描述

本項目將使用LM35溫度傳感器監測食品行業冷庫的溫度。冷藏單元的預定義閾值溫度值由操作員設置。在 PyCharm 軟件的幫助下，當前值顯示在監視器上。如果超過閾值溫度，則會通過 Twilio 向用戶發送 SMS 警報以及存儲單元的當前溫度讀數值。此外，LED 的點亮和蜂鳴器發出的蜂鳴聲提醒周圍的人員或操作員。溫度傳感器發送的數據存儲在 Bolt Cloud 中，這有助于使用機器學習隨時間進行溫度預測。Z-Score 分析有助于檢測溫度的任何突然變化。

電路連接

1 、LM 35溫度傳感器是本項目的主要單元。LM35 傳感器為 A0 引腳提供模擬輸入。Bolt 內置了模數轉換器 (ADC)，可將此模擬值轉換為數字值。然后這個數字值被發送到 Bolt Cloud。

握住 LM35 傳感器，使傳感器上寫有 LM35 文字的平面朝向您。在此配置中，左側引腳為VCC ，中間引腳為V 輸出，右側引腳為接地引腳。

連接面包板上的 LM35 傳感器。公頭/公頭跳線用于將 VCC 連接到 Bolt wifi 模塊的5V 引腳，V 輸出連接到A0 引腳，將 LM35 的接地引腳連接到 Bolt wifi 模塊的接地引腳 (GND) 。

2. LED 的短腿連接到面包板上的GND 引腳，而較長的腿連接到串聯的 330ohm 電阻的一端。電阻的另一端使用公/公跳線連接到GPIO （通用輸入/輸出）引腳 1 。

3. Buzzer的長短腳分別通過公/公跳線連接到GPIO 引腳 4和GND 引腳。

4.最后，使用 USB-A 轉 Micro-USB 電纜將 Bolt wifi 模塊連接到筆記本電腦/個人電腦/充電器，以便通電。

設備

完整的電路

?

設置 ProductOnBolt Cloud

1.設置 Bolt wifi 模塊。

2.在這里，您將找到您的 Bolt API 密鑰和 Bolt 設備 ID。保存。

3.轉到 PRODUCTS 并制作一個新產品，為其命名并選擇如下所示的選項，然后單擊 DONE。

4.進入配置，然后設置硬件并編寫預測溫度的代碼，然后保存。

5.將產品鏈接到設備并部署配置。

制作 TWILIO 帳戶

Twilio 是第三方 SMS 功能提供商。它是一家云通信平臺即服務 (PaaS) 公司。Twilio 允許軟件開發人員以編程方式撥打和接聽電話，還可以使用其 Web 服務 API 發送和接收文本消息。

1.使用您的郵件 ID 和頁面上提供的其他詳細信息進行注冊。然后 Twilio 將在您的郵件 ID 上發送一封驗證郵件，該郵件必須經過驗證。

2.然后提供將發送 OTP 的手機號碼進行驗證。

現在，您已入職。回答幾個問題，如下所示。

3.對于編碼，選擇“是”。然后選擇您喜歡的編碼語言。我選擇 Python。對于目標，選擇“在項目中使用 Twilio”。對于消息傳遞，請從下面給出的選項中進行選擇。我為我的項目選擇了SMS 。

4.現在，轉到 Twilio Dashboard 并選擇將提供試用號的選項。選擇那個號碼。該號碼將與您聯系。

5.最后，您將在 Twilio Dashboard 上獲得您的 Account SID、Auth Token、您的手機號碼和試用號。保存。

下載并安裝 PYCHARM

PyCharm is an integrated development environment (IDE) used in computer programming, specifically for the Python language. It is developed by the Czech company JetBrains.It provides code analysis, a graphical debugger, etc. and supports Data Science with Anaconda. PyCharm is cross-platform, with Windows, macOS and Linux versions.

1. Download PyCharm by visiting http://www.jetbrains.com/pycharm/ and click DOWNLOAD (shown in black). Then select the type of OS you have and then download either OPEN SOURCE (COMMUNITY) which is completely free or PROFESSIONAL which is for free trail. I took open source. Then click Download button and wait for it to download.

2. When it gets downloaded. Double click the file stored on your computer. Then follow the steps shown in the images below to install PyCharm.

3. Click NEXT. Select the folder where you want to install PyCharm. Select 64 bit launcher. Create a folder with name 'JetBrains'. Wait for PyCharm to install.

4. Click FINISH once installed. Click OK for 'Do not import settings'. Accept the terms and con5.ditions shown. Finally, click on 'Don't send'.

5. Now, open PyCharm. Select the screen colour DARCULA or INTELLIJ. Skip the next step (If you want, you can install UI themes as also). Then, wait for PyCharm to load. Then select the option CreateNewProject. Then give a name to your project. Make a new file as shown below and name it. Then you can start with your coding.

THEORY

Z-SCORE ANALYSIS

Z-scoreanalysis is used for anomaly detection i.e. anything outside of normal. When temperature increases or decreases suddenly, it is called anomaly and Z-score analysis detects this. Upper and lower bounds are found out using multiplication factor, frame size and input values from LM 35 sensor.

Mn is the mean of input data collected. Vi is the variance, r is the frame size i.e. number of input data points collected. I have collected 5 data points. C is the multiplication factor which is taken as 3 in my project. Tn gives the upper and lower threshold depending on + and - respectively in the formula shown.

WORKING

LM 35 sensor collects data every 5 minutes and sends it to Bolt Cloud.

I placed the setup in open for getting normal temperature. Then I put the setup in cold area quickly so that temperature decreases suddenly.

Then again, I placed the setup in open for normal temperature. Then, I took the setup in a hot area quickly so that temperature rise is quick.

The code checks if the temperature is within threshold limits or not by using data from LM 35 sensor. If it isn't in the limits, then Twilio sends the message on our mobile regarding threshold getting crossed.

The code also checks if there is sudden change in temperature (increase or decrease) beyond the upper and lower bounds. If there is, then Twilio sends the message on our mobile regarding sudden change in temperature.

Prediction graphs gets plotted using data from LM35 sensor to predict temperature in near future using machine learning. Sudden changes were made to temperature values to see the affect on the prediction.

The detailed explanation of code is given along with the code itself.

PREDICTIONGRAPH

PredictedPoints shows number of future data points of temperature taht need to be predicted. No. of Polynomial Coefficients (value of n) indicates how many coefficients or elements should be present in the function. Number of polynomial coefficients decide the nature of predict history graph (shown in red). This graph gets plotted based on the nature of actual graph (shown in blue) and no. of polynomial coefficients. It helps to tune the machine learning model. Frame size indicates how many points on actual graph should be taken into consideration for plotting predict history graph. Predicted Graph (shown in blue) shows the value and nature of future data points and graphs respectively. The 3 parameters shown in the figure have to changed and set such that predicted data most closely imitates the actual data. This lead to future data being most accurate.

OUTPUT

CodedOutputRegardingThresholdCrossing and Sudden Decrease in Temperature

MessageRegardingThresholdCrossing and Sudden Decrease in Temperature

CodedOutputRegardingThresholdCrossing and Sudden Increase in Temperature

MessageRegardingThresholdCrossing and Sudden Decrease in Temperature

?