter
Prior art date
Application number
TW108121985A
Other languages
Chinese (zh)
Inventor
謝文彬
吳若慧
吳振成
康志強
Original Assignee
謝文彬
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 謝文彬 filed Critical 謝文彬
Priority to TW108121985A priority Critical patent/TW202100472A/en
Priority to CN202010557796.6A priority patent/CN112132700A/en
Publication of TW202100472A publication Critical patent/TW202100472A/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A20/00—Water conservation; Efficient water supply; Efficient water use
    • Y02A20/152—Water filtration

Abstract

Provided is a formula decision making method for controlling water quality. The formula decision making method is implemented by virtue of a control system. The formula decision making method for controlling water quality includes the following step: (a) generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model which outputs at least one piece of formula data according to the input water quality parameter variation value, wherein each piece of the predicted formula data includes a predicted component of a formula and a predicted component proportion. By using the formula decision making method for controlling water quality, provided by this invention, a more accurate predicted formula which is capable of conforming to the economic benefit may be provided, the stability of a water quality treatment system may be more favorably improved, and problems caused by the abnormality of a water quality monitoring system may be eliminated.

Description

用於調控水質的配方決策方法Formula decision method for regulating water quality

本發明是有關於一種配方決策方法，特別是指一種用於調控水質的配方決策方法。The invention relates to a formula decision-making method, in particular to a formula decision-making method for regulating water quality.

現有水質調整或廢水處理大多採用被動式控管與監測方式，也就是在加入水處理藥劑配方至原水(或廢水)後，透過一監測系統來監測原水的變化，之後再經由此監測結果決定下一階段的處理，故需透過重複循環處理後才能達到預定的水質標準。然而，上述處理方式可能因等待監測結果至決定下一階段處理期間所耗費的時間過長，導致原水的性質又發生變化，同時也造成後續加入的藥劑配方無法即時因應此水質變化進行調整，致使配方精準度下降；其次，耗費時間過長還可能會導致處理系統或監測系統須隨時停機而讓處理或監測系統不穩定、產生異常、甚或損壞。Existing water quality adjustments or wastewater treatment mostly adopt passive control and monitoring methods, that is, after adding a water treatment agent formula to the raw water (or wastewater), a monitoring system is used to monitor the changes in the raw water, and then the monitoring results are used to determine the next Therefore, it is necessary to repeat the treatment cycle to reach the predetermined water quality standard. However, the above-mentioned treatment method may take too long to wait for the monitoring result to decide the next stage of treatment, which may cause the nature of the raw water to change again, and at the same time, the subsequent pharmaceutical formulations cannot be adjusted in response to this water quality change. The accuracy of the formulation decreases; secondly, too long a time consuming may also cause the processing system or monitoring system to be shut down at any time, making the processing or monitoring system unstable, abnormal, or even damaged.

隨著自動化及人工智慧的發展，目前也有關於水質監測方法的相關專利被提出，例如中華民國專利公告案TW I658273便提出一種水質監測方法，主要是透過監測一待測水體的多個水質參數與其預設安全值來判斷該待測水體的水質狀態；若水質狀態處於危險狀態則產生警示訊息至一行動裝置，以及啟動示警裝置或令水質改善模組啟動水質優化設備。上述水質監測方法僅能解決後端監測所遇到的問題，無法有效提供能即時因應水質變化的藥劑配方以及改善處理或監測系統的穩定性。With the development of automation and artificial intelligence, related patents on water quality monitoring methods have been proposed. For example, the Republic of China Patent Announcement TW I658273 proposed a water quality monitoring method, which mainly monitors multiple water quality parameters of a water body to be tested and its The preset safety value is used to determine the water quality status of the water body to be tested; if the water quality status is in a dangerous state, a warning message is generated to a mobile device, and the warning device is activated or the water quality improvement module activates the water quality optimization equipment. The above-mentioned water quality monitoring methods can only solve the problems encountered in back-end monitoring, and cannot effectively provide pharmaceutical formulations that can respond to changes in water quality immediately and improve the stability of the treatment or monitoring system.

由以上說明可知，現有水質調整方法仍須再設法因應各種水質提供更即時且精準的藥劑配方，同時改善處理系統或監測系統的穩定性。It can be seen from the above description that the existing water quality adjustment methods still need to try to provide more immediate and accurate pharmaceutical formulations in response to various water quality, while improving the stability of the treatment system or monitoring system.

因此，本發明之目的，即在提供一種能提供即時且精準的藥劑配方、並能同時改善處理或監測系統之穩定性的用於調控水質的配方決策方法。Therefore, the purpose of the present invention is to provide a formulation decision-making method for regulating water quality that can provide instant and accurate pharmaceutical formulations and improve the stability of the treatment or monitoring system.

於是，本發明用於調控水質的配方決策方法，藉由一調控系統來實施，該用於調控水質的配方決策方法包含以下步驟：(a)根據一期望的水質參數變化數值，利用一用於依據所輸入之水質參數變化數值輸出至少一配方資料的配方決策模型，產生至少一預測的配方資料，其中，每一預測的配方資料包含配方的一預測成分組成與一預測成分比例。Therefore, the formula decision method for regulating water quality of the present invention is implemented by a regulating system. The formula decision method for regulating water quality includes the following steps: (a) According to a desired water quality parameter change value, use a A formula decision model that outputs at least one formula data according to the input water quality parameter change value generates at least one predicted formula data, wherein each predicted formula data includes a predicted ingredient composition and a predicted ingredient ratio of the formula.

本發明之功效在於：本發明配方決策方法是藉由該配方決策模型而能有效地提供即時且精準的預測配方，且運用此預測配方進行水質調整可以避免處理系統或監測系統發生異常或損壞，並提升處理系統或監測系統的穩定性。The effect of the present invention is that the formula decision method of the present invention can effectively provide a real-time and accurate prediction formula through the formula decision model, and the use of this prediction formula for water quality adjustment can avoid abnormalities or damages to the treatment system or the monitoring system. And improve the stability of the processing system or monitoring system.

本發明將就以下實施例作進一步說明，但應瞭解的是，該實施例僅為例示說明之用，而不應被解釋為本發明實施之限制。The present invention will be further described in the following examples, but it should be understood that the examples are only for illustrative purposes and should not be construed as limiting the implementation of the present invention.

在本發明被詳細描述之前，應當注意在以下的說明內容中，類似的元件是以相同的編號來表示。Before the present invention is described in detail, it should be noted that in the following description, similar elements are represented by the same numbers.

於本文中所提到的「水質參數」涵蓋所有廢水或原水處理可能運用的水質參數，例如但不限於濁度、色度、有機物含量、金屬含量、無機物(如亞硝酸、硝酸鹽、銨離子、硫化氫、磷酸鹽、矽酸鹽、硫酸鹽等)含量、pH值、導電度、水溫、菌數(如大腸菌群、藍細菌、金黃色葡萄球菌、綠膿桿菌等)、溶氧量、硬度、氨濃度、氯濃度、鹽度、總懸浮固體濃度、氧化還原電位、化學需氧量及生物需氧量等。The "water quality parameters" mentioned in this article cover all water quality parameters that may be used in wastewater or raw water treatment, such as but not limited to turbidity, chroma, organic content, metal content, and inorganic substances (such as nitrous acid, nitrate, ammonium ion) , Hydrogen sulfide, phosphate, silicate, sulfate, etc.) content, pH, conductivity, water temperature, bacterial count (such as coliforms, cyanobacteria, Staphylococcus aureus, Pseudomonas aeruginosa, etc.), dissolved oxygen , Hardness, ammonia concentration, chlorine concentration, salinity, total suspended solids concentration, redox potential, chemical oxygen demand and biological oxygen demand, etc.

在本文中所提到的「成分」涵蓋各種化學藥劑、生物菌種、其他水質處理試劑等。The "ingredients" mentioned in this article cover various chemical agents, biological strains, and other water treatment reagents.

參閱圖1，實施本發明用於調控水質的配方決策方法的一實施例的一調控系統100，該調控系統100包含一電腦裝置1、一通訊連接該電腦裝置1的一水質處理裝置2，及一通訊連接該電腦裝置1的監測裝置3，該電腦裝置1包含一輸入單元11、一儲存運算單元12、一輸出單元13、及一通訊單元14。1, a control system 100 implementing an embodiment of the formula decision method for regulating water quality of the present invention, the control system 100 includes a computer device 1, a water quality treatment device 2 communicatively connected to the computer device 1, and A monitoring device 3 communicatively connected to the computer device 1, the computer device 1 includes an input unit 11, a storage operation unit 12, an output unit 13, and a communication unit 14.

該輸入單元11用於輸入一相關於水質參數變化數值的輸入訊號至該儲存運算單元12。該輸入單元11還可以選擇地輸入另一相關於輸入其他資料的輸入訊號至該儲存運算單元12，該其他資料包含一或多組配方資料。The input unit 11 is used to input an input signal related to the change value of the water quality parameter to the storage operation unit 12. The input unit 11 can also optionally input another input signal related to inputting other data to the storage operation unit 12, and the other data includes one or more sets of recipe data.

該儲存運算單元12儲存有一用於依據所輸入之水質參數變化數值輸出至少一配方資料的配方決策模型。該儲存運算單元12還儲存有多組訓練資料，每組訓練資料包括配方的一成分組成與一成分比例，以及該成分組成與成分比例所對應的一水質參數變化數值。The storage operation unit 12 stores a formula decision model for outputting at least one formula data according to the input water quality parameter change value. The storage computing unit 12 also stores multiple sets of training data. Each set of training data includes a component composition and a component ratio of the formula, and a water quality parameter change value corresponding to the component composition and the component ratio.

該輸出單元13用於輸出該配方決策模型所產生的至少一配方資料至該通訊單元14。The output unit 13 is used to output at least one recipe data generated by the recipe decision model to the communication unit 14.

該通訊單元14是通訊連接至該水質處理裝置2及該監測裝置3，以將該配方資料傳送至該水質處理裝置2及該監測裝置3。The communication unit 14 is communicatively connected to the water quality treatment device 2 and the monitoring device 3 to transmit the formula data to the water quality treatment device 2 and the monitoring device 3.

該水質處理裝置2泛指任何用於處理調控水質的裝置，例如但不限於物理處理單元、化學處理單元、生物處理單元、污泥處理單元、水回收處理單元等，上述單元可以單獨使用或組合使用。該水質處理裝置2較佳為自動化裝置。The water quality treatment device 2 generally refers to any device for treating and regulating water quality, such as but not limited to physical treatment unit, chemical treatment unit, biological treatment unit, sludge treatment unit, water recovery treatment unit, etc. The above units can be used alone or in combination use. The water quality treatment device 2 is preferably an automated device.

該監測裝置3用於監控檢測水質(包含調配前及調配後的水質)，並產生一實際水質參數。The monitoring device 3 is used to monitor and detect the water quality (including the water quality before and after the deployment), and generate an actual water quality parameter.

參閱圖2，本發明用於調控水質的配方決策方法之實施例1，藉由上述電腦裝置1來實施並包含一步驟21、一步驟22、一步驟23、一步驟24及一步驟25。Referring to FIG. 2, Embodiment 1 of the formula decision method for regulating water quality of the present invention is implemented by the above-mentioned computer device 1 and includes a step 21, a step 22, a step 23, a step 24, and a step 25.

在該步驟21中，該電腦裝置1之儲存運算單元12根據該等訓練資料，利用一機器學習演算法，建立該配方決策模型。該機器學習演算法例如但不限於：LinearRegression、RidgeCV、Ridge、AdaBoostRegressor、RandomForestRegressor、BaggingRegressor、ExtraTreeRegressor、XGBRegressor、GradientBoostingRegressor、Lasso、ElasticNet等，其中，當水質參數變化值為色度去除率或有機物II去除率時，機器學習演算法以XGBRegressor為最佳；當水質參數變化值為有機物I去除率或濁度去除率時，機器學習演算法以GradientBoostingRegressor為最佳。In this step 21, the storage computing unit 12 of the computer device 1 uses a machine learning algorithm to establish the formula decision model based on the training data. The machine learning algorithm is for example, but not limited to: LinearRegression, RidgeCV, Ridge, AdaBoostRegressor, RandomForestRegressor, BaggingRegressor, ExtraTreeRegressor, XGBRegressor, GradientBoostingRegressor, Lasso, ElasticNet, etc., where, when the water quality parameter change value is the color removal rate or the organic matter II When the water quality parameter change value is the organic matter I removal rate or turbidity removal rate, the machine learning algorithm uses the GradientBoostingRegressor as the best.

在步驟22中，該電腦裝置1之儲存運算單元12根據一回應於該輸入單元11之輸入訊號而產生之期望的水質參數變化數值，利用該配方決策模型，產生至少一個預測的配方資料，其中，每一個預測的配方資料包含配方的一預測成分組成與一預測成分比例。該步驟22的實施態樣例如但不限於下表1的具體例1至4： [表1] 具體例 編號 期望的水質參數變化數值 預測的配方資料：成分(比例) 1 色度去除率78 配方資料1：A藥劑(206.0)、B藥劑(14.40) 配方資料2：A藥劑(102.0)、B藥劑(19.55) 2 濁度去除率95 配方資料1：C藥劑(42.0)、D藥劑(208.0) 配方資料2：C藥劑(22.0)、D藥劑(178.0) 配方資料3：C藥劑(32.0)、D藥劑(193.0)   3 有機物I去除率28 配方資料1：A藥劑(102.0)、B藥劑(19.550) 配方資料2：A藥劑(128.0)、B藥劑(22.125) 配方資料3：A藥劑(154.0)、B藥劑(24.700) 4 有機物II去除率95 配方資料1：C藥劑(287.0)、D藥劑(208.0) 配方資料2：C藥劑(302.0)、D藥劑(228.0) 配方資料3：C藥劑(262.0)、D藥劑(223.0) 配方資料4：C藥劑(42.0)、D藥劑(208.0) In step 22, the storage computing unit 12 of the computer device 1 generates at least one predicted formula data based on a desired water quality parameter change value generated in response to the input signal of the input unit 11, using the formula decision model, wherein , Each predicted recipe data includes a predicted ingredient composition and a predicted ingredient ratio of the recipe. The implementation aspects of this step 22 are, for example, but not limited to, specific examples 1 to 4 in Table 1 below: [Table 1] Specific case number Formulation data for numerical prediction of expected water quality parameter changes: composition (proportion) 1 Chroma removal rate 78 Formulation data 1: A medicine (206.0), B medicine (14.40) Formulation data 2: A medicine (102.0), B medicine (19.55) 2 Turbidity removal rate 95 Formulation data 1: C medicine (42.0), D medicine (208.0) Formulation data 2: C medicine (22.0), D medicine (178.0) Formulation data 3: C medicine (32.0), D medicine (193.0) 3 Organic matter I removal rate 28 Formulation data 1: A medicine (102.0), B medicine (19.550) Formulation data 2: A medicine (128.0), B medicine (22.125) Formulation data 3: A medicine (154.0), B medicine (24.700) 4 Organic matter II removal rate 95 Formulation data 1: C medicine (287.0), D medicine (208.0) Formulation data 2: C medicine (302.0), D medicine (228.0) Formulation data 3: C medicine (262.0), D medicine (223.0) Formulation data 4: C medicine (42.0), D medicine (208.0)

以具體例1為例，當輸入色度去除率78作為期望的水質參數變化數值時，利用該配方決策模型將會產生2種配方資料，每種配方資料都包含成分組成與成分比例。Taking specific example 1 as an example, when the chromaticity removal rate 78 is input as the desired water quality parameter change value, the formula decision model will generate two formula data, each of which contains ingredient composition and ingredient ratio.

另外值得一提的是，該配方決策模型除了可以進行如步驟22流程，也可以根據欲使用的配方資料來獲得預測的水質參數變化數值，例如下表2的具體例I至IV： [表2] 具體例編號 欲使用的配方資料 預測的水質參數變化數值 成分組成 成分比例 I A藥劑與B藥劑 154：24.7 色度去除率 82.60577 II A藥劑與B藥劑 150：23 有機物I去除率 31.96711 III C藥劑與D藥劑 5.2：220 濁度去除率 90.78616 IV C藥劑與D藥劑 4.5：220 有機物II去除率 31.99935 It is also worth mentioning that the formula decision model can not only perform the process as in step 22, but also obtain the predicted water quality parameter change value according to the formula data to be used, for example, specific examples I to IV in Table 2 below: [Table 2 ] Specific case number Formulation information to be used Predicted changes in water quality parameters Composition component ratio I Medicine A and Medicine B 154: 24.7 Chroma removal rate 82.60577 II Medicine A and Medicine B 150:23 Organic matter I removal rate 31.96711 III Medicine C and Medicine D 5.2: 220 Turbidity removal rate 90.78616 IV Medicine C and Medicine D 4.5: 220 Organic matter II removal rate 31.99935

以具體例I為例，當輸入成分組成為A藥劑與B藥劑、以及成分比例為154：24.7時，可以利用該配方決策模型獲得預測的水質參數變化數值，即色度去除率82.60577。Taking specific example I as an example, when the input component composition is A drug and B drug, and the component ratio is 154:24.7, the formula decision model can be used to obtain the predicted water quality parameter change value, that is, the color removal rate of 82.60577.

在該步驟23中，該水質處理裝置2根據該步驟22所產生的至少一個預測的配方資料中之一目標配方資料對水質進行調控，以獲得一調配後的水質。值得一提的是，在本實施方式中，目標配方資料的決定可以是經由一使用者選取或是隨機選取。此外，該步驟23是根據該目標配方資料先配製藥劑配方，之後再經由該水質處理裝置2以機器自動化方式將藥劑配方加入欲處理的水中進行水質調控。其中，該水質處理裝置2可以是任何市售的水質處理裝置，甚至也可以使用任何市售的自動加藥裝置(圖未示)來將藥劑配方加入欲處理的水中進行水質調控。在本發明之其他實施方式中，該步驟23也可藉由人工方式先根據該目標配方資料配製藥劑配方後，再將藥劑配方加入欲處理的水中進行水質調控。In the step 23, the water quality treatment device 2 adjusts the water quality according to one of the target formula data in the at least one predicted formula data generated in the step 22 to obtain a formulated water quality. It is worth mentioning that, in this embodiment, the determination of the target formula data can be selected by a user or selected randomly. In addition, in step 23, a pharmaceutical formula is first prepared according to the target formula data, and then the pharmaceutical formula is added to the water to be treated through the water quality treatment device 2 in a machine-automated manner for water quality control. Wherein, the water quality treatment device 2 can be any commercially available water quality treatment device, and even any commercially available automatic dosing device (not shown) can be used to add a pharmaceutical formula to the water to be treated for water quality control. In other embodiments of the present invention, the step 23 can also manually prepare a pharmaceutical formula according to the target formula data, and then add the pharmaceutical formula to the water to be treated for water quality control.

在該步驟24中，該監測裝置3量測該調配後之水質以獲得一實際的水質參數變化數值。該監測裝置3可以是任何市售已知的水質監測裝置。在本發明之其他實施方式中，該步驟24也可藉由人工方式來量測該調配後之水質。In step 24, the monitoring device 3 measures the adjusted water quality to obtain an actual water quality parameter change value. The monitoring device 3 can be any commercially known water quality monitoring device. In other embodiments of the present invention, the step 24 can also manually measure the water quality after the preparation.

在該步驟25中，該電腦裝置1的儲存運算單元12根據該步驟24所獲得的該實際的水質參數變化數值，利用該機器學習演算法，修正該配方決策模型。在本實施例中，該機器學習演算法的實施態樣可以例如但不限於線性迴歸、類神經網路(NN)等。In this step 25, the storage computing unit 12 of the computer device 1 uses the machine learning algorithm to modify the formula decision model according to the actual water quality parameter change value obtained in the step 24. In this embodiment, the implementation of the machine learning algorithm can be, for example, but not limited to, linear regression, neural network (NN), etc.

參閱圖3，值得特別說明的是，該步驟25還進一步包含一子步驟251及一子步驟252。Referring to FIG. 3, it is worth noting that the step 25 further includes a sub-step 251 and a sub-step 252.

在該子步驟251中，根據該步驟24所量測的實際的水質參數變化數值與該步驟22所輸入的期望的水質參數變化數值來獲得一損失函數。在本實施例中，該子步驟251的實施態樣例如但不限於利用線性迴歸或類神經網路來實施。In this sub-step 251, a loss function is obtained according to the actual water quality parameter change value measured in step 24 and the expected water quality parameter change value input in step 22. In this embodiment, the implementation of the sub-step 251 is, for example, but not limited to, using linear regression or neural network-like implementation.

在該子步驟252中，根據該子步驟251的損失函數來修正該配方決策模型。In the sub-step 252, the formula decision model is modified according to the loss function of the sub-step 251.

參閱圖4，本發明用於調控水質的配方決策方法之實施例2，藉由上述調控系統100來實施並包含一步驟31、一步驟32、一步驟33、一步驟34、一步驟35及一步驟36。4, the second embodiment of the formulation decision method for regulating water quality of the present invention is implemented by the aforementioned regulating system 100 and includes one step 31, one step 32, one step 33, one step 34, one step 35, and one step. Step 36.

該步驟31與上述實施例1的步驟21類似，不同處在於：每組訓練資料還包括一配方成本，在該步驟31中，該電腦裝置1係根據每組包含該配方成本的訓練資料，利用該機器學習演算法，建立該配方決策模型。This step 31 is similar to step 21 of the above-mentioned embodiment 1, the difference is: each set of training data also includes a formula cost. In this step 31, the computer device 1 uses each set of training data that contains the formula cost. The machine learning algorithm establishes the formula decision model.

該步驟32與上述實施例1的步驟22類似，不同處在於：每一預測的配方資料包含配方的一預測成分組成、一預測成份比例與一預測配方成本。該步驟32的實施態樣例如但不限於下表3的具體例5至8： [表3] 具體例 編號 期望的水質參數變化數值 預測的配方資料：成分(比例)、成本(單位為元/m3 ) 5 色度去除率78 配方資料1：A藥劑(206.0)、B藥劑(14.40)、成本2.80472 配方資料2：A藥劑(102.0)、B藥劑(19.55)、成本1.49804 6 濁度去除率95 配方資料1：C藥劑(42.0)、D藥劑(208.0)、成本4.832 配方資料2：C藥劑(22.0)、D藥劑(178.0)、成本3.912 配方資料3：C藥劑(32.0)、D藥劑(193.0)、成本4.372 7 有機物I去除率28 配方資料1：A藥劑(102.0)、B藥劑(19.550)、成本1.49804 配方資料2：A藥劑(128.0)、B藥劑(22.125)、成本1.85870 配方資料3：A藥劑(154.0)、B藥劑(24.700)、成本2.21936 8 有機物II去除率95 配方資料1：C藥劑(287.0)、D藥劑(208.0)、成本8.752 配方資料2：C藥劑(302.0)、D藥劑(228.0)、成本9.392 配方資料3：C藥劑(262.0)、D藥劑(223.0)、成本8.652 配方資料4：C藥劑(42.0)、D藥劑(208.0)、成本4.832 This step 32 is similar to step 22 of the above-mentioned embodiment 1, except that: each predicted recipe data includes a predicted ingredient composition, a predicted ingredient ratio, and a predicted recipe cost of the recipe. The implementation of this step 32 is, for example, but not limited to, specific examples 5 to 8 in Table 3 below: [Table 3] Specific case number Formulation data for numerical prediction of expected water quality parameter changes: composition (proportion), cost (unit: yuan/m 3 ) 5 Chroma removal rate 78 Formulation data 1: A medicine (206.0), B medicine (14.40), cost 2.80472 Formulation data 2: A medicine (102.0), B medicine (19.55), cost 1.49904 6 Turbidity removal rate 95 Formulation data 1: C medicine (42.0), D medicine (208.0), cost 4.832 Formulation data 2: C medicine (22.0), D medicine (178.0), cost 3.912 Formulation data 3: C medicine (32.0) , D medicine (193.0), cost 4.372 7 Organic matter I removal rate 28 Formulation data 1: A medicine (102.0), B medicine (19.550), cost 1.49804 Formulation data 2: A medicine (128.0), B medicine (22.125), cost 1.85870 Formulation data 3: A medicine (154.0) , B medicine (24.700), cost 2.21936 8 Organic matter II removal rate 95 Formulation data 1: C medicine (287.0), D medicine (208.0), cost 8.752 Formulation data 2: C medicine (302.0), D medicine (228.0), cost 9.392 Formulation data 3: C medicine (262.0) , D medicine (223.0), cost 8.652 Formulation data 4: C medicine (42.0), D medicine (208.0), cost 4.832

該步驟33是自該至少一個預測的配方資料，選取出一對應有最少預測配方成本的目標配方資料。以上表3的具體例5至8為例進行選取，該步驟33的實施態樣例如但不限於下表4的具體例5至8： [表4] 具體例 編號 對應有最少預測配方成本的目標配方資料 5 配方資料2：A藥劑(102.0)、B藥劑(19.55)、成本1.49804 6 配方資料2：C藥劑(22.0)、D藥劑(178.0)、成本3.912 7 配方資料1：A藥劑(102.0)、B藥劑(19.550)、成本1.49804 8 配方資料4：C藥劑(42.0)、D藥劑(208.0)、成本4.832 In step 33, from the at least one predicted recipe data, a target recipe data corresponding to the least predicted recipe cost is selected. Specific examples 5 to 8 in Table 3 above are selected as examples, and the implementation aspects of step 33 are, for example, but not limited to, specific examples 5 to 8 in Table 4 below: [Table 4] Specific case number Corresponding to the target formula data with the least predicted formula cost 5 Formulation data 2: A drug (102.0), B drug (19.55), cost 1.49904 6 Formulation data 2: C medicine (22.0), D medicine (178.0), cost 3.912 7 Formulation data 1: Medicine A (102.0), Medicine B (19.550), cost 1.49904 8 Formulation data 4: C medicine (42.0), D medicine (208.0), cost 4.832

該步驟34與上述實施例1的步驟23類似，不同處在於:該步驟34是根據該目標配方資料對水質進行調控。This step 34 is similar to the step 23 of the aforementioned embodiment 1, except that the step 34 is to adjust the water quality according to the target formula data.

該步驟35及36分別與上述實施例1的步驟24及25類似。The steps 35 and 36 are similar to the steps 24 and 25 of the first embodiment described above, respectively.

另外值得一提的是，本發明配方決策方法也可用於檢視監測裝置3是否有鈍化現象，具體實施態樣例如：該電腦裝置1透過輸入欲使用的配方資料(包含成分組成與成分比例)，利用該配方決策模型來取得預測的水質參數變化數值，而後依據該配方資料配製藥劑配方並對水質進行調控，以獲得一調配後的水質；接著，該監測裝置3量測該調配後之水質，以獲得一實際的水質參數變化數值；該電腦裝置1根據該預測的水質參數變化數值與該實際的水質參數變化數值獲得一數據偏差值；最後根據該數據偏差值，判斷是否需經由該電腦裝置1產生監測裝置3有鈍化現象的警示。It is also worth mentioning that the formula decision method of the present invention can also be used to check whether the monitoring device 3 has passivation. For example, the computer device 1 inputs formula data (including ingredient composition and ingredient ratio) to be used. The formula decision model is used to obtain the predicted water quality parameter change value, and then the pharmaceutical formula is prepared according to the formula data and the water quality is adjusted to obtain a adjusted water quality; then, the monitoring device 3 measures the adjusted water quality, To obtain an actual water quality parameter change value; the computer device 1 obtains a data deviation value according to the predicted water quality parameter change value and the actual water quality parameter change value; finally, according to the data deviation value, it is determined whether the computer device needs to be passed through 1 Produce a warning that the monitoring device 3 has passivation.

綜上所述，本發明配方決策方法是藉由該配方決策模型而能有效地提供即時且精準的預測配方，且運用此預設測配方進行水質調整可以避免處理系統或監測系統發生異常或損壞，並提升處理系統或監測系統的穩定性，故確實能達成本發明之目的。To sum up, the formula decision method of the present invention can effectively provide real-time and accurate prediction formula through the formula decision model, and the use of this preset measurement formula for water quality adjustment can avoid abnormalities or damages to the treatment system or the monitoring system , And improve the stability of the processing system or monitoring system, so it can indeed achieve the purpose of the invention.

惟以上所述者，僅為本發明之實施例而已，當不能以此限定本發明實施之範圍，凡是依本發明申請專利範圍及專利說明書內容所作之簡單的等效變化與修飾，皆仍屬本發明專利涵蓋之範圍內。However, the above are only examples of the present invention. When the scope of implementation of the present invention cannot be limited by this, all simple equivalent changes and modifications made in accordance with the scope of the patent application of the present invention and the content of the patent specification still belong to This invention patent covers the scope.

1:電腦裝置 11:輸入單元 12:儲存運算單元 13:輸出單元 14:通訊單元 2:水質處理裝置 21~25:步驟 3:監測裝置 251~252:子步驟 31~36:步驟 100:調控系統 1: computer device 11: Input unit 12: Storage computing unit 13: output unit 14: Communication unit 2: Water quality treatment device 21~25: Step 3: Monitoring device 251~252: Substep 31~36: Step 100: control system

本發明之其他的特徵及功效，將於參照圖式的實施方式中清楚地呈現，其中： 圖1是一方塊圖，說明用以實施本發明用於調控水質的配方決策方法的一調控系統； 圖2是一流程圖，說明本發明用於調控水質的配方決策方法的實施例1的一配方決策方法的流程； 圖3是一流程圖，說明該實施例1之步驟25的子步驟流程；及 圖4是一流程圖，說明本發明用於調控水質的配方決策方法的實施例2的一配方決策方法的流程。Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, in which: Figure 1 is a block diagram illustrating a control system for implementing the formulation decision method for regulating water quality of the present invention; 2 is a flowchart illustrating the flow of a formula decision method of Embodiment 1 of the formula decision method for regulating water quality of the present invention; Figure 3 is a flowchart illustrating the sub-step flow of step 25 of the embodiment 1; and FIG. 4 is a flowchart illustrating the flow of a formula decision method of Embodiment 2 of the formula decision method for regulating water quality of the present invention.

21~25:步驟 21~25: Step

Claims (7)

一種用於調控水質的配方決策方法，藉由一調控系統來實施，該用於調控水質的配方決策方法包含以下步驟： (a)   根據一期望的水質參數變化數值，利用一用於依據所輸入之水質參數變化數值輸出至少一配方資料的配方決策模型，產生至少一預測的配方資料，其中，每一預測的配方資料包含配方的一預測成分組成與一預測成分比例。A formula decision method for regulating water quality is implemented by a regulating system. The formula decision method for regulating water quality includes the following steps: (a) According to an expected water quality parameter change value, a formula decision model for outputting at least one formula data according to the input water quality parameter change value is used to generate at least one predicted formula data, wherein each predicted formula data A predicted component composition and a predicted component ratio of the formula are included. 如請求項1所述的用於調控水質的配方決策方法，該調控系統儲存有多組訓練資料，每組訓練資料包括配方的一成分組成與一成分比例，以及該成份組成與該成分比例所對應的一水質參數變化數值，在步驟(a)之前，還包含以下步驟： (a0)  根據該等訓練資料，利用一機器學習演算法，建立該配方決策模型。The formula decision-making method for regulating water quality as described in claim 1, the regulating system stores multiple sets of training data, each set of training data includes a component composition and a component ratio of the formula, and the composition of the component and the ratio of the component The corresponding change value of a water quality parameter, before step (a), also includes the following steps: (a0) Based on the training data, a machine learning algorithm is used to establish the formula decision model. 如請求項1或2所述的用於調控水質的配方決策方法，其中，該水質參數變化數值中的水質參數是選自於濁度、色度、有機物含量、金屬含量、無機物含量、pH值、導電度、水溫、菌數、溶氧量、硬度、氨濃度、氯濃度、鹽度、總懸浮固體濃度、氧化還原電位、化學需氧量及生物需氧量之至少一者。The formula decision method for regulating water quality according to claim 1 or 2, wherein the water quality parameter in the water quality parameter change value is selected from the group consisting of turbidity, chroma, organic content, metal content, inorganic content, pH value , Electrical conductivity, water temperature, bacterial count, dissolved oxygen, hardness, ammonia concentration, chlorine concentration, salinity, total suspended solids concentration, redox potential, chemical oxygen demand and biological oxygen demand at least one. 如請求項2所述的用於調控水質的配方決策方法，於步驟(a)後還包含以下步驟： (b)  根據該步驟(a)的該等預測的配方資料中之一預測配方資料對水質進行調控，以獲得一調配後的水質； (c)  量測該調配後之水質以獲得一實際的水質參數變化數值； (d)  根據該實際的水質參數變化數值，利用該機器學習演算法，修正該配方決策模型。The formula decision-making method for regulating water quality as described in claim 2 further includes the following steps after step (a): (b) Regulate the water quality according to one of the predicted formula data in this step (a) to obtain a water quality after blending; (c) Measure the water quality after the deployment to obtain an actual water quality parameter change value; (d) According to the actual water quality parameter change value, use the machine learning algorithm to modify the formula decision model. 如請求項2所述的用於調控水質的配方決策方法，每組訓練資料還包括一配方成本，其中： 在該步驟(a0)中，根據每組包含該配方成本的訓練資料，利用該機器學習演算法，建立該配方決策模型； 在該步驟(a)中，每一預測的配方資料還包含一預測配方成本； 該用於調控水質的配方決策方法在該步驟(a)後還包含以下步驟； (e)   自該至少一預測的配方資料，選取出一對應有最少預測配方成本的目標配方資料。For the formula decision-making method for regulating water quality as described in claim 2, each set of training data also includes a formula cost, including: In this step (a0), use the machine learning algorithm to establish the formula decision model according to each set of training data containing the formula cost; In this step (a), each predicted recipe data also includes a predicted recipe cost; The formula decision-making method for regulating water quality further includes the following steps after step (a); (e) From the at least one predicted recipe data, select a target recipe data corresponding to the least predicted recipe cost. 如請求項5所述的用於調控水質的配方決策方法，在該步驟(e)之後，還包含以下步驟： (f)   根據步驟(b)的該目標配方資料對水質進行調控，以獲得一調配後的水質； (g)   量測該調配後之水質以獲得一實際的水質參數變化數值； (h)   根據該實際的水質參數變化數值，利用該機器學習演算法，修正該配方決策模型。The formula decision method for regulating water quality as described in claim 5, after this step (e), further includes the following steps: (f) Adjust the water quality according to the target formula data in step (b) to obtain a formulated water quality; (g) Measure the water quality after the deployment to obtain an actual water quality parameter change value; (h) According to the actual water quality parameter change value, the machine learning algorithm is used to modify the formula decision model. 如請求項6所述的用於調控水質的配方決策方法，其中，步驟(h)包含以下子步驟： (h-1) 根據該實際的水質參數變化數值與期望的該水質參數變化數值獲得一損失函數； (h-2) 根據該損失函數修正該配方決策模型。The formula decision-making method for regulating water quality according to claim 6, wherein step (h) includes the following sub-steps: (h-1) Obtain a loss function according to the actual water quality parameter change value and the expected water quality parameter change value; (h-2) Revise the formula decision model according to the loss function.
TW108121985A 2019-06-24 2019-06-24 Formula decision making method for controlling water quality generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model TW202100472A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
TW108121985A TW202100472A (en) 2019-06-24 2019-06-24 Formula decision making method for controlling water quality generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model
CN202010557796.6A CN112132700A (en) 2019-06-24 2020-06-18 Formula decision method for regulating and controlling water quality

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW108121985A TW202100472A (en) 2019-06-24 2019-06-24 Formula decision making method for controlling water quality generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model

Publications (1)

Publication Number Publication Date
TW202100472A true TW202100472A (en) 2021-01-01

Family

ID=73851432

Family Applications (1)

Application Number Title Priority Date Filing Date
TW108121985A TW202100472A (en) 2019-06-24 2019-06-24 Formula decision making method for controlling water quality generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model

Country Status (2)

Country Link
CN (1) CN112132700A (en)
TW (1) TW202100472A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI769902B (en) * 2021-07-30 2022-07-01 馮世宏 Cloud-based wastewater water quality monitoring and automatic dosing system

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114453141B (en) * 2022-01-17 2024-02-02 淮北矿业股份有限公司 Flotation agent for washing coking coal and preparation method thereof
CN117049617B (en) * 2023-08-14 2024-02-02 江苏秦郡环保科技有限公司 Sewage treatment control method and system for garbage incinerator slag

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7720615B2 (en) * 2006-03-20 2010-05-18 Sensis Corporation System for detection and prediction of water quality events
CN107219779A (en) * 2016-03-22 2017-09-29 周楠 The addition method and device of a kind of water treatment agent
CN109143840A (en) * 2018-09-18 2019-01-04 湖南柿竹园有色金属有限责任公司 A kind of mine tailing wastewater processing dosing closed loop uniform recipe design technology

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI769902B (en) * 2021-07-30 2022-07-01 馮世宏 Cloud-based wastewater water quality monitoring and automatic dosing system

Also Published As

Publication number Publication date
CN112132700A (en) 2020-12-25

Similar Documents

Publication Publication Date Title
TW202100472A (en) Formula decision making method for controlling water quality generating at least one piece of predicted formula data according to a desired water quality parameter variation value and utilizing a formula decision making model
US10640404B2 (en) Water treatment system and method
CA2863441C (en) Method of controlling addition rate of an odor control chemical
CN106007202A (en) Intelligent control system for sewage treatment plant key process parameters and control method of intelligent control system
CN109032061A (en) A kind of dosing intelligence control system for water process
dos Santos et al. Intelligent system for improving dosage control
JP6816131B2 (en) Neutralization plant
CN112759063A (en) Pre-ozone adding control method and control system thereof
CN214693460U (en) Ozone dosing control system in advance
CN202671285U (en) Automatic medicine feeding device for industrial sewage of coal-fired power plant
WO2013009106A9 (en) Method and apparatus for measuring a residual chloride concentration using a sensor, and water treatment system using the method and apparatus
SE1300156A1 (en) Method to optimize the chemical precipitation process in water and wastewater treatment plants
KR100573189B1 (en) Method and device for pH control of water supply facilities using tuning method of 2-DOF PID controller by neural network
CN110655229A (en) System for simulating and correcting capacity of water treatment plant
CN107662964A (en) Generating means of aqua sterilisa and preparation method thereof
CN105502689A (en) Coal slime water biological flocculation technology optimizing method
KR101334693B1 (en) Sensor and regression model based method of determining for injection amount of a coagulant, and purified-water treatment system using the same
CN108073073A (en) A kind of design method of ozone sterilization water-purifier
CN114997064A (en) Intelligent control model for high-toxicity sewage dosing treatment and algorithm thereof
Schmidt et al. Influence of molecular noise on the growth of single cells and bacterial populations
KR20190000499A (en) Water purification plant operation decision support system
JP2022117306A (en) Estimation device, prediction device, control device, estimation system, prediction system, control system, estimation program, prediction program, control program, estimation method, prediction method and control method
EP4046080A1 (en) A method and system for monitoring a process
Mussared et al. Implementation of predictive alum dose control systems
CN1257110C (en) Flocculating dispensing compound control method